mirror-github-bspeice/fractorium
1
0
Fork 0
mirror of https://bitbucket.org/mfeemster/fractorium.git synced 2025-01-22 05:30:06 -05:00
Code Issues Actions Packages Projects Releases Wiki Activity
75ffc2f665
fractorium/Source/Ember/Variations04.h

5769 lines
179 KiB
C
Raw Normal View History

Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
#pragma once
#include "Variation.h"
namespace EmberNs
{
/// <summary>
/// eSwirl.
/// </summary>
template <typename T>
--Code changes -Split VariationList into .h/cpp files. --Make VariationList.cpp the only place where the variations files are included from. This speeds up the build, but variations can no longer be directly instantiated. --Variations are no longer exported.
2016-02-18 21:58:24 -05:00
class ESwirlVariation : public ParametricVariation<T>
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
public:
--Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums.
2016-01-04 19:50:15 -05:00
ESwirlVariation(T weight = 1.0) : ParametricVariation<T>("eSwirl", eVariationId::VAR_ESWIRL, weight, true)
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
Init();
}
PARVARCOPY(ESwirlVariation)
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
T tmp = helper.m_PrecalcSumSquares + 1;
T tmp2 = 2 * helper.In.x;
--Code changes -Make usage of VarFuncs more consistent by moving some global functions there. -Make CircleTrans1 a little safer by calling Zeps() on m_Sc during precalc. -Begin prepping for 0.9.9.5 release.
2016-03-04 21:54:06 -05:00
T xmax = (VarFuncs<T>::SafeSqrt(tmp + tmp2) + VarFuncs<T>::SafeSqrt(tmp - tmp2)) * T(0.5);
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
ClampGteRef<T>(xmax, -1);
--User changes -Add new variations: bubbleT3D, crob, hexaplay3D, hexcrop, hexes, hexnix3D, loonie2, loonie3, nBlur, octapol and synth. -Allow for pre/post versions of dc_bubble, dc_cylinder and dc_linear whereas before they were omitted. -When saving a file with multiple embers in it, detect if time values are all the same and if so, start them at zero and increment by 1 for each ember. -Allow for numerous quality increases to be coalesced into one. It will pick up at the end of the current render. -Show selection highlight on variations tree in response to mouse hover. This makes it easier to see for which variation or param the current mouse wheel action will apply. -Make default temporal samples be 100, whereas before it was 1000 which was overkill. -Require the shift key to be held with delete for deleting an ember to prevent it from triggering when the user enters delete in the edit box. -This wasn't otherwise fixable without writing a lot more code. --Bug fixes -EmberGenome was crashing when generating a sequence from a source file with more than 2 embers in it. -EmberGenome was improperly handling the first frame of a merge after the last frame of the loop. -These bugs were due to a previous commit. Revert parts of that commit. -Prevent a zoom value of less than 0 when reading from xml. -Slight optimization of the crescents, and mask variations, if the compiler wasn't doing it already. -Unique file naming was broken because it was looking for _# and the default names ended with -#. -Disallow renaming of an ember in the library tree to an empty string. -Severe bug that prevented some variations from being read correctly from params generated outside this program. -Severe OpenCL randomization bug. The first x coordinates of the first points in the first kernel call of the first ember of a render since the OpenCL renderer object was created were not random and were mostly -1. -Severe bug when populating xform selection distributions that could sometimes cause a crash due to roundoff error. Fix by using double. -Limit the max number of variations in a random ember to MAX_CL_VARS, which is 8. This ensures they'll look the same on CPU and GPU. -Prevent user from saving stylesheet to default.qss, it's a special reserved filename. --Code changes -Generalize using the running sum output point inside of a variation for all cases: pre, reg and post. -Allow for array variables in variations where the address of each element is stored in m_Params. -Qualify all math functions with std:: -No longer use our own Clamp() in OpenCL, instead use the standard clamp(). -Redesign how functions are used in the variations OpenCL code. -Add tests to EmberTester to verify some of the new functionality. -Place more const and override qualifiers on functions where appropriate. -Add a global rand with a lock to be used very sparingly. -Use a map instead of a vector for bad param names in Xml parsing. -Prefix affine interpolation mode defines with "AFFINE_" to make their purpose more clear. -Allow for variations that change state during iteration by sending a separate copy of the ember to each rendering thread. -Implement this same functionality with a local struct in OpenCL. It's members are the total of all variables that need to change state within an ember. -Add Contains() function to Utils.h. -EmberRender: print names of kernels being printed with --dump_kernel option. -Clean up EmberTester to handle some of the recent changes. -Fix various casts. -Replace % 2 with & 1, even though the compiler was likely doing this already. -Add new file Variations06.h to accommodate new variations. -General cleanup.
2015-11-22 17:15:07 -05:00
T mu = std::acosh(xmax);
T nu = std::acos(Clamp<T>(helper.In.x / xmax, -1, 1));//-Pi < nu < Pi.
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
if (helper.In.y < 0)
nu *= -1;
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
nu = nu + mu * m_Out + m_In / mu;
--User changes -Add new variations: bubbleT3D, crob, hexaplay3D, hexcrop, hexes, hexnix3D, loonie2, loonie3, nBlur, octapol and synth. -Allow for pre/post versions of dc_bubble, dc_cylinder and dc_linear whereas before they were omitted. -When saving a file with multiple embers in it, detect if time values are all the same and if so, start them at zero and increment by 1 for each ember. -Allow for numerous quality increases to be coalesced into one. It will pick up at the end of the current render. -Show selection highlight on variations tree in response to mouse hover. This makes it easier to see for which variation or param the current mouse wheel action will apply. -Make default temporal samples be 100, whereas before it was 1000 which was overkill. -Require the shift key to be held with delete for deleting an ember to prevent it from triggering when the user enters delete in the edit box. -This wasn't otherwise fixable without writing a lot more code. --Bug fixes -EmberGenome was crashing when generating a sequence from a source file with more than 2 embers in it. -EmberGenome was improperly handling the first frame of a merge after the last frame of the loop. -These bugs were due to a previous commit. Revert parts of that commit. -Prevent a zoom value of less than 0 when reading from xml. -Slight optimization of the crescents, and mask variations, if the compiler wasn't doing it already. -Unique file naming was broken because it was looking for _# and the default names ended with -#. -Disallow renaming of an ember in the library tree to an empty string. -Severe bug that prevented some variations from being read correctly from params generated outside this program. -Severe OpenCL randomization bug. The first x coordinates of the first points in the first kernel call of the first ember of a render since the OpenCL renderer object was created were not random and were mostly -1. -Severe bug when populating xform selection distributions that could sometimes cause a crash due to roundoff error. Fix by using double. -Limit the max number of variations in a random ember to MAX_CL_VARS, which is 8. This ensures they'll look the same on CPU and GPU. -Prevent user from saving stylesheet to default.qss, it's a special reserved filename. --Code changes -Generalize using the running sum output point inside of a variation for all cases: pre, reg and post. -Allow for array variables in variations where the address of each element is stored in m_Params. -Qualify all math functions with std:: -No longer use our own Clamp() in OpenCL, instead use the standard clamp(). -Redesign how functions are used in the variations OpenCL code. -Add tests to EmberTester to verify some of the new functionality. -Place more const and override qualifiers on functions where appropriate. -Add a global rand with a lock to be used very sparingly. -Use a map instead of a vector for bad param names in Xml parsing. -Prefix affine interpolation mode defines with "AFFINE_" to make their purpose more clear. -Allow for variations that change state during iteration by sending a separate copy of the ember to each rendering thread. -Implement this same functionality with a local struct in OpenCL. It's members are the total of all variables that need to change state within an ember. -Add Contains() function to Utils.h. -EmberRender: print names of kernels being printed with --dump_kernel option. -Clean up EmberTester to handle some of the recent changes. -Fix various casts. -Replace % 2 with & 1, even though the compiler was likely doing this already. -Add new file Variations06.h to accommodate new variations. -General cleanup.
2015-11-22 17:15:07 -05:00
helper.Out.x = m_Weight * std::cosh(mu) * std::cos(nu);
helper.Out.y = m_Weight * std::sinh(mu) * std::sin(nu);
--User changes -Allow for pausing the renderer in the main window. This makes is more efficient when entering many parameters, such as when following a tutorial. -Add support for new variations: erf, gamma, jac_cn, jac_dn, jac_sn, logDB, pressure_wave, pRose3D, splits3D, w, waves2b, x, xerf, y, z. -Inform user of the start and stop of file parsing in EmberAnimate because the files could potentially be very large. -Move the follwing fields to a new table called Animation: Interpolation, Affine Interpolation, Temporal Samples, Temporal Filter Width, Temporal Filter Type. -These currently have no effect on the interactive renderer and instead are used when running flames through EmberGenome to generate sequences, and then animating them in Fractorium or EmberAnimate. -Add new parameter overrides for EmberRender and EmberAnimate which directly assign values to all flames being rendered, rather than scale: --quality --demin --demax --Bug fixes -Left pad instead of right pad names of sequence outputs from EmberGenome. -Unique file naming was broken for files which already had an underscore in them. -Properly report that png is the default format of EmberRender and EmberAnimate output instead of erroneously claiming it was jpg. -Make command line programs search these folders in this order for the palette file: ./ ~/.fractorium ~/.config/fractorium /usr/share/fractorium /usr/local/share/fractorium -Fix possible bad values in hexes. -Better assignment of Z variables. -Fix boarders due to use of poorly implemented rint() function from flam3. Use std::rint() now. -wedge_sph was completely wrong due to having accidentally swapped the mapping of two parameters. -Make juliascope juliascope_power parameter be of type REAL_NONZERO since it's used as a denominator. -Make Z assignment compatible with the originals in: -arch, bcircle, bCollide, bent, bent2, bisplit, blob, blur_linear, blur_square, bMod, boarders, boarders2, bSwirl, bTransform, butterfly, cardioid, cell, circleblur, circlize, circlize2, circus, collideoscope, cos, cosine, cosh, coth, cpow, cpow2, crescents, cropn, csc, csch, curl, curve, dc_gridout, deltaa, diamond, disc2, eclipse, eCollide, edisc, eJulia, elliptic, eMod, eMotion, ennepers, epispiral, ePush, eRotate, eScale, eSwirl, ex, exp, expo, exponential, fan, fdisc, fibonacci, fibonacci2, fisheye, flipcircle, flipy, flower, flux, funnel, glynnia, GlynnSim1, GlynnSim2, GlynnSim3, gridout, handkerchief, heart, hole, idisc, julia, julian2, juliaNab, kaleidoscope, lazyTravis, Lissajous, mask, MobiusN, mobius_strip, modulus, murl, murl2, npolar, ortho, oscilloscope, parabola, perspective, petal, phoenix_julia, pie (was also inconsistent between cpu and gpu), poincare, popcorn, popcorn2, power, pow_block, rational3, rays, rblur, rings, rippled, roundspher, sec, secant2, sigmoid, sin, sineblur, sinh, sinusgrid, sphericaln, spiralwing, spirograph, split, squarize, squirrel, squish, sschecks, starblur, stripes, stwin, super_shape, tan, tancos, tangent, tanh, TwinTrian, twoface, unpolar, waves, wavesn, wedge_julia, whorl, xheart, zblur, zscale. --Code changes -Generalize Variation::PrecalcHelper() and rename to PrePostPrecalcHelper(). --Do the same for the OpenCL version and rename it PrePostPrecalcOpenCLString(). -Rename Variation::m_AssignType to m_PrePostAssignType since it's only relevant to pre/post variations.
2016-01-29 20:02:15 -05:00
helper.Out.z = DefaultZ(helper);
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
}
--Code changes -Better use of const functions. -Ember.pro should not define BUILDING_EMBERCL, instead define BUILDING_EMBER.
2015-06-29 23:13:53 -04:00
virtual string OpenCLString() const override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
ostringstream ss, ss2;
0.4.1.3 Beta 10/14/2014 --User Changes Size is no longer fixed to the window size. Size scaling is done differently in the final render dialog. This fixes several bugs. Remove Xml saving size from settings and options dialog, it no longer applies. Final render can be broken into strips. Set default save path to the desktop if none is found in the settings file. Set default output size to 1920x1080 if none is found in the settings file. --Bug Fixes Better memory size reporting in final render dialog. --Code Changes Migrate to C++11, Qt 5.3.1, and Visual Studio 2013. Change most instances of unsigned int to size_t, and int to intmax_t. Add m_OrigPixPerUnit and m_ScaleType to Ember for scaling purposes. Replace some sprintf_s() calls in XmlToEmber with ostringstream. Move more non-templated members into RendererBase. Add CopyVec() overload that takes a per element function pointer. Add vector Memset(). Replace '&' with '+' instead of "&amp;" in XmlToEmber for much faster parsing. Break strips rendering out into EmberCommon and call from EmberRender and Fractorium. Make AddAndWriteBuffer() just call WriteBuffer(). Make AddAndWriteImage() delete the existing image first before replacing it. Add SetOutputTexture() to RendererCL to support making new textures in response to resize events. Remove multiple return statements in RendererCL, and replace with a bool that tracks results. Add ToDouble(), MakeEnd(), ToString() and Exists() wrappers in Fractorium. Add Size() wrapper in EmberFile. Make QString function arguments const QString&, and string with const string&. Make ShowCritical() wrapper for invoking a message box from another thread. Add combo box to TwoButtonWidget and rename.
2014-10-14 11:53:15 -04:00
intmax_t i = 0, varIndex = IndexInXform();
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
ss2 << "_" << XformIndexInEmber() << "]";
string index = ss2.str();
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
string weight = WeightDefineString();
string in = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
string out = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
ss << "\t{\n"
<< "\t\treal_t tmp = precalcSumSquares + 1;\n"
<< "\t\treal_t tmp2 = 2 * vIn.x;\n"
Ditch T(f) define in opencl programs in preference to already-defined real_t. Also removed some erroneously-added double-'T(' statements
2015-02-26 16:18:55 -05:00
<< "\t\treal_t xmax = (SafeSqrt(tmp + tmp2) + SafeSqrt(tmp - tmp2)) * (real_t)(0.5);\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\n"
<< "\t\tif (xmax < 1)\n"
<< "\t\t xmax = 1;\n"
<< "\n"
<< "\t\treal_t mu = acosh(xmax);\n"
--User changes -Add new variations: bubbleT3D, crob, hexaplay3D, hexcrop, hexes, hexnix3D, loonie2, loonie3, nBlur, octapol and synth. -Allow for pre/post versions of dc_bubble, dc_cylinder and dc_linear whereas before they were omitted. -When saving a file with multiple embers in it, detect if time values are all the same and if so, start them at zero and increment by 1 for each ember. -Allow for numerous quality increases to be coalesced into one. It will pick up at the end of the current render. -Show selection highlight on variations tree in response to mouse hover. This makes it easier to see for which variation or param the current mouse wheel action will apply. -Make default temporal samples be 100, whereas before it was 1000 which was overkill. -Require the shift key to be held with delete for deleting an ember to prevent it from triggering when the user enters delete in the edit box. -This wasn't otherwise fixable without writing a lot more code. --Bug fixes -EmberGenome was crashing when generating a sequence from a source file with more than 2 embers in it. -EmberGenome was improperly handling the first frame of a merge after the last frame of the loop. -These bugs were due to a previous commit. Revert parts of that commit. -Prevent a zoom value of less than 0 when reading from xml. -Slight optimization of the crescents, and mask variations, if the compiler wasn't doing it already. -Unique file naming was broken because it was looking for _# and the default names ended with -#. -Disallow renaming of an ember in the library tree to an empty string. -Severe bug that prevented some variations from being read correctly from params generated outside this program. -Severe OpenCL randomization bug. The first x coordinates of the first points in the first kernel call of the first ember of a render since the OpenCL renderer object was created were not random and were mostly -1. -Severe bug when populating xform selection distributions that could sometimes cause a crash due to roundoff error. Fix by using double. -Limit the max number of variations in a random ember to MAX_CL_VARS, which is 8. This ensures they'll look the same on CPU and GPU. -Prevent user from saving stylesheet to default.qss, it's a special reserved filename. --Code changes -Generalize using the running sum output point inside of a variation for all cases: pre, reg and post. -Allow for array variables in variations where the address of each element is stored in m_Params. -Qualify all math functions with std:: -No longer use our own Clamp() in OpenCL, instead use the standard clamp(). -Redesign how functions are used in the variations OpenCL code. -Add tests to EmberTester to verify some of the new functionality. -Place more const and override qualifiers on functions where appropriate. -Add a global rand with a lock to be used very sparingly. -Use a map instead of a vector for bad param names in Xml parsing. -Prefix affine interpolation mode defines with "AFFINE_" to make their purpose more clear. -Allow for variations that change state during iteration by sending a separate copy of the ember to each rendering thread. -Implement this same functionality with a local struct in OpenCL. It's members are the total of all variables that need to change state within an ember. -Add Contains() function to Utils.h. -EmberRender: print names of kernels being printed with --dump_kernel option. -Clean up EmberTester to handle some of the recent changes. -Fix various casts. -Replace % 2 with & 1, even though the compiler was likely doing this already. -Add new file Variations06.h to accommodate new variations. -General cleanup.
2015-11-22 17:15:07 -05:00
<< "\t\treal_t nu = acos(clamp(vIn.x / xmax, -(real_t)(1.0), (real_t)(1.0)));\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\n"
<< "\t\tif (vIn.y < 0)\n"
<< "\t\t nu *= -1;\n"
<< "\n"
<< "\t\tnu = nu + mu * " << out << " + " << in << " / mu;\n"
<< "\n"
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
<< "\t\tvOut.x = " << weight << " * cosh(mu) * cos(nu);\n"
<< "\t\tvOut.y = " << weight << " * sinh(mu) * sin(nu);\n"
--User changes -Allow for pausing the renderer in the main window. This makes is more efficient when entering many parameters, such as when following a tutorial. -Add support for new variations: erf, gamma, jac_cn, jac_dn, jac_sn, logDB, pressure_wave, pRose3D, splits3D, w, waves2b, x, xerf, y, z. -Inform user of the start and stop of file parsing in EmberAnimate because the files could potentially be very large. -Move the follwing fields to a new table called Animation: Interpolation, Affine Interpolation, Temporal Samples, Temporal Filter Width, Temporal Filter Type. -These currently have no effect on the interactive renderer and instead are used when running flames through EmberGenome to generate sequences, and then animating them in Fractorium or EmberAnimate. -Add new parameter overrides for EmberRender and EmberAnimate which directly assign values to all flames being rendered, rather than scale: --quality --demin --demax --Bug fixes -Left pad instead of right pad names of sequence outputs from EmberGenome. -Unique file naming was broken for files which already had an underscore in them. -Properly report that png is the default format of EmberRender and EmberAnimate output instead of erroneously claiming it was jpg. -Make command line programs search these folders in this order for the palette file: ./ ~/.fractorium ~/.config/fractorium /usr/share/fractorium /usr/local/share/fractorium -Fix possible bad values in hexes. -Better assignment of Z variables. -Fix boarders due to use of poorly implemented rint() function from flam3. Use std::rint() now. -wedge_sph was completely wrong due to having accidentally swapped the mapping of two parameters. -Make juliascope juliascope_power parameter be of type REAL_NONZERO since it's used as a denominator. -Make Z assignment compatible with the originals in: -arch, bcircle, bCollide, bent, bent2, bisplit, blob, blur_linear, blur_square, bMod, boarders, boarders2, bSwirl, bTransform, butterfly, cardioid, cell, circleblur, circlize, circlize2, circus, collideoscope, cos, cosine, cosh, coth, cpow, cpow2, crescents, cropn, csc, csch, curl, curve, dc_gridout, deltaa, diamond, disc2, eclipse, eCollide, edisc, eJulia, elliptic, eMod, eMotion, ennepers, epispiral, ePush, eRotate, eScale, eSwirl, ex, exp, expo, exponential, fan, fdisc, fibonacci, fibonacci2, fisheye, flipcircle, flipy, flower, flux, funnel, glynnia, GlynnSim1, GlynnSim2, GlynnSim3, gridout, handkerchief, heart, hole, idisc, julia, julian2, juliaNab, kaleidoscope, lazyTravis, Lissajous, mask, MobiusN, mobius_strip, modulus, murl, murl2, npolar, ortho, oscilloscope, parabola, perspective, petal, phoenix_julia, pie (was also inconsistent between cpu and gpu), poincare, popcorn, popcorn2, power, pow_block, rational3, rays, rblur, rings, rippled, roundspher, sec, secant2, sigmoid, sin, sineblur, sinh, sinusgrid, sphericaln, spiralwing, spirograph, split, squarize, squirrel, squish, sschecks, starblur, stripes, stwin, super_shape, tan, tancos, tangent, tanh, TwinTrian, twoface, unpolar, waves, wavesn, wedge_julia, whorl, xheart, zblur, zscale. --Code changes -Generalize Variation::PrecalcHelper() and rename to PrePostPrecalcHelper(). --Do the same for the OpenCL version and rename it PrePostPrecalcOpenCLString(). -Rename Variation::m_AssignType to m_PrePostAssignType since it's only relevant to pre/post variations.
2016-01-29 20:02:15 -05:00
<< "\t\tvOut.z = " << DefaultZCl()
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t}\n";
return ss.str();
}
--User changes -Add new variations: bubbleT3D, crob, hexaplay3D, hexcrop, hexes, hexnix3D, loonie2, loonie3, nBlur, octapol and synth. -Allow for pre/post versions of dc_bubble, dc_cylinder and dc_linear whereas before they were omitted. -When saving a file with multiple embers in it, detect if time values are all the same and if so, start them at zero and increment by 1 for each ember. -Allow for numerous quality increases to be coalesced into one. It will pick up at the end of the current render. -Show selection highlight on variations tree in response to mouse hover. This makes it easier to see for which variation or param the current mouse wheel action will apply. -Make default temporal samples be 100, whereas before it was 1000 which was overkill. -Require the shift key to be held with delete for deleting an ember to prevent it from triggering when the user enters delete in the edit box. -This wasn't otherwise fixable without writing a lot more code. --Bug fixes -EmberGenome was crashing when generating a sequence from a source file with more than 2 embers in it. -EmberGenome was improperly handling the first frame of a merge after the last frame of the loop. -These bugs were due to a previous commit. Revert parts of that commit. -Prevent a zoom value of less than 0 when reading from xml. -Slight optimization of the crescents, and mask variations, if the compiler wasn't doing it already. -Unique file naming was broken because it was looking for _# and the default names ended with -#. -Disallow renaming of an ember in the library tree to an empty string. -Severe bug that prevented some variations from being read correctly from params generated outside this program. -Severe OpenCL randomization bug. The first x coordinates of the first points in the first kernel call of the first ember of a render since the OpenCL renderer object was created were not random and were mostly -1. -Severe bug when populating xform selection distributions that could sometimes cause a crash due to roundoff error. Fix by using double. -Limit the max number of variations in a random ember to MAX_CL_VARS, which is 8. This ensures they'll look the same on CPU and GPU. -Prevent user from saving stylesheet to default.qss, it's a special reserved filename. --Code changes -Generalize using the running sum output point inside of a variation for all cases: pre, reg and post. -Allow for array variables in variations where the address of each element is stored in m_Params. -Qualify all math functions with std:: -No longer use our own Clamp() in OpenCL, instead use the standard clamp(). -Redesign how functions are used in the variations OpenCL code. -Add tests to EmberTester to verify some of the new functionality. -Place more const and override qualifiers on functions where appropriate. -Add a global rand with a lock to be used very sparingly. -Use a map instead of a vector for bad param names in Xml parsing. -Prefix affine interpolation mode defines with "AFFINE_" to make their purpose more clear. -Allow for variations that change state during iteration by sending a separate copy of the ember to each rendering thread. -Implement this same functionality with a local struct in OpenCL. It's members are the total of all variables that need to change state within an ember. -Add Contains() function to Utils.h. -EmberRender: print names of kernels being printed with --dump_kernel option. -Clean up EmberTester to handle some of the recent changes. -Fix various casts. -Replace % 2 with & 1, even though the compiler was likely doing this already. -Add new file Variations06.h to accommodate new variations. -General cleanup.
2015-11-22 17:15:07 -05:00
virtual vector<string> OpenCLGlobalFuncNames() const override
{
return vector<string> { "SafeSqrt" };
}
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
protected:
void Init()
{
string prefix = Prefix();
m_Params.clear();
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
m_Params.push_back(ParamWithName<T>(&m_In, prefix + "eSwirl_in"));
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
m_Params.push_back(ParamWithName<T>(&m_Out, prefix + "eSwirl_out"));
}
private:
T m_In;
T m_Out;
};
--User changes: -Show common folder locations such as documents, downloads, pictures in the sidebar in all file dialogs. -Warning message about exceeding memory in final render dialog now suggests strips as the solution to the problem. -Strips now has a tooltip explaining what it does. -Allow more digits in the spinners on the color section the flame tab. -Add manually adjustable size spinners in the final render dialog. Percentage scale and absolute size are fully synced. -Default prefix in final render is now the filename when doing animations (coming from sequence section of the library tab). -Changed the elliptic variation back to using a less precise version for float, and a more precise version for double. The last release had it always using double. -New applied xaos table that shows a read-only view of actual weights by taking the base xform weights and multiplying them by the xaos values. -New table in the xaos tab that gives a graphical representation of the probability that each xform is chosen, with and without xaos. -Add button to transpose the xaos rows and columns. -Add support for importing .chaos files from Chaotica. --Pasting back to Chaotica will work for most, but not all, variations due to incompatible parameter names in some. -Curves are now splines instead of Bezier. This adds compatibility with Chaotica, but breaks it for Apophysis. Xmls are still pastable, but the color curves will look different. --The curve editor on the palette tab can now add points by clicking on the lines and remove points by clicking on the points themselves, just like Chaotica. --Splines are saved in four new xml fields: overall_curve, red_curve, green_curve and blue_curve. -Allow for specifying the percentage of a sub batch each thread should iterate through per kernel call when running with OpenCL. This gives a roughly 1% performance increase due to having to make less kernel calls while iterating. --This field is present for interactive editing (where it's not very useful) and in the final render dialog. --On the command line, this is specified as --sbpctth for EmberRender and EmberAnimate. -Allow double clicking to toggle the supersample field in the flame tab between 1 and 2 for easily checking the effect of the field. -When showing affine values as polar coordinates, show angles normalized to 360 to match Chaotica. -Fuse Count spinner now toggles between 15 and 100 when double clicking for easily checking the effect of the field. -Added field for limiting the range in the x and y direction that the initial points are chosen from. -Added a field called K2 which is an alternative way to set brightness, ignored when zero. --This has no effect for many variations, but hs a noticeable effect for some. -Added new variations: arcsech arcsech2 arcsinh arctanh asteria block bwraps_rand circlecrop2 coth_spiral crackle2 depth_blur depth_blur2 depth_gaussian depth_gaussian2 depth_ngon depth_ngon2 depth_sine depth_sine2 dragonfire dspherical dust excinis exp2 flipx flowerdb foci_p gaussian glynnia2 glynnsim4 glynnsim5 henon henon hex_rand hex_truchet hypershift lazyjess lens lozi lozi modulusx modulusy oscilloscope2 point_symmetry pointsymmetry projective pulse rotate scry2 shift smartshape spher squares starblur2 swirl3 swirl3r tanh_spiral target0 target2 tile_hlp truchet_glyph truchet_inv truchet_knot unicorngaloshen vibration vibration2 --hex_truchet, hex_rand should always use double. They are extremely sensitive. --Bug fixes: -Bounds sign was flipped for x coordinate of world space when center was not zero. -Right clicking and dragging spinner showed menu on mouse up, even if it was very far away. -Text boxes for size in final render dialog were hard to type in. Same bug as xform weight used to be so fix the same way. -Fix spelling to be plural in toggle color speed box. -Stop using the blank user palette to generate flames. Either put colored palettes in it, or exclude it from randoms. -Clicking the random palette button for a palette file with only one palette in it would freeze the program. -Clicking none scale in final render did not re-render the preview. -Use less precision on random xaos. No need for 12 decimal places. -The term sub batch is overloaded in the options dialog. Change the naming and tooltip of those settings for cpu and opencl. --Also made clear in the tooltip for the default opencl quality setting that the value is per device. -The arrows spinner in palette editor appears like a read-only label. Made it look like a spinner. -Fix border colors for various spin boxes and table headers in the style sheet. Requires reload. -Fix a bug in the bwraps variation which would produce different results than Chaotica and Apophysis. -Synth was allowed to be selected for random flame generation when using an Nvidia card but it shouldn't have been because Nvidia has a hard time compiling synth. -A casting bug in the OpenCL kernels for log scaling and density filtering was preventing successful compilations on Intel iGPUs. Fixed even though we don't support anything other than AMD and Nvidia. -Palette rotation (click and drag) position was not being reset when loading a new flame. -When the xform circles were hidden, opening and closing the options dialog would improperly reshow them. -Double click toggle was broken on integer spin boxes. -Fixed tab order of some controls. -Creating a palette from a jpg in the palette editor only produced a single color. --Needed to package imageformats/qjpeg.dll with the Windows installer. -The basic memory benchmark test flame was not really testing memory. Make it more spread out. -Remove the temporal samples field from the flame tab, it was never used because it's only an animation parameter which is specified in the final render dialog or on the command line with EmberAnimate. --Code changes: -Add IsEmpty() to Palette to determine if a palette is all black. -Attempt to avoid selecting a blank palette in PaletteList::GetRandomPalette(). -Add function ScanForChaosNodes() and some associated helper functions in XmlToEmber. -Make variation param name correction be case insensitive in XmlToEmber. -Report error when assigning a variation param value in XmlToEmber. -Add SubBatchPercentPerThread() method to RendererCL. -Override enterEvent() and leaveEvent() in DoubleSpinBox and SpinBox to prevent the context menu from showing up on right mouse up after already leaving the spinner. -Filtering the mouse wheel event in TableWidget no longer appears to be needed. It was probably an old Qt bug that has been fixed. -Gui/ember syncing code in the final render dialog needed to be reworked to accommodate absolute sizes.
2019-04-13 22:00:46 -04:00
/// <summary>
/// lazyjess.
/// By FarDareisMai.
/// </summary>
template <typename T>
class LazyJessVariation : public ParametricVariation<T>
{
public:
LazyJessVariation(T weight = 1.0) : ParametricVariation<T>("lazyjess", eVariationId::VAR_LAZYJESS, weight, true, true)
{
Init();
}
PARVARCOPY(LazyJessVariation)
virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
{
T theta, sina, cosa;
T x = helper.In.x;
T y = helper.In.y;
T modulus = helper.m_PrecalcSqrtSumSquares;
// n==2 requires a special case
if (m_N == T(2))
{
if (std::abs(x) < m_Weight) // If the input point falls inside the designated area...
{
// // ...then rotate it.
theta = std::atan2(y, x) + m_Spin;
sina = std::sin(theta);
cosa = std::cos(theta);
x = m_Weight * modulus * cosa;
y = m_Weight * modulus * sina;
if (std::abs(x) < m_Weight)
{
helper.Out.x = x;
helper.Out.y = y;
}
else // If it is now part of a corner that falls outside the designated area...
{
// ...then flip and rotate into place.
theta = std::atan2(y, x) - m_Spin + m_CornerRotation;
sina = std::sin(theta);
cosa = std::cos(theta);
helper.Out.x = m_Weight * modulus * cosa;
helper.Out.y = -(m_Weight * modulus * sina);
}
}
else
{
modulus = 1 + m_Space / Zeps(modulus);
helper.Out.x = m_Weight * modulus * x;
helper.Out.y = m_Weight * modulus * y;
}
}
else
{
// Calculate the distance r from origin to the edge of the polygon at the angle of the input point.
theta = std::atan2(y, x) + M_2PI;
T theta_diff = std::fmod(theta + m_HalfSlice, m_PieSlice);
T r = m_Weight * T(M_SQRT2) * m_SinVertex / Zeps(std::sin(T(M_PI) - theta_diff - m_Vertex));
if (modulus < r)
{
// Again, rotating points within designated area.
theta = std::atan2(y, x) + m_Spin + M_2PI;
sina = std::sin(theta);
cosa = std::cos(theta);
x = m_Weight * modulus * cosa;
y = m_Weight * modulus * sina;
// Calculating r and modulus for our rotated point.
theta_diff = std::fmod(theta + m_HalfSlice, m_PieSlice);
r = m_Weight * T(M_SQRT2) * m_SinVertex / Zeps(std::sin(T(M_PI) - theta_diff - m_Vertex));
modulus = VarFuncs<T>::Hypot(x, y);
if (modulus < r)
{
helper.Out.x = x;
helper.Out.y = y;
}
else
{
// Again, flipping and rotating corners that fall outside the designated area.
theta = std::atan2(y, x) - m_Spin + m_CornerRotation + M_2PI;
sina = std::sin(theta);
cosa = std::cos(theta);
helper.Out.x = m_Weight * modulus * cosa;
helper.Out.y = -(m_Weight * modulus * sina);
}
}
else
{
modulus = 1 + m_Space / Zeps(modulus);
helper.Out.x = m_Weight * modulus * x;
helper.Out.y = m_Weight * modulus * y;
}
}
helper.Out.z = DefaultZ(helper);
}
virtual string OpenCLString() const override
{
ostringstream ss, ss2;
intmax_t i = 0, varIndex = IndexInXform();
ss2 << "_" << XformIndexInEmber() << "]";
string index = ss2.str();
string weight = WeightDefineString();
string n = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string spin = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string space = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string corner = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string vertex = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string sinvertex = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string pieslice = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string halfslice = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string cornerrotation = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
ss << "\t{\n"
<< "\t\treal_t theta, sina, cosa;\n"
<< "\t\treal_t x = vIn.x;\n"
<< "\t\treal_t y = vIn.y;\n"
<< "\t\treal_t modulus = precalcSqrtSumSquares;\n"
<< "\n"
<< "\t\tif (" << n << " == 2.0)\n"
<< "\t\t{\n"
<< "\t\t if (fabs(x) < " << weight << ")\n"
<< "\t\t {\n"
<< "\t\t theta = atan2(y, x) + " << spin << ";\n"
<< "\t\t sina = sin(theta);\n"
<< "\t\t cosa = cos(theta);\n"
<< "\t\t x = " << weight << " * modulus * cosa;\n"
<< "\t\t y = " << weight << " * modulus * sina;\n"
<< "\n"
<< "\t\t if (fabs(x) < " << weight << ")\n"
<< "\t\t {\n"
<< "\t\t vOut.x = x;\n"
<< "\t\t vOut.y = y;\n"
<< "\t\t }\n"
<< "\t\t else\n"
<< "\t\t {\n"
<< "\t\t theta = atan2(y, x) - " << spin << " + " << cornerrotation << ";\n"
<< "\t\t sina = sin(theta);\n"
<< "\t\t cosa = cos(theta);\n"
<< "\t\t vOut.x = " << weight << " * modulus * cosa;\n"
<< "\t\t vOut.y = -(" << weight << " * modulus * sina);\n"
<< "\t\t }\n"
<< "\t\t }\n"
<< "\t\t else\n"
<< "\t\t {\n"
<< "\t\t modulus = 1 + " << space << " / Zeps(modulus);\n"
<< "\t\t vOut.x = " << weight << " * modulus * x;\n"
<< "\t\t vOut.y = " << weight << " * modulus * y;\n"
<< "\t\t }\n"
<< "\t\t}\n"
<< "\t\telse\n"
<< "\t\t{\n"
<< "\t\t theta = atan2(y, x) + M_2PI;\n"
<< "\t\t real_t theta_diff = fmod(theta + " << halfslice << ", " << pieslice << ");\n"
<< "\t\t real_t r = " << weight << " * M_SQRT2 * " << sinvertex << " / Zeps(sin(MPI - theta_diff - " << vertex << "));\n"
<< "\n"
<< "\t\t if (modulus < r)\n"
<< "\t\t {\n"
<< "\t\t theta = atan2(y, x) + " << spin << " + M_2PI;\n"
<< "\t\t sina = sin(theta);\n"
<< "\t\t cosa = cos(theta);\n"
<< "\t\t x = " << weight << " * modulus * cosa;\n"
<< "\t\t y = " << weight << " * modulus * sina;\n"
<< "\t\t theta_diff = fmod(theta + " << halfslice << ", " << pieslice << ");\n"
<< "\t\t r = " << weight << " * M_SQRT2 * " << sinvertex << " / Zeps(sin(MPI - theta_diff - " << vertex << "));\n"
<< "\t\t modulus = Hypot(x, y);\n"
<< "\n"
<< "\t\t if (modulus < r)\n"
<< "\t\t {\n"
<< "\t\t vOut.x = x;\n"
<< "\t\t vOut.y = y;\n"
<< "\t\t }\n"
<< "\t\t else\n"
<< "\t\t {\n"
<< "\t\t theta = atan2(y, x) - " << spin << " + " << cornerrotation << " + M_2PI;\n"
<< "\t\t sina = sin(theta);\n"
<< "\t\t cosa = cos(theta);\n"
<< "\t\t vOut.x = " << weight << " * modulus * cosa;\n"
<< "\t\t vOut.y = -(" << weight << " * modulus * sina);\n"
<< "\t\t }\n"
<< "\t\t }\n"
<< "\t\t else\n"
<< "\t\t {\n"
<< "\t\t modulus = 1 + " << space << " / Zeps(modulus);\n"
<< "\t\t vOut.x = " << weight << " * modulus * x;\n"
<< "\t\t vOut.y = " << weight << " * modulus * y;\n"
<< "\t\t }\n"
<< "\t\t}\n"
<< "\t\tvOut.z = " << DefaultZCl()
<< "\t}\n";
return ss.str();
}
virtual void Precalc() override
{
m_Vertex = T(M_PI) * (m_N - 2) / Zeps(2 * m_N);
m_SinVertex = std::sin(m_Vertex);
m_PieSlice = M_2PI / Zeps(m_N);
m_HalfSlice = m_PieSlice / 2;
m_CornerRotation = (m_Corner - 1) * m_PieSlice;
}
virtual vector<string> OpenCLGlobalFuncNames() const override
{
return vector<string> { "Zeps", "Hypot" };
}
protected:
void Init()
{
string prefix = Prefix();
m_Params.clear();
m_Params.push_back(ParamWithName<T>(&m_N, prefix + "lazyjess_n", 4, eParamType::INTEGER_NONZERO, 2));
m_Params.push_back(ParamWithName<T>(&m_Spin, prefix + "lazyjess_spin", T(M_PI), eParamType::REAL_CYCLIC, 0, M_2PI));
m_Params.push_back(ParamWithName<T>(&m_Space, prefix + "lazyjess_space"));
m_Params.push_back(ParamWithName<T>(&m_Corner, prefix + "lazyjess_corner", 1, eParamType::INTEGER_NONZERO));
m_Params.push_back(ParamWithName<T>(true, &m_Vertex, prefix + "lazyjess_vertex"));//Precalc.
m_Params.push_back(ParamWithName<T>(true, &m_SinVertex, prefix + "lazyjess_sin_vertex"));
m_Params.push_back(ParamWithName<T>(true, &m_PieSlice, prefix + "lazyjess_pie_slice"));
m_Params.push_back(ParamWithName<T>(true, &m_HalfSlice, prefix + "lazyjess_half_slice"));
m_Params.push_back(ParamWithName<T>(true, &m_CornerRotation, prefix + "lazyjess_corner_rotation"));
}
private:
T m_N;
T m_Spin;
T m_Space;
T m_Corner;
T m_Vertex;//Precalc.
T m_SinVertex;
T m_PieSlice;
T m_HalfSlice;
T m_CornerRotation;
};
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
/// <summary>
/// lazyTravis.
/// </summary>
template <typename T>
--Code changes -Split VariationList into .h/cpp files. --Make VariationList.cpp the only place where the variations files are included from. This speeds up the build, but variations can no longer be directly instantiated. --Variations are no longer exported.
2016-02-18 21:58:24 -05:00
class LazyTravisVariation : public ParametricVariation<T>
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
public:
--Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums.
2016-01-04 19:50:15 -05:00
LazyTravisVariation(T weight = 1.0) : ParametricVariation<T>("lazyTravis", eVariationId::VAR_LAZY_TRAVIS, weight)
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
Init();
}
PARVARCOPY(LazyTravisVariation)
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
--User changes -Add post_smartcrop. --Bug fixes -Fix bug in crackle. -Wrong point assignment in hexaplay3D, hexnix3D. -Improper Z assignment in rblur. -Fix inconsistency with original in circlecrop. -Put EMBER_ROOT bakc to ./../../../ in default.pri. This is TBD. --Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums. -Optimize crob, nBlur. -Fix weird assignment statement in falloff3. -Cleanup in VarFuncs::SimplexNoise3D(). -Replace fabs() with std::abs(). -General cleanup.
2016-01-12 23:42:12 -05:00
T x = std::abs(helper.In.x);
T y = std::abs(helper.In.y);
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
T s;
T p;
T x2, y2;
if (x > m_Weight || y > m_Weight)
{
if (x > y)
{
s = x;
if (helper.In.x > 0)
p = s + helper.In.y + s * m_Out4;
else
p = 5 * s - helper.In.y + s * m_Out4;
}
else
{
s = y;
if (helper.In.y > 0)
p = 3 * s - helper.In.x + s * m_Out4;
else
p = 7 * s + helper.In.x + s * m_Out4;
}
p = fmod(p, s * 8);
if (p <= 2 * s)
{
x2 = s + m_Space;
y2 = -(1 * s - p);
y2 = y2 + y2 / s * m_Space;
}
else if (p <= 4 * s)
{
y2 = s + m_Space;
x2 = (3 * s - p);
x2 = x2 + x2 / s * m_Space;
}
else if (p <= 6 * s)
{
x2 = -(s + m_Space);
y2 = (5 * s - p);
y2 = y2 + y2 / s * m_Space;
}
else
{
y2 = -(s + m_Space);
x2 = -(7 * s - p);
x2 = x2 + x2 / s * m_Space;
}
helper.Out.x = m_Weight * x2;
helper.Out.y = m_Weight * y2;
}
else
{
if (x > y)
{
s = x;
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
if (helper.In.x > 0)
p = s + helper.In.y + s * m_In4;
else
p = 5 * s - helper.In.y + s * m_In4;
}
else
{
s = y;
if (helper.In.y > 0)
p = 3 * s - helper.In.x + s * m_In4;
else
p = 7 * s + helper.In.x + s * m_In4;
}
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
p = fmod(p, s * 8);
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
if (p <= 2 * s)
{
helper.Out.x = m_Weight * s;
helper.Out.y = -(m_Weight * (s - p));
}
else if (p <= 4 * s)
{
helper.Out.x = m_Weight * (3 * s - p);
helper.Out.y = m_Weight * s;
}
else if (p <= 6 * s)
{
helper.Out.x = -(m_Weight * s);
helper.Out.y = m_Weight * (5 * s - p);
}
else
{
helper.Out.x = -(m_Weight * (7 * s - p));
helper.Out.y = -(m_Weight * s);
}
}
--User changes -Allow for pausing the renderer in the main window. This makes is more efficient when entering many parameters, such as when following a tutorial. -Add support for new variations: erf, gamma, jac_cn, jac_dn, jac_sn, logDB, pressure_wave, pRose3D, splits3D, w, waves2b, x, xerf, y, z. -Inform user of the start and stop of file parsing in EmberAnimate because the files could potentially be very large. -Move the follwing fields to a new table called Animation: Interpolation, Affine Interpolation, Temporal Samples, Temporal Filter Width, Temporal Filter Type. -These currently have no effect on the interactive renderer and instead are used when running flames through EmberGenome to generate sequences, and then animating them in Fractorium or EmberAnimate. -Add new parameter overrides for EmberRender and EmberAnimate which directly assign values to all flames being rendered, rather than scale: --quality --demin --demax --Bug fixes -Left pad instead of right pad names of sequence outputs from EmberGenome. -Unique file naming was broken for files which already had an underscore in them. -Properly report that png is the default format of EmberRender and EmberAnimate output instead of erroneously claiming it was jpg. -Make command line programs search these folders in this order for the palette file: ./ ~/.fractorium ~/.config/fractorium /usr/share/fractorium /usr/local/share/fractorium -Fix possible bad values in hexes. -Better assignment of Z variables. -Fix boarders due to use of poorly implemented rint() function from flam3. Use std::rint() now. -wedge_sph was completely wrong due to having accidentally swapped the mapping of two parameters. -Make juliascope juliascope_power parameter be of type REAL_NONZERO since it's used as a denominator. -Make Z assignment compatible with the originals in: -arch, bcircle, bCollide, bent, bent2, bisplit, blob, blur_linear, blur_square, bMod, boarders, boarders2, bSwirl, bTransform, butterfly, cardioid, cell, circleblur, circlize, circlize2, circus, collideoscope, cos, cosine, cosh, coth, cpow, cpow2, crescents, cropn, csc, csch, curl, curve, dc_gridout, deltaa, diamond, disc2, eclipse, eCollide, edisc, eJulia, elliptic, eMod, eMotion, ennepers, epispiral, ePush, eRotate, eScale, eSwirl, ex, exp, expo, exponential, fan, fdisc, fibonacci, fibonacci2, fisheye, flipcircle, flipy, flower, flux, funnel, glynnia, GlynnSim1, GlynnSim2, GlynnSim3, gridout, handkerchief, heart, hole, idisc, julia, julian2, juliaNab, kaleidoscope, lazyTravis, Lissajous, mask, MobiusN, mobius_strip, modulus, murl, murl2, npolar, ortho, oscilloscope, parabola, perspective, petal, phoenix_julia, pie (was also inconsistent between cpu and gpu), poincare, popcorn, popcorn2, power, pow_block, rational3, rays, rblur, rings, rippled, roundspher, sec, secant2, sigmoid, sin, sineblur, sinh, sinusgrid, sphericaln, spiralwing, spirograph, split, squarize, squirrel, squish, sschecks, starblur, stripes, stwin, super_shape, tan, tancos, tangent, tanh, TwinTrian, twoface, unpolar, waves, wavesn, wedge_julia, whorl, xheart, zblur, zscale. --Code changes -Generalize Variation::PrecalcHelper() and rename to PrePostPrecalcHelper(). --Do the same for the OpenCL version and rename it PrePostPrecalcOpenCLString(). -Rename Variation::m_AssignType to m_PrePostAssignType since it's only relevant to pre/post variations.
2016-01-29 20:02:15 -05:00
helper.Out.z = DefaultZ(helper);
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
}
--Code changes -Better use of const functions. -Ember.pro should not define BUILDING_EMBERCL, instead define BUILDING_EMBER.
2015-06-29 23:13:53 -04:00
virtual string OpenCLString() const override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
ostringstream ss, ss2;
0.4.1.3 Beta 10/14/2014 --User Changes Size is no longer fixed to the window size. Size scaling is done differently in the final render dialog. This fixes several bugs. Remove Xml saving size from settings and options dialog, it no longer applies. Final render can be broken into strips. Set default save path to the desktop if none is found in the settings file. Set default output size to 1920x1080 if none is found in the settings file. --Bug Fixes Better memory size reporting in final render dialog. --Code Changes Migrate to C++11, Qt 5.3.1, and Visual Studio 2013. Change most instances of unsigned int to size_t, and int to intmax_t. Add m_OrigPixPerUnit and m_ScaleType to Ember for scaling purposes. Replace some sprintf_s() calls in XmlToEmber with ostringstream. Move more non-templated members into RendererBase. Add CopyVec() overload that takes a per element function pointer. Add vector Memset(). Replace '&' with '+' instead of "&amp;" in XmlToEmber for much faster parsing. Break strips rendering out into EmberCommon and call from EmberRender and Fractorium. Make AddAndWriteBuffer() just call WriteBuffer(). Make AddAndWriteImage() delete the existing image first before replacing it. Add SetOutputTexture() to RendererCL to support making new textures in response to resize events. Remove multiple return statements in RendererCL, and replace with a bool that tracks results. Add ToDouble(), MakeEnd(), ToString() and Exists() wrappers in Fractorium. Add Size() wrapper in EmberFile. Make QString function arguments const QString&, and string with const string&. Make ShowCritical() wrapper for invoking a message box from another thread. Add combo box to TwoButtonWidget and rename.
2014-10-14 11:53:15 -04:00
intmax_t i = 0, varIndex = IndexInXform();
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
ss2 << "_" << XformIndexInEmber() << "]";
string index = ss2.str();
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
string weight = WeightDefineString();
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
string spinIn = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
string spinOut = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
string space = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string in4 = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string out4 = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
ss << "\t{\n"
<< "\t\treal_t x = fabs(vIn.x);\n"
<< "\t\treal_t y = fabs(vIn.y);\n"
<< "\t\treal_t s;\n"
<< "\t\treal_t p;\n"
<< "\t\treal_t x2, y2;\n"
<< "\n"
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
<< "\t\tif (x > " << weight << " || y > " << weight << ")\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t\t{\n"
<< "\t\t if (x > y)\n"
<< "\t\t {\n"
<< "\t\t s = x;\n"
<< "\n"
<< "\t\t if (vIn.x > 0)\n"
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
<< "\t\t p = fma(s, " << out4 << ", s + vIn.y);\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t\t else\n"
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
<< "\t\t p = fma((real_t)(5.0), s, fma(s, " << out4 << ", -vIn.y));\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t\t }\n"
<< "\t\t else\n"
<< "\t\t {\n"
<< "\t\t s = y;\n"
<< "\n"
<< "\t\t if (vIn.y > 0)\n"
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
<< "\t\t p = fma((real_t)(3.0), s, fma(s, " << out4 << ", -vIn.x));\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t\t else\n"
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
<< "\t\t p = fma((real_t)(7.0), s, fma(s, " << out4 << ", vIn.x));\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t\t }\n"
<< "\n"
<< "\t\t p = fmod(p, s * 8);\n"
<< "\n"
<< "\t\t if (p <= 2 * s)\n"
<< "\t\t {\n"
<< "\t\t x2 = s + " << space << ";\n"
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
<< "\t\t y2 = -fma((real_t)(1.0), s, -p);\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t\t y2 = y2 + y2 / s * " << space << ";\n"
<< "\t\t }\n"
<< "\t\t else if (p <= 4 * s)\n"
<< "\t\t {\n"
<< "\t\t y2 = s + " << space << ";\n"
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
<< "\t\t x2 = fma((real_t)(3.0), s, -p);\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t\t x2 = x2 + x2 / s * " << space << ";\n"
<< "\t\t }\n"
<< "\t\t else if (p <= 6 * s)\n"
<< "\t\t {\n"
<< "\t\t x2 = -(s + " << space << ");\n"
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
<< "\t\t y2 = fma((real_t)(5.0), s, -p);\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t\t y2 = y2 + y2 / s * " << space << ";\n"
<< "\t\t }\n"
<< "\t\t else\n"
<< "\t\t {\n"
<< "\t\t y2 = -(s + " << space << ");\n"
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
<< "\t\t x2 = -fma((real_t)(7.0), s, -p);\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t\t x2 = x2 + x2 / s * " << space << ";\n"
<< "\t\t }\n"
<< "\n"
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
<< "\t\t vOut.x = " << weight << " * x2;\n"
<< "\t\t vOut.y = " << weight << " * y2;\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t\t}\n"
<< "\t\telse\n"
<< "\t\t{\n"
<< "\t\t if (x > y)\n"
<< "\t\t {\n"
<< "\t\t s = x;\n"
<< "\n"
<< "\t\t if (vIn.x > 0)\n"
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
<< "\t\t p = fma(s, " << in4 << ", s + vIn.y);\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t\t else\n"
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
<< "\t\t p = fma((real_t)(5.0), s, fma(s, " << in4 << ", -vIn.y));\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t\t }\n"
<< "\t\t else\n"
<< "\t\t {\n"
<< "\t\t s = y;\n"
<< "\n"
<< "\t\t if (vIn.y > 0)\n"
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
<< "\t\t p = fma((real_t)(3.0), s, fma(s, " << in4 << ", -vIn.x));\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t\t else\n"
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
<< "\t\t p = fma((real_t)(7.0), s, fma(s, " << in4 << ", vIn.x));\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t\t }\n"
<< "\n"
<< "\t\t p = fmod(p, s * 8);\n"
<< "\n"
<< "\t\t if (p <= 2 * s)\n"
<< "\t\t {\n"
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
<< "\t\t vOut.x = " << weight << " * s;\n"
<< "\t\t vOut.y = -(" << weight << " * (s - p));\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t\t }\n"
<< "\t\t else if (p <= 4 * s)\n"
<< "\t\t {\n"
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
<< "\t\t vOut.x = " << weight << " * fma((real_t)(3.0), s, -p);\n"
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
<< "\t\t vOut.y = " << weight << " * s;\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t\t }\n"
<< "\t\t else if (p <= 6 * s)\n"
<< "\t\t {\n"
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
<< "\t\t vOut.x = -(" << weight << " * s);\n"
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
<< "\t\t vOut.y = " << weight << " * fma((real_t)(5.0), s, -p);\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t\t }\n"
<< "\t\t else\n"
<< "\t\t {\n"
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
<< "\t\t vOut.x = -(" << weight << " * fma((real_t)(7.0), s, -p));\n"
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
<< "\t\t vOut.y = -(" << weight << " * s);\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t\t }\n"
<< "\t\t}\n"
<< "\n"
--User changes -Allow for pausing the renderer in the main window. This makes is more efficient when entering many parameters, such as when following a tutorial. -Add support for new variations: erf, gamma, jac_cn, jac_dn, jac_sn, logDB, pressure_wave, pRose3D, splits3D, w, waves2b, x, xerf, y, z. -Inform user of the start and stop of file parsing in EmberAnimate because the files could potentially be very large. -Move the follwing fields to a new table called Animation: Interpolation, Affine Interpolation, Temporal Samples, Temporal Filter Width, Temporal Filter Type. -These currently have no effect on the interactive renderer and instead are used when running flames through EmberGenome to generate sequences, and then animating them in Fractorium or EmberAnimate. -Add new parameter overrides for EmberRender and EmberAnimate which directly assign values to all flames being rendered, rather than scale: --quality --demin --demax --Bug fixes -Left pad instead of right pad names of sequence outputs from EmberGenome. -Unique file naming was broken for files which already had an underscore in them. -Properly report that png is the default format of EmberRender and EmberAnimate output instead of erroneously claiming it was jpg. -Make command line programs search these folders in this order for the palette file: ./ ~/.fractorium ~/.config/fractorium /usr/share/fractorium /usr/local/share/fractorium -Fix possible bad values in hexes. -Better assignment of Z variables. -Fix boarders due to use of poorly implemented rint() function from flam3. Use std::rint() now. -wedge_sph was completely wrong due to having accidentally swapped the mapping of two parameters. -Make juliascope juliascope_power parameter be of type REAL_NONZERO since it's used as a denominator. -Make Z assignment compatible with the originals in: -arch, bcircle, bCollide, bent, bent2, bisplit, blob, blur_linear, blur_square, bMod, boarders, boarders2, bSwirl, bTransform, butterfly, cardioid, cell, circleblur, circlize, circlize2, circus, collideoscope, cos, cosine, cosh, coth, cpow, cpow2, crescents, cropn, csc, csch, curl, curve, dc_gridout, deltaa, diamond, disc2, eclipse, eCollide, edisc, eJulia, elliptic, eMod, eMotion, ennepers, epispiral, ePush, eRotate, eScale, eSwirl, ex, exp, expo, exponential, fan, fdisc, fibonacci, fibonacci2, fisheye, flipcircle, flipy, flower, flux, funnel, glynnia, GlynnSim1, GlynnSim2, GlynnSim3, gridout, handkerchief, heart, hole, idisc, julia, julian2, juliaNab, kaleidoscope, lazyTravis, Lissajous, mask, MobiusN, mobius_strip, modulus, murl, murl2, npolar, ortho, oscilloscope, parabola, perspective, petal, phoenix_julia, pie (was also inconsistent between cpu and gpu), poincare, popcorn, popcorn2, power, pow_block, rational3, rays, rblur, rings, rippled, roundspher, sec, secant2, sigmoid, sin, sineblur, sinh, sinusgrid, sphericaln, spiralwing, spirograph, split, squarize, squirrel, squish, sschecks, starblur, stripes, stwin, super_shape, tan, tancos, tangent, tanh, TwinTrian, twoface, unpolar, waves, wavesn, wedge_julia, whorl, xheart, zblur, zscale. --Code changes -Generalize Variation::PrecalcHelper() and rename to PrePostPrecalcHelper(). --Do the same for the OpenCL version and rename it PrePostPrecalcOpenCLString(). -Rename Variation::m_AssignType to m_PrePostAssignType since it's only relevant to pre/post variations.
2016-01-29 20:02:15 -05:00
<< "\t\tvOut.z = " << DefaultZCl()
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t}\n";
return ss.str();
}
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
virtual void Precalc() override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
m_In4 = 4 * m_SpinIn;
m_Out4 = 4 * m_SpinOut;
}
protected:
void Init()
{
string prefix = Prefix();
m_Params.clear();
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
m_Params.push_back(ParamWithName<T>(&m_SpinIn, prefix + "lazyTravis_spin_in", 1, eParamType::REAL_CYCLIC, 0, 2));
--Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums.
2016-01-04 19:50:15 -05:00
m_Params.push_back(ParamWithName<T>(&m_SpinOut, prefix + "lazyTravis_spin_out", 0, eParamType::REAL_CYCLIC, 0, 2));
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
m_Params.push_back(ParamWithName<T>(&m_Space, prefix + "lazyTravis_space"));
m_Params.push_back(ParamWithName<T>(true, &m_In4, prefix + "lazyTravis_in4"));//Precalc.
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
m_Params.push_back(ParamWithName<T>(true, &m_Out4, prefix + "lazyTravis_out4"));
}
private:
T m_SpinIn;
T m_SpinOut;
T m_Space;
T m_In4;//Precalc.
T m_Out4;
};
/// <summary>
/// squish.
/// </summary>
template <typename T>
--Code changes -Split VariationList into .h/cpp files. --Make VariationList.cpp the only place where the variations files are included from. This speeds up the build, but variations can no longer be directly instantiated. --Variations are no longer exported.
2016-02-18 21:58:24 -05:00
class SquishVariation : public ParametricVariation<T>
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
public:
--Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums.
2016-01-04 19:50:15 -05:00
SquishVariation(T weight = 1.0) : ParametricVariation<T>("squish", eVariationId::VAR_SQUISH, weight)
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
Init();
}
PARVARCOPY(SquishVariation)
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
--User changes -Add post_smartcrop. --Bug fixes -Fix bug in crackle. -Wrong point assignment in hexaplay3D, hexnix3D. -Improper Z assignment in rblur. -Fix inconsistency with original in circlecrop. -Put EMBER_ROOT bakc to ./../../../ in default.pri. This is TBD. --Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums. -Optimize crob, nBlur. -Fix weird assignment statement in falloff3. -Cleanup in VarFuncs::SimplexNoise3D(). -Replace fabs() with std::abs(). -General cleanup.
2016-01-12 23:42:12 -05:00
T x = std::abs(helper.In.x);
T y = std::abs(helper.In.y);
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
T s;
T p;
if (x > y)
{
s = x;
if (helper.In.x > 0)
p = helper.In.y;
else
p = 4 * s - helper.In.y;
}
else
{
s = y;
if (helper.In.y > 0)
p = 2 * s - helper.In.x;
else
p = 6 * s + helper.In.x;
}
p = m_InvPower * (p + 8 * s * Floor<T>(m_Power * rand.Frand01<T>()));
if (p <= s)
{
helper.Out.x = m_Weight * s;
helper.Out.y = m_Weight * p;
}
else if (p <= 3 * s)
{
helper.Out.x = m_Weight * (2 * s - p);
helper.Out.y = m_Weight * s;
}
else if (p <= 5 * s)
{
helper.Out.x = -(m_Weight * s);
helper.Out.y = m_Weight * (4 * s - p);
}
else if (p <= 7 * s)
{
helper.Out.x = -(m_Weight * (6 * s - p));
helper.Out.y = -(m_Weight * s);
}
else
{
helper.Out.x = m_Weight * s;
helper.Out.y = (m_Weight * (8 * s - p));
}
--User changes -Allow for pausing the renderer in the main window. This makes is more efficient when entering many parameters, such as when following a tutorial. -Add support for new variations: erf, gamma, jac_cn, jac_dn, jac_sn, logDB, pressure_wave, pRose3D, splits3D, w, waves2b, x, xerf, y, z. -Inform user of the start and stop of file parsing in EmberAnimate because the files could potentially be very large. -Move the follwing fields to a new table called Animation: Interpolation, Affine Interpolation, Temporal Samples, Temporal Filter Width, Temporal Filter Type. -These currently have no effect on the interactive renderer and instead are used when running flames through EmberGenome to generate sequences, and then animating them in Fractorium or EmberAnimate. -Add new parameter overrides for EmberRender and EmberAnimate which directly assign values to all flames being rendered, rather than scale: --quality --demin --demax --Bug fixes -Left pad instead of right pad names of sequence outputs from EmberGenome. -Unique file naming was broken for files which already had an underscore in them. -Properly report that png is the default format of EmberRender and EmberAnimate output instead of erroneously claiming it was jpg. -Make command line programs search these folders in this order for the palette file: ./ ~/.fractorium ~/.config/fractorium /usr/share/fractorium /usr/local/share/fractorium -Fix possible bad values in hexes. -Better assignment of Z variables. -Fix boarders due to use of poorly implemented rint() function from flam3. Use std::rint() now. -wedge_sph was completely wrong due to having accidentally swapped the mapping of two parameters. -Make juliascope juliascope_power parameter be of type REAL_NONZERO since it's used as a denominator. -Make Z assignment compatible with the originals in: -arch, bcircle, bCollide, bent, bent2, bisplit, blob, blur_linear, blur_square, bMod, boarders, boarders2, bSwirl, bTransform, butterfly, cardioid, cell, circleblur, circlize, circlize2, circus, collideoscope, cos, cosine, cosh, coth, cpow, cpow2, crescents, cropn, csc, csch, curl, curve, dc_gridout, deltaa, diamond, disc2, eclipse, eCollide, edisc, eJulia, elliptic, eMod, eMotion, ennepers, epispiral, ePush, eRotate, eScale, eSwirl, ex, exp, expo, exponential, fan, fdisc, fibonacci, fibonacci2, fisheye, flipcircle, flipy, flower, flux, funnel, glynnia, GlynnSim1, GlynnSim2, GlynnSim3, gridout, handkerchief, heart, hole, idisc, julia, julian2, juliaNab, kaleidoscope, lazyTravis, Lissajous, mask, MobiusN, mobius_strip, modulus, murl, murl2, npolar, ortho, oscilloscope, parabola, perspective, petal, phoenix_julia, pie (was also inconsistent between cpu and gpu), poincare, popcorn, popcorn2, power, pow_block, rational3, rays, rblur, rings, rippled, roundspher, sec, secant2, sigmoid, sin, sineblur, sinh, sinusgrid, sphericaln, spiralwing, spirograph, split, squarize, squirrel, squish, sschecks, starblur, stripes, stwin, super_shape, tan, tancos, tangent, tanh, TwinTrian, twoface, unpolar, waves, wavesn, wedge_julia, whorl, xheart, zblur, zscale. --Code changes -Generalize Variation::PrecalcHelper() and rename to PrePostPrecalcHelper(). --Do the same for the OpenCL version and rename it PrePostPrecalcOpenCLString(). -Rename Variation::m_AssignType to m_PrePostAssignType since it's only relevant to pre/post variations.
2016-01-29 20:02:15 -05:00
helper.Out.z = DefaultZ(helper);
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
}
--Code changes -Better use of const functions. -Ember.pro should not define BUILDING_EMBERCL, instead define BUILDING_EMBER.
2015-06-29 23:13:53 -04:00
virtual string OpenCLString() const override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
ostringstream ss, ss2;
0.4.1.3 Beta 10/14/2014 --User Changes Size is no longer fixed to the window size. Size scaling is done differently in the final render dialog. This fixes several bugs. Remove Xml saving size from settings and options dialog, it no longer applies. Final render can be broken into strips. Set default save path to the desktop if none is found in the settings file. Set default output size to 1920x1080 if none is found in the settings file. --Bug Fixes Better memory size reporting in final render dialog. --Code Changes Migrate to C++11, Qt 5.3.1, and Visual Studio 2013. Change most instances of unsigned int to size_t, and int to intmax_t. Add m_OrigPixPerUnit and m_ScaleType to Ember for scaling purposes. Replace some sprintf_s() calls in XmlToEmber with ostringstream. Move more non-templated members into RendererBase. Add CopyVec() overload that takes a per element function pointer. Add vector Memset(). Replace '&' with '+' instead of "&amp;" in XmlToEmber for much faster parsing. Break strips rendering out into EmberCommon and call from EmberRender and Fractorium. Make AddAndWriteBuffer() just call WriteBuffer(). Make AddAndWriteImage() delete the existing image first before replacing it. Add SetOutputTexture() to RendererCL to support making new textures in response to resize events. Remove multiple return statements in RendererCL, and replace with a bool that tracks results. Add ToDouble(), MakeEnd(), ToString() and Exists() wrappers in Fractorium. Add Size() wrapper in EmberFile. Make QString function arguments const QString&, and string with const string&. Make ShowCritical() wrapper for invoking a message box from another thread. Add combo box to TwoButtonWidget and rename.
2014-10-14 11:53:15 -04:00
intmax_t i = 0, varIndex = IndexInXform();
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
ss2 << "_" << XformIndexInEmber() << "]";
string index = ss2.str();
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
string weight = WeightDefineString();
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
string power = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
string invPower = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
ss << "\t{\n"
<< "\t\treal_t x = fabs(vIn.x);\n"
<< "\t\treal_t y = fabs(vIn.y);\n"
<< "\t\treal_t s;\n"
<< "\t\treal_t p;\n"
<< "\n"
<< "\t\tif (x > y)\n"
<< "\t\t{\n"
<< "\t\t s = x;\n"
<< "\n"
<< "\t\t if (vIn.x > 0)\n"
<< "\t\t p = vIn.y;\n"
<< "\t\t else\n"
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
<< "\t\t p = fma((real_t)(4.0), s, -vIn.y);\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t\t}\n"
<< "\t\telse\n"
<< "\t\t{\n"
<< "\t\t s = y;\n"
<< "\n"
<< "\t\t if (vIn.y > 0)\n"
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
<< "\t\t p = fma((real_t)(2.0), s, -vIn.x);\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t\t else\n"
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
<< "\t\t p = fma((real_t)(6.0), s, vIn.x);\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t\t}\n"
<< "\n"
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
<< "\t\tp = " << invPower << " * fma((real_t)(8.0), s * floor(" << power << " * MwcNext01(mwc)), p);\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\n"
<< "\t\tif (p <= s)\n"
<< "\t\t{\n"
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
<< "\t\t vOut.x = " << weight << " * s;\n"
<< "\t\t vOut.y = " << weight << " * p;\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t\t}\n"
<< "\t\telse if (p <= 3 * s)\n"
<< "\t\t{\n"
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
<< "\t\t vOut.x = " << weight << " * fma((real_t)(2.0), s, -p);\n"
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
<< "\t\t vOut.y = " << weight << " * s;\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t\t}\n"
<< "\t\telse if (p <= 5 * s)\n"
<< "\t\t{\n"
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
<< "\t\t vOut.x = -(" << weight << " * s);\n"
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
<< "\t\t vOut.y = " << weight << " * fma((real_t)(4.0), s, -p);\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t\t}\n"
<< "\t\telse if (p <= 7 * s)\n"
<< "\t\t{\n"
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
<< "\t\t vOut.x = -(" << weight << " * fma((real_t)(6.0), s, -p));\n"
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
<< "\t\t vOut.y = -(" << weight << " * s);\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t\t}\n"
<< "\t\telse\n"
<< "\t\t{\n"
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
<< "\t\t vOut.x = " << weight << " * s;\n"
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
<< "\t\t vOut.y = -(" << weight << " * fma((real_t)(8.0), s, -p));\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t\t}\n"
<< "\n"
--User changes -Allow for pausing the renderer in the main window. This makes is more efficient when entering many parameters, such as when following a tutorial. -Add support for new variations: erf, gamma, jac_cn, jac_dn, jac_sn, logDB, pressure_wave, pRose3D, splits3D, w, waves2b, x, xerf, y, z. -Inform user of the start and stop of file parsing in EmberAnimate because the files could potentially be very large. -Move the follwing fields to a new table called Animation: Interpolation, Affine Interpolation, Temporal Samples, Temporal Filter Width, Temporal Filter Type. -These currently have no effect on the interactive renderer and instead are used when running flames through EmberGenome to generate sequences, and then animating them in Fractorium or EmberAnimate. -Add new parameter overrides for EmberRender and EmberAnimate which directly assign values to all flames being rendered, rather than scale: --quality --demin --demax --Bug fixes -Left pad instead of right pad names of sequence outputs from EmberGenome. -Unique file naming was broken for files which already had an underscore in them. -Properly report that png is the default format of EmberRender and EmberAnimate output instead of erroneously claiming it was jpg. -Make command line programs search these folders in this order for the palette file: ./ ~/.fractorium ~/.config/fractorium /usr/share/fractorium /usr/local/share/fractorium -Fix possible bad values in hexes. -Better assignment of Z variables. -Fix boarders due to use of poorly implemented rint() function from flam3. Use std::rint() now. -wedge_sph was completely wrong due to having accidentally swapped the mapping of two parameters. -Make juliascope juliascope_power parameter be of type REAL_NONZERO since it's used as a denominator. -Make Z assignment compatible with the originals in: -arch, bcircle, bCollide, bent, bent2, bisplit, blob, blur_linear, blur_square, bMod, boarders, boarders2, bSwirl, bTransform, butterfly, cardioid, cell, circleblur, circlize, circlize2, circus, collideoscope, cos, cosine, cosh, coth, cpow, cpow2, crescents, cropn, csc, csch, curl, curve, dc_gridout, deltaa, diamond, disc2, eclipse, eCollide, edisc, eJulia, elliptic, eMod, eMotion, ennepers, epispiral, ePush, eRotate, eScale, eSwirl, ex, exp, expo, exponential, fan, fdisc, fibonacci, fibonacci2, fisheye, flipcircle, flipy, flower, flux, funnel, glynnia, GlynnSim1, GlynnSim2, GlynnSim3, gridout, handkerchief, heart, hole, idisc, julia, julian2, juliaNab, kaleidoscope, lazyTravis, Lissajous, mask, MobiusN, mobius_strip, modulus, murl, murl2, npolar, ortho, oscilloscope, parabola, perspective, petal, phoenix_julia, pie (was also inconsistent between cpu and gpu), poincare, popcorn, popcorn2, power, pow_block, rational3, rays, rblur, rings, rippled, roundspher, sec, secant2, sigmoid, sin, sineblur, sinh, sinusgrid, sphericaln, spiralwing, spirograph, split, squarize, squirrel, squish, sschecks, starblur, stripes, stwin, super_shape, tan, tancos, tangent, tanh, TwinTrian, twoface, unpolar, waves, wavesn, wedge_julia, whorl, xheart, zblur, zscale. --Code changes -Generalize Variation::PrecalcHelper() and rename to PrePostPrecalcHelper(). --Do the same for the OpenCL version and rename it PrePostPrecalcOpenCLString(). -Rename Variation::m_AssignType to m_PrePostAssignType since it's only relevant to pre/post variations.
2016-01-29 20:02:15 -05:00
<< "\t\tvOut.z = " << DefaultZCl()
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t}\n";
return ss.str();
}
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
virtual void Precalc() override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
m_InvPower = 1 / m_Power;
}
protected:
void Init()
{
string prefix = Prefix();
m_Params.clear();
--Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums.
2016-01-04 19:50:15 -05:00
m_Params.push_back(ParamWithName<T>(&m_Power, prefix + "squish_power", 2, eParamType::INTEGER, 2, T(INT_MAX)));
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
m_Params.push_back(ParamWithName<T>(true, &m_InvPower, prefix + "squish_inv_power"));//Precalc.
}
private:
T m_Power;
T m_InvPower;//Precalc.
};
/// <summary>
/// circus.
/// </summary>
template <typename T>
--Code changes -Split VariationList into .h/cpp files. --Make VariationList.cpp the only place where the variations files are included from. This speeds up the build, but variations can no longer be directly instantiated. --Variations are no longer exported.
2016-02-18 21:58:24 -05:00
class CircusVariation : public ParametricVariation<T>
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
public:
--Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums.
2016-01-04 19:50:15 -05:00
CircusVariation(T weight = 1.0) : ParametricVariation<T>("circus", eVariationId::VAR_CIRCUS, weight, true, true, true)
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
Init();
}
PARVARCOPY(CircusVariation)
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
T r = helper.m_PrecalcSqrtSumSquares;
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
if (r <= 1)
r *= m_Scale;
else
r *= m_InvScale;
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
helper.Out.x = m_Weight * r * helper.m_PrecalcCosa;
helper.Out.y = m_Weight * r * helper.m_PrecalcSina;
--User changes -Allow for pausing the renderer in the main window. This makes is more efficient when entering many parameters, such as when following a tutorial. -Add support for new variations: erf, gamma, jac_cn, jac_dn, jac_sn, logDB, pressure_wave, pRose3D, splits3D, w, waves2b, x, xerf, y, z. -Inform user of the start and stop of file parsing in EmberAnimate because the files could potentially be very large. -Move the follwing fields to a new table called Animation: Interpolation, Affine Interpolation, Temporal Samples, Temporal Filter Width, Temporal Filter Type. -These currently have no effect on the interactive renderer and instead are used when running flames through EmberGenome to generate sequences, and then animating them in Fractorium or EmberAnimate. -Add new parameter overrides for EmberRender and EmberAnimate which directly assign values to all flames being rendered, rather than scale: --quality --demin --demax --Bug fixes -Left pad instead of right pad names of sequence outputs from EmberGenome. -Unique file naming was broken for files which already had an underscore in them. -Properly report that png is the default format of EmberRender and EmberAnimate output instead of erroneously claiming it was jpg. -Make command line programs search these folders in this order for the palette file: ./ ~/.fractorium ~/.config/fractorium /usr/share/fractorium /usr/local/share/fractorium -Fix possible bad values in hexes. -Better assignment of Z variables. -Fix boarders due to use of poorly implemented rint() function from flam3. Use std::rint() now. -wedge_sph was completely wrong due to having accidentally swapped the mapping of two parameters. -Make juliascope juliascope_power parameter be of type REAL_NONZERO since it's used as a denominator. -Make Z assignment compatible with the originals in: -arch, bcircle, bCollide, bent, bent2, bisplit, blob, blur_linear, blur_square, bMod, boarders, boarders2, bSwirl, bTransform, butterfly, cardioid, cell, circleblur, circlize, circlize2, circus, collideoscope, cos, cosine, cosh, coth, cpow, cpow2, crescents, cropn, csc, csch, curl, curve, dc_gridout, deltaa, diamond, disc2, eclipse, eCollide, edisc, eJulia, elliptic, eMod, eMotion, ennepers, epispiral, ePush, eRotate, eScale, eSwirl, ex, exp, expo, exponential, fan, fdisc, fibonacci, fibonacci2, fisheye, flipcircle, flipy, flower, flux, funnel, glynnia, GlynnSim1, GlynnSim2, GlynnSim3, gridout, handkerchief, heart, hole, idisc, julia, julian2, juliaNab, kaleidoscope, lazyTravis, Lissajous, mask, MobiusN, mobius_strip, modulus, murl, murl2, npolar, ortho, oscilloscope, parabola, perspective, petal, phoenix_julia, pie (was also inconsistent between cpu and gpu), poincare, popcorn, popcorn2, power, pow_block, rational3, rays, rblur, rings, rippled, roundspher, sec, secant2, sigmoid, sin, sineblur, sinh, sinusgrid, sphericaln, spiralwing, spirograph, split, squarize, squirrel, squish, sschecks, starblur, stripes, stwin, super_shape, tan, tancos, tangent, tanh, TwinTrian, twoface, unpolar, waves, wavesn, wedge_julia, whorl, xheart, zblur, zscale. --Code changes -Generalize Variation::PrecalcHelper() and rename to PrePostPrecalcHelper(). --Do the same for the OpenCL version and rename it PrePostPrecalcOpenCLString(). -Rename Variation::m_AssignType to m_PrePostAssignType since it's only relevant to pre/post variations.
2016-01-29 20:02:15 -05:00
helper.Out.z = DefaultZ(helper);
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
}
--Code changes -Better use of const functions. -Ember.pro should not define BUILDING_EMBERCL, instead define BUILDING_EMBER.
2015-06-29 23:13:53 -04:00
virtual string OpenCLString() const override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
ostringstream ss, ss2;
0.4.1.3 Beta 10/14/2014 --User Changes Size is no longer fixed to the window size. Size scaling is done differently in the final render dialog. This fixes several bugs. Remove Xml saving size from settings and options dialog, it no longer applies. Final render can be broken into strips. Set default save path to the desktop if none is found in the settings file. Set default output size to 1920x1080 if none is found in the settings file. --Bug Fixes Better memory size reporting in final render dialog. --Code Changes Migrate to C++11, Qt 5.3.1, and Visual Studio 2013. Change most instances of unsigned int to size_t, and int to intmax_t. Add m_OrigPixPerUnit and m_ScaleType to Ember for scaling purposes. Replace some sprintf_s() calls in XmlToEmber with ostringstream. Move more non-templated members into RendererBase. Add CopyVec() overload that takes a per element function pointer. Add vector Memset(). Replace '&' with '+' instead of "&amp;" in XmlToEmber for much faster parsing. Break strips rendering out into EmberCommon and call from EmberRender and Fractorium. Make AddAndWriteBuffer() just call WriteBuffer(). Make AddAndWriteImage() delete the existing image first before replacing it. Add SetOutputTexture() to RendererCL to support making new textures in response to resize events. Remove multiple return statements in RendererCL, and replace with a bool that tracks results. Add ToDouble(), MakeEnd(), ToString() and Exists() wrappers in Fractorium. Add Size() wrapper in EmberFile. Make QString function arguments const QString&, and string with const string&. Make ShowCritical() wrapper for invoking a message box from another thread. Add combo box to TwoButtonWidget and rename.
2014-10-14 11:53:15 -04:00
intmax_t i = 0, varIndex = IndexInXform();
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
ss2 << "_" << XformIndexInEmber() << "]";
string index = ss2.str();
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
string weight = WeightDefineString();
string scale = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
string invScale = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
ss << "\t{\n"
<< "\t\treal_t r = precalcSqrtSumSquares;\n"
<< "\n"
<< "\t\tif (r <= 1)\n"
<< "\t\t r *= " << scale << ";\n"
<< "\t\telse\n"
<< "\t\t r *= " << invScale << ";\n"
<< "\n"
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
<< "\t\tvOut.x = " << weight << " * r * precalcCosa;\n"
<< "\t\tvOut.y = " << weight << " * r * precalcSina;\n"
--User changes -Allow for pausing the renderer in the main window. This makes is more efficient when entering many parameters, such as when following a tutorial. -Add support for new variations: erf, gamma, jac_cn, jac_dn, jac_sn, logDB, pressure_wave, pRose3D, splits3D, w, waves2b, x, xerf, y, z. -Inform user of the start and stop of file parsing in EmberAnimate because the files could potentially be very large. -Move the follwing fields to a new table called Animation: Interpolation, Affine Interpolation, Temporal Samples, Temporal Filter Width, Temporal Filter Type. -These currently have no effect on the interactive renderer and instead are used when running flames through EmberGenome to generate sequences, and then animating them in Fractorium or EmberAnimate. -Add new parameter overrides for EmberRender and EmberAnimate which directly assign values to all flames being rendered, rather than scale: --quality --demin --demax --Bug fixes -Left pad instead of right pad names of sequence outputs from EmberGenome. -Unique file naming was broken for files which already had an underscore in them. -Properly report that png is the default format of EmberRender and EmberAnimate output instead of erroneously claiming it was jpg. -Make command line programs search these folders in this order for the palette file: ./ ~/.fractorium ~/.config/fractorium /usr/share/fractorium /usr/local/share/fractorium -Fix possible bad values in hexes. -Better assignment of Z variables. -Fix boarders due to use of poorly implemented rint() function from flam3. Use std::rint() now. -wedge_sph was completely wrong due to having accidentally swapped the mapping of two parameters. -Make juliascope juliascope_power parameter be of type REAL_NONZERO since it's used as a denominator. -Make Z assignment compatible with the originals in: -arch, bcircle, bCollide, bent, bent2, bisplit, blob, blur_linear, blur_square, bMod, boarders, boarders2, bSwirl, bTransform, butterfly, cardioid, cell, circleblur, circlize, circlize2, circus, collideoscope, cos, cosine, cosh, coth, cpow, cpow2, crescents, cropn, csc, csch, curl, curve, dc_gridout, deltaa, diamond, disc2, eclipse, eCollide, edisc, eJulia, elliptic, eMod, eMotion, ennepers, epispiral, ePush, eRotate, eScale, eSwirl, ex, exp, expo, exponential, fan, fdisc, fibonacci, fibonacci2, fisheye, flipcircle, flipy, flower, flux, funnel, glynnia, GlynnSim1, GlynnSim2, GlynnSim3, gridout, handkerchief, heart, hole, idisc, julia, julian2, juliaNab, kaleidoscope, lazyTravis, Lissajous, mask, MobiusN, mobius_strip, modulus, murl, murl2, npolar, ortho, oscilloscope, parabola, perspective, petal, phoenix_julia, pie (was also inconsistent between cpu and gpu), poincare, popcorn, popcorn2, power, pow_block, rational3, rays, rblur, rings, rippled, roundspher, sec, secant2, sigmoid, sin, sineblur, sinh, sinusgrid, sphericaln, spiralwing, spirograph, split, squarize, squirrel, squish, sschecks, starblur, stripes, stwin, super_shape, tan, tancos, tangent, tanh, TwinTrian, twoface, unpolar, waves, wavesn, wedge_julia, whorl, xheart, zblur, zscale. --Code changes -Generalize Variation::PrecalcHelper() and rename to PrePostPrecalcHelper(). --Do the same for the OpenCL version and rename it PrePostPrecalcOpenCLString(). -Rename Variation::m_AssignType to m_PrePostAssignType since it's only relevant to pre/post variations.
2016-01-29 20:02:15 -05:00
<< "\t\tvOut.z = " << DefaultZCl()
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t}\n";
return ss.str();
}
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
virtual void Precalc() override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
m_InvScale = 1 / m_Scale;
}
protected:
void Init()
{
string prefix = Prefix();
m_Params.clear();
m_Params.push_back(ParamWithName<T>(&m_Scale, prefix + "circus_scale", 1));
m_Params.push_back(ParamWithName<T>(true, &m_InvScale, prefix + "circus_inv_power"));//Precalc.
}
private:
T m_Scale;
T m_InvScale;//Precalc.
};
/// <summary>
/// tancos.
/// </summary>
template <typename T>
--Code changes -Split VariationList into .h/cpp files. --Make VariationList.cpp the only place where the variations files are included from. This speeds up the build, but variations can no longer be directly instantiated. --Variations are no longer exported.
2016-02-18 21:58:24 -05:00
class TancosVariation : public Variation<T>
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
public:
--Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums.
2016-01-04 19:50:15 -05:00
TancosVariation(T weight = 1.0) : Variation<T>("tancos", eVariationId::VAR_TANCOS, weight, true) { }
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
VARCOPY(TancosVariation)
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
Numerous fixes 0.4.0.5 Beta 07/18/2014 --User Changes Allow for vibrancy values > 1. Add flatten and unflatten menu items. Automatically flatten like Apophysis does. Add plugin and new_linear tags to Xml to be compatible with Apophysis. --Bug Fixes Fix blur, blur3d, bubble, cropn, cross, curl, curl3d, epispiral, ho, julia3d, julia3dz, loonie, mirror_x, mirror_y, mirror_z, rotate_x, sinusoidal, spherical, spherical3d, stripes. Unique filename on final render was completely broken. Two severe OpenCL bugs. Random seeds were biased and fusing was being reset too often leading to results that differ from the CPU. Subtle, but sometimes severe bug in the setup of the xaos weights. Use properly defined epsilon by getting the value from std::numeric_limits, rather than hard coding 1e-6 or 1e-10. Omit incorrect usage of epsilon everywhere. It should not be automatically added to denominators. Rather, it should only be used if the denominator is zero. Force final render progress bars to 100 on completion. Sometimes they didn't seem to make it there. Make variation name and params comparisons be case insensitive. --Code Changes Make ForEach and FindIf wrappers around std::for_each and std::find_if.
2014-07-19 02:33:18 -04:00
T d = Zeps(helper.m_PrecalcSumSquares);
06/09/2017 --User changes -dark.qss is now per-OS. -Properly set/reload palette when coming from the palette editor. The latter must be done if they've modified the current palette even if they've clicked cancel. --Bug fixes -Make the following variations safer by using Zeps(): conic, bipolar, edisc, whorl, tan, csc, cot, tanh, sech, csch, coth, auger, bwraps, hypertile3d, hypertile3d1, ortho, poincare, rational3, barycentroid, sschecks, cscq, cschq, scry_3D, splitbrdr, hexcrop, nblur, crob. -Fix bug enabling/disabling overwrite button in palette editor. -Small optimization for gdoffs, use precalcAtanYX. -Properly propagate z through circlesplit, cylinder2 and tile_log variations. -Some values in truchet_fill could've been NaN. --Code changes -Make most installation files read only. -Qualify many calls with std:: to ensure they're not colliding with glm:: -Use auto in more places.
2017-06-09 22:38:06 -04:00
helper.Out.x = (m_Weight / d) * (std::tanh(d) * (2 * helper.In.x));
helper.Out.y = (m_Weight / d) * (std::cos(d) * (2 * helper.In.y));
--User changes -Allow for pausing the renderer in the main window. This makes is more efficient when entering many parameters, such as when following a tutorial. -Add support for new variations: erf, gamma, jac_cn, jac_dn, jac_sn, logDB, pressure_wave, pRose3D, splits3D, w, waves2b, x, xerf, y, z. -Inform user of the start and stop of file parsing in EmberAnimate because the files could potentially be very large. -Move the follwing fields to a new table called Animation: Interpolation, Affine Interpolation, Temporal Samples, Temporal Filter Width, Temporal Filter Type. -These currently have no effect on the interactive renderer and instead are used when running flames through EmberGenome to generate sequences, and then animating them in Fractorium or EmberAnimate. -Add new parameter overrides for EmberRender and EmberAnimate which directly assign values to all flames being rendered, rather than scale: --quality --demin --demax --Bug fixes -Left pad instead of right pad names of sequence outputs from EmberGenome. -Unique file naming was broken for files which already had an underscore in them. -Properly report that png is the default format of EmberRender and EmberAnimate output instead of erroneously claiming it was jpg. -Make command line programs search these folders in this order for the palette file: ./ ~/.fractorium ~/.config/fractorium /usr/share/fractorium /usr/local/share/fractorium -Fix possible bad values in hexes. -Better assignment of Z variables. -Fix boarders due to use of poorly implemented rint() function from flam3. Use std::rint() now. -wedge_sph was completely wrong due to having accidentally swapped the mapping of two parameters. -Make juliascope juliascope_power parameter be of type REAL_NONZERO since it's used as a denominator. -Make Z assignment compatible with the originals in: -arch, bcircle, bCollide, bent, bent2, bisplit, blob, blur_linear, blur_square, bMod, boarders, boarders2, bSwirl, bTransform, butterfly, cardioid, cell, circleblur, circlize, circlize2, circus, collideoscope, cos, cosine, cosh, coth, cpow, cpow2, crescents, cropn, csc, csch, curl, curve, dc_gridout, deltaa, diamond, disc2, eclipse, eCollide, edisc, eJulia, elliptic, eMod, eMotion, ennepers, epispiral, ePush, eRotate, eScale, eSwirl, ex, exp, expo, exponential, fan, fdisc, fibonacci, fibonacci2, fisheye, flipcircle, flipy, flower, flux, funnel, glynnia, GlynnSim1, GlynnSim2, GlynnSim3, gridout, handkerchief, heart, hole, idisc, julia, julian2, juliaNab, kaleidoscope, lazyTravis, Lissajous, mask, MobiusN, mobius_strip, modulus, murl, murl2, npolar, ortho, oscilloscope, parabola, perspective, petal, phoenix_julia, pie (was also inconsistent between cpu and gpu), poincare, popcorn, popcorn2, power, pow_block, rational3, rays, rblur, rings, rippled, roundspher, sec, secant2, sigmoid, sin, sineblur, sinh, sinusgrid, sphericaln, spiralwing, spirograph, split, squarize, squirrel, squish, sschecks, starblur, stripes, stwin, super_shape, tan, tancos, tangent, tanh, TwinTrian, twoface, unpolar, waves, wavesn, wedge_julia, whorl, xheart, zblur, zscale. --Code changes -Generalize Variation::PrecalcHelper() and rename to PrePostPrecalcHelper(). --Do the same for the OpenCL version and rename it PrePostPrecalcOpenCLString(). -Rename Variation::m_AssignType to m_PrePostAssignType since it's only relevant to pre/post variations.
2016-01-29 20:02:15 -05:00
helper.Out.z = DefaultZ(helper);
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
}
--Code changes -Better use of const functions. -Ember.pro should not define BUILDING_EMBERCL, instead define BUILDING_EMBER.
2015-06-29 23:13:53 -04:00
virtual string OpenCLString() const override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
ostringstream ss;
0.4.1.3 Beta 10/14/2014 --User Changes Size is no longer fixed to the window size. Size scaling is done differently in the final render dialog. This fixes several bugs. Remove Xml saving size from settings and options dialog, it no longer applies. Final render can be broken into strips. Set default save path to the desktop if none is found in the settings file. Set default output size to 1920x1080 if none is found in the settings file. --Bug Fixes Better memory size reporting in final render dialog. --Code Changes Migrate to C++11, Qt 5.3.1, and Visual Studio 2013. Change most instances of unsigned int to size_t, and int to intmax_t. Add m_OrigPixPerUnit and m_ScaleType to Ember for scaling purposes. Replace some sprintf_s() calls in XmlToEmber with ostringstream. Move more non-templated members into RendererBase. Add CopyVec() overload that takes a per element function pointer. Add vector Memset(). Replace '&' with '+' instead of "&amp;" in XmlToEmber for much faster parsing. Break strips rendering out into EmberCommon and call from EmberRender and Fractorium. Make AddAndWriteBuffer() just call WriteBuffer(). Make AddAndWriteImage() delete the existing image first before replacing it. Add SetOutputTexture() to RendererCL to support making new textures in response to resize events. Remove multiple return statements in RendererCL, and replace with a bool that tracks results. Add ToDouble(), MakeEnd(), ToString() and Exists() wrappers in Fractorium. Add Size() wrapper in EmberFile. Make QString function arguments const QString&, and string with const string&. Make ShowCritical() wrapper for invoking a message box from another thread. Add combo box to TwoButtonWidget and rename.
2014-10-14 11:53:15 -04:00
intmax_t varIndex = IndexInXform();
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
string weight = WeightDefineString();
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
ss << "\t{\n"
Numerous fixes 0.4.0.5 Beta 07/18/2014 --User Changes Allow for vibrancy values > 1. Add flatten and unflatten menu items. Automatically flatten like Apophysis does. Add plugin and new_linear tags to Xml to be compatible with Apophysis. --Bug Fixes Fix blur, blur3d, bubble, cropn, cross, curl, curl3d, epispiral, ho, julia3d, julia3dz, loonie, mirror_x, mirror_y, mirror_z, rotate_x, sinusoidal, spherical, spherical3d, stripes. Unique filename on final render was completely broken. Two severe OpenCL bugs. Random seeds were biased and fusing was being reset too often leading to results that differ from the CPU. Subtle, but sometimes severe bug in the setup of the xaos weights. Use properly defined epsilon by getting the value from std::numeric_limits, rather than hard coding 1e-6 or 1e-10. Omit incorrect usage of epsilon everywhere. It should not be automatically added to denominators. Rather, it should only be used if the denominator is zero. Force final render progress bars to 100 on completion. Sometimes they didn't seem to make it there. Make variation name and params comparisons be case insensitive. --Code Changes Make ForEach and FindIf wrappers around std::for_each and std::find_if.
2014-07-19 02:33:18 -04:00
<< "\t\treal_t d = Zeps(precalcSumSquares);\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\n"
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
<< "\t\tvOut.x = (" << weight << " / d) * (tanh(d) * ((real_t)(2.0) * vIn.x));\n"
<< "\t\tvOut.y = (" << weight << " / d) * (cos(d) * ((real_t)(2.0) * vIn.y));\n"
--User changes -Allow for pausing the renderer in the main window. This makes is more efficient when entering many parameters, such as when following a tutorial. -Add support for new variations: erf, gamma, jac_cn, jac_dn, jac_sn, logDB, pressure_wave, pRose3D, splits3D, w, waves2b, x, xerf, y, z. -Inform user of the start and stop of file parsing in EmberAnimate because the files could potentially be very large. -Move the follwing fields to a new table called Animation: Interpolation, Affine Interpolation, Temporal Samples, Temporal Filter Width, Temporal Filter Type. -These currently have no effect on the interactive renderer and instead are used when running flames through EmberGenome to generate sequences, and then animating them in Fractorium or EmberAnimate. -Add new parameter overrides for EmberRender and EmberAnimate which directly assign values to all flames being rendered, rather than scale: --quality --demin --demax --Bug fixes -Left pad instead of right pad names of sequence outputs from EmberGenome. -Unique file naming was broken for files which already had an underscore in them. -Properly report that png is the default format of EmberRender and EmberAnimate output instead of erroneously claiming it was jpg. -Make command line programs search these folders in this order for the palette file: ./ ~/.fractorium ~/.config/fractorium /usr/share/fractorium /usr/local/share/fractorium -Fix possible bad values in hexes. -Better assignment of Z variables. -Fix boarders due to use of poorly implemented rint() function from flam3. Use std::rint() now. -wedge_sph was completely wrong due to having accidentally swapped the mapping of two parameters. -Make juliascope juliascope_power parameter be of type REAL_NONZERO since it's used as a denominator. -Make Z assignment compatible with the originals in: -arch, bcircle, bCollide, bent, bent2, bisplit, blob, blur_linear, blur_square, bMod, boarders, boarders2, bSwirl, bTransform, butterfly, cardioid, cell, circleblur, circlize, circlize2, circus, collideoscope, cos, cosine, cosh, coth, cpow, cpow2, crescents, cropn, csc, csch, curl, curve, dc_gridout, deltaa, diamond, disc2, eclipse, eCollide, edisc, eJulia, elliptic, eMod, eMotion, ennepers, epispiral, ePush, eRotate, eScale, eSwirl, ex, exp, expo, exponential, fan, fdisc, fibonacci, fibonacci2, fisheye, flipcircle, flipy, flower, flux, funnel, glynnia, GlynnSim1, GlynnSim2, GlynnSim3, gridout, handkerchief, heart, hole, idisc, julia, julian2, juliaNab, kaleidoscope, lazyTravis, Lissajous, mask, MobiusN, mobius_strip, modulus, murl, murl2, npolar, ortho, oscilloscope, parabola, perspective, petal, phoenix_julia, pie (was also inconsistent between cpu and gpu), poincare, popcorn, popcorn2, power, pow_block, rational3, rays, rblur, rings, rippled, roundspher, sec, secant2, sigmoid, sin, sineblur, sinh, sinusgrid, sphericaln, spiralwing, spirograph, split, squarize, squirrel, squish, sschecks, starblur, stripes, stwin, super_shape, tan, tancos, tangent, tanh, TwinTrian, twoface, unpolar, waves, wavesn, wedge_julia, whorl, xheart, zblur, zscale. --Code changes -Generalize Variation::PrecalcHelper() and rename to PrePostPrecalcHelper(). --Do the same for the OpenCL version and rename it PrePostPrecalcOpenCLString(). -Rename Variation::m_AssignType to m_PrePostAssignType since it's only relevant to pre/post variations.
2016-01-29 20:02:15 -05:00
<< "\t\tvOut.z = " << DefaultZCl()
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t}\n";
return ss.str();
}
--User changes -Add new variations: bubbleT3D, crob, hexaplay3D, hexcrop, hexes, hexnix3D, loonie2, loonie3, nBlur, octapol and synth. -Allow for pre/post versions of dc_bubble, dc_cylinder and dc_linear whereas before they were omitted. -When saving a file with multiple embers in it, detect if time values are all the same and if so, start them at zero and increment by 1 for each ember. -Allow for numerous quality increases to be coalesced into one. It will pick up at the end of the current render. -Show selection highlight on variations tree in response to mouse hover. This makes it easier to see for which variation or param the current mouse wheel action will apply. -Make default temporal samples be 100, whereas before it was 1000 which was overkill. -Require the shift key to be held with delete for deleting an ember to prevent it from triggering when the user enters delete in the edit box. -This wasn't otherwise fixable without writing a lot more code. --Bug fixes -EmberGenome was crashing when generating a sequence from a source file with more than 2 embers in it. -EmberGenome was improperly handling the first frame of a merge after the last frame of the loop. -These bugs were due to a previous commit. Revert parts of that commit. -Prevent a zoom value of less than 0 when reading from xml. -Slight optimization of the crescents, and mask variations, if the compiler wasn't doing it already. -Unique file naming was broken because it was looking for _# and the default names ended with -#. -Disallow renaming of an ember in the library tree to an empty string. -Severe bug that prevented some variations from being read correctly from params generated outside this program. -Severe OpenCL randomization bug. The first x coordinates of the first points in the first kernel call of the first ember of a render since the OpenCL renderer object was created were not random and were mostly -1. -Severe bug when populating xform selection distributions that could sometimes cause a crash due to roundoff error. Fix by using double. -Limit the max number of variations in a random ember to MAX_CL_VARS, which is 8. This ensures they'll look the same on CPU and GPU. -Prevent user from saving stylesheet to default.qss, it's a special reserved filename. --Code changes -Generalize using the running sum output point inside of a variation for all cases: pre, reg and post. -Allow for array variables in variations where the address of each element is stored in m_Params. -Qualify all math functions with std:: -No longer use our own Clamp() in OpenCL, instead use the standard clamp(). -Redesign how functions are used in the variations OpenCL code. -Add tests to EmberTester to verify some of the new functionality. -Place more const and override qualifiers on functions where appropriate. -Add a global rand with a lock to be used very sparingly. -Use a map instead of a vector for bad param names in Xml parsing. -Prefix affine interpolation mode defines with "AFFINE_" to make their purpose more clear. -Allow for variations that change state during iteration by sending a separate copy of the ember to each rendering thread. -Implement this same functionality with a local struct in OpenCL. It's members are the total of all variables that need to change state within an ember. -Add Contains() function to Utils.h. -EmberRender: print names of kernels being printed with --dump_kernel option. -Clean up EmberTester to handle some of the recent changes. -Fix various casts. -Replace % 2 with & 1, even though the compiler was likely doing this already. -Add new file Variations06.h to accommodate new variations. -General cleanup.
2015-11-22 17:15:07 -05:00
virtual vector<string> OpenCLGlobalFuncNames() const override
{
return vector<string> { "Zeps" };
}
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
};
/// <summary>
/// rippled.
/// </summary>
template <typename T>
--Code changes -Split VariationList into .h/cpp files. --Make VariationList.cpp the only place where the variations files are included from. This speeds up the build, but variations can no longer be directly instantiated. --Variations are no longer exported.
2016-02-18 21:58:24 -05:00
class RippledVariation : public Variation<T>
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
public:
--Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums.
2016-01-04 19:50:15 -05:00
RippledVariation(T weight = 1.0) : Variation<T>("rippled", eVariationId::VAR_RIPPLED, weight, true) { }
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
VARCOPY(RippledVariation)
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
Numerous fixes 0.4.0.5 Beta 07/18/2014 --User Changes Allow for vibrancy values > 1. Add flatten and unflatten menu items. Automatically flatten like Apophysis does. Add plugin and new_linear tags to Xml to be compatible with Apophysis. --Bug Fixes Fix blur, blur3d, bubble, cropn, cross, curl, curl3d, epispiral, ho, julia3d, julia3dz, loonie, mirror_x, mirror_y, mirror_z, rotate_x, sinusoidal, spherical, spherical3d, stripes. Unique filename on final render was completely broken. Two severe OpenCL bugs. Random seeds were biased and fusing was being reset too often leading to results that differ from the CPU. Subtle, but sometimes severe bug in the setup of the xaos weights. Use properly defined epsilon by getting the value from std::numeric_limits, rather than hard coding 1e-6 or 1e-10. Omit incorrect usage of epsilon everywhere. It should not be automatically added to denominators. Rather, it should only be used if the denominator is zero. Force final render progress bars to 100 on completion. Sometimes they didn't seem to make it there. Make variation name and params comparisons be case insensitive. --Code Changes Make ForEach and FindIf wrappers around std::for_each and std::find_if.
2014-07-19 02:33:18 -04:00
T d = Zeps(helper.m_PrecalcSumSquares);
06/09/2017 --User changes -dark.qss is now per-OS. -Properly set/reload palette when coming from the palette editor. The latter must be done if they've modified the current palette even if they've clicked cancel. --Bug fixes -Make the following variations safer by using Zeps(): conic, bipolar, edisc, whorl, tan, csc, cot, tanh, sech, csch, coth, auger, bwraps, hypertile3d, hypertile3d1, ortho, poincare, rational3, barycentroid, sschecks, cscq, cschq, scry_3D, splitbrdr, hexcrop, nblur, crob. -Fix bug enabling/disabling overwrite button in palette editor. -Small optimization for gdoffs, use precalcAtanYX. -Properly propagate z through circlesplit, cylinder2 and tile_log variations. -Some values in truchet_fill could've been NaN. --Code changes -Make most installation files read only. -Qualify many calls with std:: to ensure they're not colliding with glm:: -Use auto in more places.
2017-06-09 22:38:06 -04:00
helper.Out.x = (m_Weight / 2) * (std::tanh(d) * (2 * helper.In.x));
helper.Out.y = (m_Weight / 2) * (std::cos(d) * (2 * helper.In.y));
--User changes -Allow for pausing the renderer in the main window. This makes is more efficient when entering many parameters, such as when following a tutorial. -Add support for new variations: erf, gamma, jac_cn, jac_dn, jac_sn, logDB, pressure_wave, pRose3D, splits3D, w, waves2b, x, xerf, y, z. -Inform user of the start and stop of file parsing in EmberAnimate because the files could potentially be very large. -Move the follwing fields to a new table called Animation: Interpolation, Affine Interpolation, Temporal Samples, Temporal Filter Width, Temporal Filter Type. -These currently have no effect on the interactive renderer and instead are used when running flames through EmberGenome to generate sequences, and then animating them in Fractorium or EmberAnimate. -Add new parameter overrides for EmberRender and EmberAnimate which directly assign values to all flames being rendered, rather than scale: --quality --demin --demax --Bug fixes -Left pad instead of right pad names of sequence outputs from EmberGenome. -Unique file naming was broken for files which already had an underscore in them. -Properly report that png is the default format of EmberRender and EmberAnimate output instead of erroneously claiming it was jpg. -Make command line programs search these folders in this order for the palette file: ./ ~/.fractorium ~/.config/fractorium /usr/share/fractorium /usr/local/share/fractorium -Fix possible bad values in hexes. -Better assignment of Z variables. -Fix boarders due to use of poorly implemented rint() function from flam3. Use std::rint() now. -wedge_sph was completely wrong due to having accidentally swapped the mapping of two parameters. -Make juliascope juliascope_power parameter be of type REAL_NONZERO since it's used as a denominator. -Make Z assignment compatible with the originals in: -arch, bcircle, bCollide, bent, bent2, bisplit, blob, blur_linear, blur_square, bMod, boarders, boarders2, bSwirl, bTransform, butterfly, cardioid, cell, circleblur, circlize, circlize2, circus, collideoscope, cos, cosine, cosh, coth, cpow, cpow2, crescents, cropn, csc, csch, curl, curve, dc_gridout, deltaa, diamond, disc2, eclipse, eCollide, edisc, eJulia, elliptic, eMod, eMotion, ennepers, epispiral, ePush, eRotate, eScale, eSwirl, ex, exp, expo, exponential, fan, fdisc, fibonacci, fibonacci2, fisheye, flipcircle, flipy, flower, flux, funnel, glynnia, GlynnSim1, GlynnSim2, GlynnSim3, gridout, handkerchief, heart, hole, idisc, julia, julian2, juliaNab, kaleidoscope, lazyTravis, Lissajous, mask, MobiusN, mobius_strip, modulus, murl, murl2, npolar, ortho, oscilloscope, parabola, perspective, petal, phoenix_julia, pie (was also inconsistent between cpu and gpu), poincare, popcorn, popcorn2, power, pow_block, rational3, rays, rblur, rings, rippled, roundspher, sec, secant2, sigmoid, sin, sineblur, sinh, sinusgrid, sphericaln, spiralwing, spirograph, split, squarize, squirrel, squish, sschecks, starblur, stripes, stwin, super_shape, tan, tancos, tangent, tanh, TwinTrian, twoface, unpolar, waves, wavesn, wedge_julia, whorl, xheart, zblur, zscale. --Code changes -Generalize Variation::PrecalcHelper() and rename to PrePostPrecalcHelper(). --Do the same for the OpenCL version and rename it PrePostPrecalcOpenCLString(). -Rename Variation::m_AssignType to m_PrePostAssignType since it's only relevant to pre/post variations.
2016-01-29 20:02:15 -05:00
helper.Out.z = DefaultZ(helper);
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
}
--Code changes -Better use of const functions. -Ember.pro should not define BUILDING_EMBERCL, instead define BUILDING_EMBER.
2015-06-29 23:13:53 -04:00
virtual string OpenCLString() const override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
ostringstream ss;
0.4.1.3 Beta 10/14/2014 --User Changes Size is no longer fixed to the window size. Size scaling is done differently in the final render dialog. This fixes several bugs. Remove Xml saving size from settings and options dialog, it no longer applies. Final render can be broken into strips. Set default save path to the desktop if none is found in the settings file. Set default output size to 1920x1080 if none is found in the settings file. --Bug Fixes Better memory size reporting in final render dialog. --Code Changes Migrate to C++11, Qt 5.3.1, and Visual Studio 2013. Change most instances of unsigned int to size_t, and int to intmax_t. Add m_OrigPixPerUnit and m_ScaleType to Ember for scaling purposes. Replace some sprintf_s() calls in XmlToEmber with ostringstream. Move more non-templated members into RendererBase. Add CopyVec() overload that takes a per element function pointer. Add vector Memset(). Replace '&' with '+' instead of "&amp;" in XmlToEmber for much faster parsing. Break strips rendering out into EmberCommon and call from EmberRender and Fractorium. Make AddAndWriteBuffer() just call WriteBuffer(). Make AddAndWriteImage() delete the existing image first before replacing it. Add SetOutputTexture() to RendererCL to support making new textures in response to resize events. Remove multiple return statements in RendererCL, and replace with a bool that tracks results. Add ToDouble(), MakeEnd(), ToString() and Exists() wrappers in Fractorium. Add Size() wrapper in EmberFile. Make QString function arguments const QString&, and string with const string&. Make ShowCritical() wrapper for invoking a message box from another thread. Add combo box to TwoButtonWidget and rename.
2014-10-14 11:53:15 -04:00
intmax_t varIndex = IndexInXform();
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
string weight = WeightDefineString();
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
ss << "\t{\n"
Numerous fixes 0.4.0.5 Beta 07/18/2014 --User Changes Allow for vibrancy values > 1. Add flatten and unflatten menu items. Automatically flatten like Apophysis does. Add plugin and new_linear tags to Xml to be compatible with Apophysis. --Bug Fixes Fix blur, blur3d, bubble, cropn, cross, curl, curl3d, epispiral, ho, julia3d, julia3dz, loonie, mirror_x, mirror_y, mirror_z, rotate_x, sinusoidal, spherical, spherical3d, stripes. Unique filename on final render was completely broken. Two severe OpenCL bugs. Random seeds were biased and fusing was being reset too often leading to results that differ from the CPU. Subtle, but sometimes severe bug in the setup of the xaos weights. Use properly defined epsilon by getting the value from std::numeric_limits, rather than hard coding 1e-6 or 1e-10. Omit incorrect usage of epsilon everywhere. It should not be automatically added to denominators. Rather, it should only be used if the denominator is zero. Force final render progress bars to 100 on completion. Sometimes they didn't seem to make it there. Make variation name and params comparisons be case insensitive. --Code Changes Make ForEach and FindIf wrappers around std::for_each and std::find_if.
2014-07-19 02:33:18 -04:00
<< "\t\treal_t d = Zeps(precalcSumSquares);\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\n"
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
<< "\t\tvOut.x = (" << weight << " / (real_t)(2.0)) * (tanh(d) * ((real_t)(2.0) * vIn.x));\n"
<< "\t\tvOut.y = (" << weight << " / (real_t)(2.0)) * (cos(d) * ((real_t)(2.0) * vIn.y));\n"
--User changes -Allow for pausing the renderer in the main window. This makes is more efficient when entering many parameters, such as when following a tutorial. -Add support for new variations: erf, gamma, jac_cn, jac_dn, jac_sn, logDB, pressure_wave, pRose3D, splits3D, w, waves2b, x, xerf, y, z. -Inform user of the start and stop of file parsing in EmberAnimate because the files could potentially be very large. -Move the follwing fields to a new table called Animation: Interpolation, Affine Interpolation, Temporal Samples, Temporal Filter Width, Temporal Filter Type. -These currently have no effect on the interactive renderer and instead are used when running flames through EmberGenome to generate sequences, and then animating them in Fractorium or EmberAnimate. -Add new parameter overrides for EmberRender and EmberAnimate which directly assign values to all flames being rendered, rather than scale: --quality --demin --demax --Bug fixes -Left pad instead of right pad names of sequence outputs from EmberGenome. -Unique file naming was broken for files which already had an underscore in them. -Properly report that png is the default format of EmberRender and EmberAnimate output instead of erroneously claiming it was jpg. -Make command line programs search these folders in this order for the palette file: ./ ~/.fractorium ~/.config/fractorium /usr/share/fractorium /usr/local/share/fractorium -Fix possible bad values in hexes. -Better assignment of Z variables. -Fix boarders due to use of poorly implemented rint() function from flam3. Use std::rint() now. -wedge_sph was completely wrong due to having accidentally swapped the mapping of two parameters. -Make juliascope juliascope_power parameter be of type REAL_NONZERO since it's used as a denominator. -Make Z assignment compatible with the originals in: -arch, bcircle, bCollide, bent, bent2, bisplit, blob, blur_linear, blur_square, bMod, boarders, boarders2, bSwirl, bTransform, butterfly, cardioid, cell, circleblur, circlize, circlize2, circus, collideoscope, cos, cosine, cosh, coth, cpow, cpow2, crescents, cropn, csc, csch, curl, curve, dc_gridout, deltaa, diamond, disc2, eclipse, eCollide, edisc, eJulia, elliptic, eMod, eMotion, ennepers, epispiral, ePush, eRotate, eScale, eSwirl, ex, exp, expo, exponential, fan, fdisc, fibonacci, fibonacci2, fisheye, flipcircle, flipy, flower, flux, funnel, glynnia, GlynnSim1, GlynnSim2, GlynnSim3, gridout, handkerchief, heart, hole, idisc, julia, julian2, juliaNab, kaleidoscope, lazyTravis, Lissajous, mask, MobiusN, mobius_strip, modulus, murl, murl2, npolar, ortho, oscilloscope, parabola, perspective, petal, phoenix_julia, pie (was also inconsistent between cpu and gpu), poincare, popcorn, popcorn2, power, pow_block, rational3, rays, rblur, rings, rippled, roundspher, sec, secant2, sigmoid, sin, sineblur, sinh, sinusgrid, sphericaln, spiralwing, spirograph, split, squarize, squirrel, squish, sschecks, starblur, stripes, stwin, super_shape, tan, tancos, tangent, tanh, TwinTrian, twoface, unpolar, waves, wavesn, wedge_julia, whorl, xheart, zblur, zscale. --Code changes -Generalize Variation::PrecalcHelper() and rename to PrePostPrecalcHelper(). --Do the same for the OpenCL version and rename it PrePostPrecalcOpenCLString(). -Rename Variation::m_AssignType to m_PrePostAssignType since it's only relevant to pre/post variations.
2016-01-29 20:02:15 -05:00
<< "\t\tvOut.z = " << DefaultZCl()
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t}\n";
return ss.str();
}
--User changes -Add new variations: bubbleT3D, crob, hexaplay3D, hexcrop, hexes, hexnix3D, loonie2, loonie3, nBlur, octapol and synth. -Allow for pre/post versions of dc_bubble, dc_cylinder and dc_linear whereas before they were omitted. -When saving a file with multiple embers in it, detect if time values are all the same and if so, start them at zero and increment by 1 for each ember. -Allow for numerous quality increases to be coalesced into one. It will pick up at the end of the current render. -Show selection highlight on variations tree in response to mouse hover. This makes it easier to see for which variation or param the current mouse wheel action will apply. -Make default temporal samples be 100, whereas before it was 1000 which was overkill. -Require the shift key to be held with delete for deleting an ember to prevent it from triggering when the user enters delete in the edit box. -This wasn't otherwise fixable without writing a lot more code. --Bug fixes -EmberGenome was crashing when generating a sequence from a source file with more than 2 embers in it. -EmberGenome was improperly handling the first frame of a merge after the last frame of the loop. -These bugs were due to a previous commit. Revert parts of that commit. -Prevent a zoom value of less than 0 when reading from xml. -Slight optimization of the crescents, and mask variations, if the compiler wasn't doing it already. -Unique file naming was broken because it was looking for _# and the default names ended with -#. -Disallow renaming of an ember in the library tree to an empty string. -Severe bug that prevented some variations from being read correctly from params generated outside this program. -Severe OpenCL randomization bug. The first x coordinates of the first points in the first kernel call of the first ember of a render since the OpenCL renderer object was created were not random and were mostly -1. -Severe bug when populating xform selection distributions that could sometimes cause a crash due to roundoff error. Fix by using double. -Limit the max number of variations in a random ember to MAX_CL_VARS, which is 8. This ensures they'll look the same on CPU and GPU. -Prevent user from saving stylesheet to default.qss, it's a special reserved filename. --Code changes -Generalize using the running sum output point inside of a variation for all cases: pre, reg and post. -Allow for array variables in variations where the address of each element is stored in m_Params. -Qualify all math functions with std:: -No longer use our own Clamp() in OpenCL, instead use the standard clamp(). -Redesign how functions are used in the variations OpenCL code. -Add tests to EmberTester to verify some of the new functionality. -Place more const and override qualifiers on functions where appropriate. -Add a global rand with a lock to be used very sparingly. -Use a map instead of a vector for bad param names in Xml parsing. -Prefix affine interpolation mode defines with "AFFINE_" to make their purpose more clear. -Allow for variations that change state during iteration by sending a separate copy of the ember to each rendering thread. -Implement this same functionality with a local struct in OpenCL. It's members are the total of all variables that need to change state within an ember. -Add Contains() function to Utils.h. -EmberRender: print names of kernels being printed with --dump_kernel option. -Clean up EmberTester to handle some of the recent changes. -Fix various casts. -Replace % 2 with & 1, even though the compiler was likely doing this already. -Add new file Variations06.h to accommodate new variations. -General cleanup.
2015-11-22 17:15:07 -05:00
virtual vector<string> OpenCLGlobalFuncNames() const override
{
return vector<string> { "Zeps" };
}
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
};
/// <summary>
/// RotateX.
Numerous fixes 0.4.0.5 Beta 07/18/2014 --User Changes Allow for vibrancy values > 1. Add flatten and unflatten menu items. Automatically flatten like Apophysis does. Add plugin and new_linear tags to Xml to be compatible with Apophysis. --Bug Fixes Fix blur, blur3d, bubble, cropn, cross, curl, curl3d, epispiral, ho, julia3d, julia3dz, loonie, mirror_x, mirror_y, mirror_z, rotate_x, sinusoidal, spherical, spherical3d, stripes. Unique filename on final render was completely broken. Two severe OpenCL bugs. Random seeds were biased and fusing was being reset too often leading to results that differ from the CPU. Subtle, but sometimes severe bug in the setup of the xaos weights. Use properly defined epsilon by getting the value from std::numeric_limits, rather than hard coding 1e-6 or 1e-10. Omit incorrect usage of epsilon everywhere. It should not be automatically added to denominators. Rather, it should only be used if the denominator is zero. Force final render progress bars to 100 on completion. Sometimes they didn't seem to make it there. Make variation name and params comparisons be case insensitive. --Code Changes Make ForEach and FindIf wrappers around std::for_each and std::find_if.
2014-07-19 02:33:18 -04:00
/// This uses in/out in a rare and different way.
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
/// </summary>
template <typename T>
--Code changes -Split VariationList into .h/cpp files. --Make VariationList.cpp the only place where the variations files are included from. This speeds up the build, but variations can no longer be directly instantiated. --Variations are no longer exported.
2016-02-18 21:58:24 -05:00
class RotateXVariation : public ParametricVariation<T>
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
public:
--Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums.
2016-01-04 19:50:15 -05:00
RotateXVariation(T weight = 1.0) : ParametricVariation<T>("rotate_x", eVariationId::VAR_ROTATE_X, weight)
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
Init();
}
PARVARCOPY(RotateXVariation)
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
Numerous fixes 0.4.0.5 Beta 07/18/2014 --User Changes Allow for vibrancy values > 1. Add flatten and unflatten menu items. Automatically flatten like Apophysis does. Add plugin and new_linear tags to Xml to be compatible with Apophysis. --Bug Fixes Fix blur, blur3d, bubble, cropn, cross, curl, curl3d, epispiral, ho, julia3d, julia3dz, loonie, mirror_x, mirror_y, mirror_z, rotate_x, sinusoidal, spherical, spherical3d, stripes. Unique filename on final render was completely broken. Two severe OpenCL bugs. Random seeds were biased and fusing was being reset too often leading to results that differ from the CPU. Subtle, but sometimes severe bug in the setup of the xaos weights. Use properly defined epsilon by getting the value from std::numeric_limits, rather than hard coding 1e-6 or 1e-10. Omit incorrect usage of epsilon everywhere. It should not be automatically added to denominators. Rather, it should only be used if the denominator is zero. Force final render progress bars to 100 on completion. Sometimes they didn't seem to make it there. Make variation name and params comparisons be case insensitive. --Code Changes Make ForEach and FindIf wrappers around std::for_each and std::find_if.
2014-07-19 02:33:18 -04:00
T z = m_RxCos * helper.In.z - m_RxSin * helper.In.y;
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
--Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums.
2016-01-04 19:50:15 -05:00
if (m_VarType == eVariationType::VARTYPE_REG)
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
--User changes -Add new variations: bubbleT3D, crob, hexaplay3D, hexcrop, hexes, hexnix3D, loonie2, loonie3, nBlur, octapol and synth. -Allow for pre/post versions of dc_bubble, dc_cylinder and dc_linear whereas before they were omitted. -When saving a file with multiple embers in it, detect if time values are all the same and if so, start them at zero and increment by 1 for each ember. -Allow for numerous quality increases to be coalesced into one. It will pick up at the end of the current render. -Show selection highlight on variations tree in response to mouse hover. This makes it easier to see for which variation or param the current mouse wheel action will apply. -Make default temporal samples be 100, whereas before it was 1000 which was overkill. -Require the shift key to be held with delete for deleting an ember to prevent it from triggering when the user enters delete in the edit box. -This wasn't otherwise fixable without writing a lot more code. --Bug fixes -EmberGenome was crashing when generating a sequence from a source file with more than 2 embers in it. -EmberGenome was improperly handling the first frame of a merge after the last frame of the loop. -These bugs were due to a previous commit. Revert parts of that commit. -Prevent a zoom value of less than 0 when reading from xml. -Slight optimization of the crescents, and mask variations, if the compiler wasn't doing it already. -Unique file naming was broken because it was looking for _# and the default names ended with -#. -Disallow renaming of an ember in the library tree to an empty string. -Severe bug that prevented some variations from being read correctly from params generated outside this program. -Severe OpenCL randomization bug. The first x coordinates of the first points in the first kernel call of the first ember of a render since the OpenCL renderer object was created were not random and were mostly -1. -Severe bug when populating xform selection distributions that could sometimes cause a crash due to roundoff error. Fix by using double. -Limit the max number of variations in a random ember to MAX_CL_VARS, which is 8. This ensures they'll look the same on CPU and GPU. -Prevent user from saving stylesheet to default.qss, it's a special reserved filename. --Code changes -Generalize using the running sum output point inside of a variation for all cases: pre, reg and post. -Allow for array variables in variations where the address of each element is stored in m_Params. -Qualify all math functions with std:: -No longer use our own Clamp() in OpenCL, instead use the standard clamp(). -Redesign how functions are used in the variations OpenCL code. -Add tests to EmberTester to verify some of the new functionality. -Place more const and override qualifiers on functions where appropriate. -Add a global rand with a lock to be used very sparingly. -Use a map instead of a vector for bad param names in Xml parsing. -Prefix affine interpolation mode defines with "AFFINE_" to make their purpose more clear. -Allow for variations that change state during iteration by sending a separate copy of the ember to each rendering thread. -Implement this same functionality with a local struct in OpenCL. It's members are the total of all variables that need to change state within an ember. -Add Contains() function to Utils.h. -EmberRender: print names of kernels being printed with --dump_kernel option. -Clean up EmberTester to handle some of the recent changes. -Fix various casts. -Replace % 2 with & 1, even though the compiler was likely doing this already. -Add new file Variations06.h to accommodate new variations. -General cleanup.
2015-11-22 17:15:07 -05:00
helper.Out.x = helper.In.x;
outPoint.m_X = 0;
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
}
else
{
helper.Out.x = helper.In.x;
}
helper.Out.y = m_RxSin * helper.In.z + m_RxCos * helper.In.y;
helper.Out.z = z;
}
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
--Code changes -Better use of const functions. -Ember.pro should not define BUILDING_EMBERCL, instead define BUILDING_EMBER.
2015-06-29 23:13:53 -04:00
virtual string OpenCLString() const override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
ostringstream ss, ss2;
More C++11
2014-09-10 01:41:26 -04:00
int i = 0;
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
ss2 << "_" << XformIndexInEmber() << "]";
string index = ss2.str();
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
string weight = WeightDefineString();
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
string rxSin = "parVars[" + ToUpper(m_Params[i++].Name()) + index;//Precalcs only, no params.
string rxCos = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
ss << "\t{\n"
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
<< "\t\treal_t z = fma(" << rxCos << ", vIn.z, -(" << rxSin << " * vIn.y));\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\n";
--Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums.
2016-01-04 19:50:15 -05:00
if (m_VarType == eVariationType::VARTYPE_REG)
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
ss <<
--User changes -Add post_smartcrop. --Bug fixes -Fix bug in crackle. -Wrong point assignment in hexaplay3D, hexnix3D. -Improper Z assignment in rblur. -Fix inconsistency with original in circlecrop. -Put EMBER_ROOT bakc to ./../../../ in default.pri. This is TBD. --Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums. -Optimize crob, nBlur. -Fix weird assignment statement in falloff3. -Cleanup in VarFuncs::SimplexNoise3D(). -Replace fabs() with std::abs(). -General cleanup.
2016-01-12 23:42:12 -05:00
"\t\tvOut.x = 0;\n"
"\t\toutPoint->m_X = vIn.x;\n";
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
}
else
{
ss <<
--User changes -Add post_smartcrop. --Bug fixes -Fix bug in crackle. -Wrong point assignment in hexaplay3D, hexnix3D. -Improper Z assignment in rblur. -Fix inconsistency with original in circlecrop. -Put EMBER_ROOT bakc to ./../../../ in default.pri. This is TBD. --Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums. -Optimize crob, nBlur. -Fix weird assignment statement in falloff3. -Cleanup in VarFuncs::SimplexNoise3D(). -Replace fabs() with std::abs(). -General cleanup.
2016-01-12 23:42:12 -05:00
"\t\tvOut.x = vIn.x;\n";
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
}
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
ss << "\t\tvOut.y = fma(" << rxSin << ", vIn.z, " << rxCos << " * vIn.y);\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t\tvOut.z = z;\n"
<< "\t}\n";
return ss.str();
}
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
virtual void Precalc() override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
--User changes -Add new variations: bubbleT3D, crob, hexaplay3D, hexcrop, hexes, hexnix3D, loonie2, loonie3, nBlur, octapol and synth. -Allow for pre/post versions of dc_bubble, dc_cylinder and dc_linear whereas before they were omitted. -When saving a file with multiple embers in it, detect if time values are all the same and if so, start them at zero and increment by 1 for each ember. -Allow for numerous quality increases to be coalesced into one. It will pick up at the end of the current render. -Show selection highlight on variations tree in response to mouse hover. This makes it easier to see for which variation or param the current mouse wheel action will apply. -Make default temporal samples be 100, whereas before it was 1000 which was overkill. -Require the shift key to be held with delete for deleting an ember to prevent it from triggering when the user enters delete in the edit box. -This wasn't otherwise fixable without writing a lot more code. --Bug fixes -EmberGenome was crashing when generating a sequence from a source file with more than 2 embers in it. -EmberGenome was improperly handling the first frame of a merge after the last frame of the loop. -These bugs were due to a previous commit. Revert parts of that commit. -Prevent a zoom value of less than 0 when reading from xml. -Slight optimization of the crescents, and mask variations, if the compiler wasn't doing it already. -Unique file naming was broken because it was looking for _# and the default names ended with -#. -Disallow renaming of an ember in the library tree to an empty string. -Severe bug that prevented some variations from being read correctly from params generated outside this program. -Severe OpenCL randomization bug. The first x coordinates of the first points in the first kernel call of the first ember of a render since the OpenCL renderer object was created were not random and were mostly -1. -Severe bug when populating xform selection distributions that could sometimes cause a crash due to roundoff error. Fix by using double. -Limit the max number of variations in a random ember to MAX_CL_VARS, which is 8. This ensures they'll look the same on CPU and GPU. -Prevent user from saving stylesheet to default.qss, it's a special reserved filename. --Code changes -Generalize using the running sum output point inside of a variation for all cases: pre, reg and post. -Allow for array variables in variations where the address of each element is stored in m_Params. -Qualify all math functions with std:: -No longer use our own Clamp() in OpenCL, instead use the standard clamp(). -Redesign how functions are used in the variations OpenCL code. -Add tests to EmberTester to verify some of the new functionality. -Place more const and override qualifiers on functions where appropriate. -Add a global rand with a lock to be used very sparingly. -Use a map instead of a vector for bad param names in Xml parsing. -Prefix affine interpolation mode defines with "AFFINE_" to make their purpose more clear. -Allow for variations that change state during iteration by sending a separate copy of the ember to each rendering thread. -Implement this same functionality with a local struct in OpenCL. It's members are the total of all variables that need to change state within an ember. -Add Contains() function to Utils.h. -EmberRender: print names of kernels being printed with --dump_kernel option. -Clean up EmberTester to handle some of the recent changes. -Fix various casts. -Replace % 2 with & 1, even though the compiler was likely doing this already. -Add new file Variations06.h to accommodate new variations. -General cleanup.
2015-11-22 17:15:07 -05:00
m_RxSin = std::sin(m_Weight * T(M_PI_2));
m_RxCos = std::cos(m_Weight * T(M_PI_2));
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
}
protected:
void Init()
{
string prefix = Prefix();
m_Params.clear();
m_Params.push_back(ParamWithName<T>(true, &m_RxSin, prefix + "rotate_x_sin"));//Precalcs only, no params.
m_Params.push_back(ParamWithName<T>(true, &m_RxCos, prefix + "rotate_x_cos"));//Original used a prefix of rx_, which is incompatible with Ember's design.
}
private:
T m_RxSin;
T m_RxCos;
};
/// <summary>
/// RotateY.
Numerous fixes 0.4.0.5 Beta 07/18/2014 --User Changes Allow for vibrancy values > 1. Add flatten and unflatten menu items. Automatically flatten like Apophysis does. Add plugin and new_linear tags to Xml to be compatible with Apophysis. --Bug Fixes Fix blur, blur3d, bubble, cropn, cross, curl, curl3d, epispiral, ho, julia3d, julia3dz, loonie, mirror_x, mirror_y, mirror_z, rotate_x, sinusoidal, spherical, spherical3d, stripes. Unique filename on final render was completely broken. Two severe OpenCL bugs. Random seeds were biased and fusing was being reset too often leading to results that differ from the CPU. Subtle, but sometimes severe bug in the setup of the xaos weights. Use properly defined epsilon by getting the value from std::numeric_limits, rather than hard coding 1e-6 or 1e-10. Omit incorrect usage of epsilon everywhere. It should not be automatically added to denominators. Rather, it should only be used if the denominator is zero. Force final render progress bars to 100 on completion. Sometimes they didn't seem to make it there. Make variation name and params comparisons be case insensitive. --Code Changes Make ForEach and FindIf wrappers around std::for_each and std::find_if.
2014-07-19 02:33:18 -04:00
/// This uses in/out in a rare and different way.
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
/// </summary>
template <typename T>
--Code changes -Split VariationList into .h/cpp files. --Make VariationList.cpp the only place where the variations files are included from. This speeds up the build, but variations can no longer be directly instantiated. --Variations are no longer exported.
2016-02-18 21:58:24 -05:00
class RotateYVariation : public ParametricVariation<T>
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
public:
--Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums.
2016-01-04 19:50:15 -05:00
RotateYVariation(T weight = 1.0) : ParametricVariation<T>("rotate_y", eVariationId::VAR_ROTATE_Y, weight)
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
Init();
}
PARVARCOPY(RotateYVariation)
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
helper.Out.x = m_RyCos * helper.In.x - m_RySin * helper.In.z;
--Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums.
2016-01-04 19:50:15 -05:00
if (m_VarType == eVariationType::VARTYPE_REG)
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
helper.Out.y = 0;
outPoint.m_Y = helper.In.y;
}
else
{
helper.Out.y = helper.In.y;
}
helper.Out.z = m_RySin * helper.In.x + m_RyCos * helper.In.z;
}
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
--Code changes -Better use of const functions. -Ember.pro should not define BUILDING_EMBERCL, instead define BUILDING_EMBER.
2015-06-29 23:13:53 -04:00
virtual string OpenCLString() const override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
ostringstream ss, ss2;
More C++11
2014-09-10 01:41:26 -04:00
int i = 0;
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
ss2 << "_" << XformIndexInEmber() << "]";
string index = ss2.str();
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
string weight = WeightDefineString();
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
string rySin = "parVars[" + ToUpper(m_Params[i++].Name()) + index;//Precalcs only, no params.
string ryCos = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
ss << "\t{\n"
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
<< "\t\tvOut.x = fma(" << ryCos << ", vIn.x, -(" << rySin << " * vIn.z));\n";
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
--Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums.
2016-01-04 19:50:15 -05:00
if (m_VarType == eVariationType::VARTYPE_REG)
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
ss <<
--User changes -Add post_smartcrop. --Bug fixes -Fix bug in crackle. -Wrong point assignment in hexaplay3D, hexnix3D. -Improper Z assignment in rblur. -Fix inconsistency with original in circlecrop. -Put EMBER_ROOT bakc to ./../../../ in default.pri. This is TBD. --Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums. -Optimize crob, nBlur. -Fix weird assignment statement in falloff3. -Cleanup in VarFuncs::SimplexNoise3D(). -Replace fabs() with std::abs(). -General cleanup.
2016-01-12 23:42:12 -05:00
"\t\tvOut.y = 0;\n"
"\t\toutPoint->m_Y = vIn.y;\n";
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
}
else
{
ss <<
--User changes -Add post_smartcrop. --Bug fixes -Fix bug in crackle. -Wrong point assignment in hexaplay3D, hexnix3D. -Improper Z assignment in rblur. -Fix inconsistency with original in circlecrop. -Put EMBER_ROOT bakc to ./../../../ in default.pri. This is TBD. --Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums. -Optimize crob, nBlur. -Fix weird assignment statement in falloff3. -Cleanup in VarFuncs::SimplexNoise3D(). -Replace fabs() with std::abs(). -General cleanup.
2016-01-12 23:42:12 -05:00
"\t\tvOut.y = vIn.y;\n";
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
}
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
ss << "\t\tvOut.z = fma(" << rySin << ", vIn.x, " << ryCos << " * vIn.z);\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t}\n";
return ss.str();
}
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
virtual void Precalc() override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
--User changes -Add new variations: bubbleT3D, crob, hexaplay3D, hexcrop, hexes, hexnix3D, loonie2, loonie3, nBlur, octapol and synth. -Allow for pre/post versions of dc_bubble, dc_cylinder and dc_linear whereas before they were omitted. -When saving a file with multiple embers in it, detect if time values are all the same and if so, start them at zero and increment by 1 for each ember. -Allow for numerous quality increases to be coalesced into one. It will pick up at the end of the current render. -Show selection highlight on variations tree in response to mouse hover. This makes it easier to see for which variation or param the current mouse wheel action will apply. -Make default temporal samples be 100, whereas before it was 1000 which was overkill. -Require the shift key to be held with delete for deleting an ember to prevent it from triggering when the user enters delete in the edit box. -This wasn't otherwise fixable without writing a lot more code. --Bug fixes -EmberGenome was crashing when generating a sequence from a source file with more than 2 embers in it. -EmberGenome was improperly handling the first frame of a merge after the last frame of the loop. -These bugs were due to a previous commit. Revert parts of that commit. -Prevent a zoom value of less than 0 when reading from xml. -Slight optimization of the crescents, and mask variations, if the compiler wasn't doing it already. -Unique file naming was broken because it was looking for _# and the default names ended with -#. -Disallow renaming of an ember in the library tree to an empty string. -Severe bug that prevented some variations from being read correctly from params generated outside this program. -Severe OpenCL randomization bug. The first x coordinates of the first points in the first kernel call of the first ember of a render since the OpenCL renderer object was created were not random and were mostly -1. -Severe bug when populating xform selection distributions that could sometimes cause a crash due to roundoff error. Fix by using double. -Limit the max number of variations in a random ember to MAX_CL_VARS, which is 8. This ensures they'll look the same on CPU and GPU. -Prevent user from saving stylesheet to default.qss, it's a special reserved filename. --Code changes -Generalize using the running sum output point inside of a variation for all cases: pre, reg and post. -Allow for array variables in variations where the address of each element is stored in m_Params. -Qualify all math functions with std:: -No longer use our own Clamp() in OpenCL, instead use the standard clamp(). -Redesign how functions are used in the variations OpenCL code. -Add tests to EmberTester to verify some of the new functionality. -Place more const and override qualifiers on functions where appropriate. -Add a global rand with a lock to be used very sparingly. -Use a map instead of a vector for bad param names in Xml parsing. -Prefix affine interpolation mode defines with "AFFINE_" to make their purpose more clear. -Allow for variations that change state during iteration by sending a separate copy of the ember to each rendering thread. -Implement this same functionality with a local struct in OpenCL. It's members are the total of all variables that need to change state within an ember. -Add Contains() function to Utils.h. -EmberRender: print names of kernels being printed with --dump_kernel option. -Clean up EmberTester to handle some of the recent changes. -Fix various casts. -Replace % 2 with & 1, even though the compiler was likely doing this already. -Add new file Variations06.h to accommodate new variations. -General cleanup.
2015-11-22 17:15:07 -05:00
m_RySin = std::sin(m_Weight * T(M_PI_2));
m_RyCos = std::cos(m_Weight * T(M_PI_2));
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
}
protected:
void Init()
{
string prefix = Prefix();
m_Params.clear();
m_Params.push_back(ParamWithName<T>(true, &m_RySin, prefix + "rotate_y_sin"));//Precalcs only, no params.
m_Params.push_back(ParamWithName<T>(true, &m_RyCos, prefix + "rotate_y_cos"));//Original used a prefix of ry_, which is incompatible with Ember's design.
}
private:
T m_RySin;
T m_RyCos;
};
/// <summary>
/// RotateZ.
/// This was originally pre and post spin_z, consolidated here to be consistent with the other rotate variations.
/// </summary>
template <typename T>
--Code changes -Split VariationList into .h/cpp files. --Make VariationList.cpp the only place where the variations files are included from. This speeds up the build, but variations can no longer be directly instantiated. --Variations are no longer exported.
2016-02-18 21:58:24 -05:00
class RotateZVariation : public ParametricVariation<T>
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
public:
--Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums.
2016-01-04 19:50:15 -05:00
RotateZVariation(T weight = 1.0) : ParametricVariation<T>("rotate_z", eVariationId::VAR_ROTATE_Z, weight)
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
Init();
}
PARVARCOPY(RotateZVariation)
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
helper.Out.x = m_RzSin * helper.In.y + m_RzCos * helper.In.x;
helper.Out.y = m_RzCos * helper.In.y - m_RzSin * helper.In.x;
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
--Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums.
2016-01-04 19:50:15 -05:00
if (m_VarType == eVariationType::VARTYPE_REG)
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
--User changes -Add new variations: bubbleT3D, crob, hexaplay3D, hexcrop, hexes, hexnix3D, loonie2, loonie3, nBlur, octapol and synth. -Allow for pre/post versions of dc_bubble, dc_cylinder and dc_linear whereas before they were omitted. -When saving a file with multiple embers in it, detect if time values are all the same and if so, start them at zero and increment by 1 for each ember. -Allow for numerous quality increases to be coalesced into one. It will pick up at the end of the current render. -Show selection highlight on variations tree in response to mouse hover. This makes it easier to see for which variation or param the current mouse wheel action will apply. -Make default temporal samples be 100, whereas before it was 1000 which was overkill. -Require the shift key to be held with delete for deleting an ember to prevent it from triggering when the user enters delete in the edit box. -This wasn't otherwise fixable without writing a lot more code. --Bug fixes -EmberGenome was crashing when generating a sequence from a source file with more than 2 embers in it. -EmberGenome was improperly handling the first frame of a merge after the last frame of the loop. -These bugs were due to a previous commit. Revert parts of that commit. -Prevent a zoom value of less than 0 when reading from xml. -Slight optimization of the crescents, and mask variations, if the compiler wasn't doing it already. -Unique file naming was broken because it was looking for _# and the default names ended with -#. -Disallow renaming of an ember in the library tree to an empty string. -Severe bug that prevented some variations from being read correctly from params generated outside this program. -Severe OpenCL randomization bug. The first x coordinates of the first points in the first kernel call of the first ember of a render since the OpenCL renderer object was created were not random and were mostly -1. -Severe bug when populating xform selection distributions that could sometimes cause a crash due to roundoff error. Fix by using double. -Limit the max number of variations in a random ember to MAX_CL_VARS, which is 8. This ensures they'll look the same on CPU and GPU. -Prevent user from saving stylesheet to default.qss, it's a special reserved filename. --Code changes -Generalize using the running sum output point inside of a variation for all cases: pre, reg and post. -Allow for array variables in variations where the address of each element is stored in m_Params. -Qualify all math functions with std:: -No longer use our own Clamp() in OpenCL, instead use the standard clamp(). -Redesign how functions are used in the variations OpenCL code. -Add tests to EmberTester to verify some of the new functionality. -Place more const and override qualifiers on functions where appropriate. -Add a global rand with a lock to be used very sparingly. -Use a map instead of a vector for bad param names in Xml parsing. -Prefix affine interpolation mode defines with "AFFINE_" to make their purpose more clear. -Allow for variations that change state during iteration by sending a separate copy of the ember to each rendering thread. -Implement this same functionality with a local struct in OpenCL. It's members are the total of all variables that need to change state within an ember. -Add Contains() function to Utils.h. -EmberRender: print names of kernels being printed with --dump_kernel option. -Clean up EmberTester to handle some of the recent changes. -Fix various casts. -Replace % 2 with & 1, even though the compiler was likely doing this already. -Add new file Variations06.h to accommodate new variations. -General cleanup.
2015-11-22 17:15:07 -05:00
helper.Out.z = helper.In.z;
outPoint.m_Z = 0;
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
}
else
{
helper.Out.z = helper.In.z;
}
}
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
--Code changes -Better use of const functions. -Ember.pro should not define BUILDING_EMBERCL, instead define BUILDING_EMBER.
2015-06-29 23:13:53 -04:00
virtual string OpenCLString() const override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
ostringstream ss, ss2;
More C++11
2014-09-10 01:41:26 -04:00
int i = 0;
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
ss2 << "_" << XformIndexInEmber() << "]";
string index = ss2.str();
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
string weight = WeightDefineString();
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
string rzSin = "parVars[" + ToUpper(m_Params[i++].Name()) + index;//Precalcs only, no params.
string rzCos = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
ss << "\t{\n"
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
<< "\t\tvOut.x = fma(" << rzSin << ", vIn.y, " << rzCos << " * vIn.x);\n"
<< "\t\tvOut.y = fma(" << rzCos << ", vIn.y, -(" << rzSin << " * vIn.x));\n";
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
--Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums.
2016-01-04 19:50:15 -05:00
if (m_VarType == eVariationType::VARTYPE_REG)
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
ss <<
--User changes -Add post_smartcrop. --Bug fixes -Fix bug in crackle. -Wrong point assignment in hexaplay3D, hexnix3D. -Improper Z assignment in rblur. -Fix inconsistency with original in circlecrop. -Put EMBER_ROOT bakc to ./../../../ in default.pri. This is TBD. --Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums. -Optimize crob, nBlur. -Fix weird assignment statement in falloff3. -Cleanup in VarFuncs::SimplexNoise3D(). -Replace fabs() with std::abs(). -General cleanup.
2016-01-12 23:42:12 -05:00
"\t\tvOut.z = 0;\n"
"\t\toutPoint->m_Z = vIn.z;\n";
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
}
else
{
ss <<
--User changes -Add post_smartcrop. --Bug fixes -Fix bug in crackle. -Wrong point assignment in hexaplay3D, hexnix3D. -Improper Z assignment in rblur. -Fix inconsistency with original in circlecrop. -Put EMBER_ROOT bakc to ./../../../ in default.pri. This is TBD. --Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums. -Optimize crob, nBlur. -Fix weird assignment statement in falloff3. -Cleanup in VarFuncs::SimplexNoise3D(). -Replace fabs() with std::abs(). -General cleanup.
2016-01-12 23:42:12 -05:00
"\t\tvOut.z = vIn.z;\n";
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
}
ss << "\t}\n";
return ss.str();
}
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
virtual void Precalc() override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
--User changes -Add new variations: bubbleT3D, crob, hexaplay3D, hexcrop, hexes, hexnix3D, loonie2, loonie3, nBlur, octapol and synth. -Allow for pre/post versions of dc_bubble, dc_cylinder and dc_linear whereas before they were omitted. -When saving a file with multiple embers in it, detect if time values are all the same and if so, start them at zero and increment by 1 for each ember. -Allow for numerous quality increases to be coalesced into one. It will pick up at the end of the current render. -Show selection highlight on variations tree in response to mouse hover. This makes it easier to see for which variation or param the current mouse wheel action will apply. -Make default temporal samples be 100, whereas before it was 1000 which was overkill. -Require the shift key to be held with delete for deleting an ember to prevent it from triggering when the user enters delete in the edit box. -This wasn't otherwise fixable without writing a lot more code. --Bug fixes -EmberGenome was crashing when generating a sequence from a source file with more than 2 embers in it. -EmberGenome was improperly handling the first frame of a merge after the last frame of the loop. -These bugs were due to a previous commit. Revert parts of that commit. -Prevent a zoom value of less than 0 when reading from xml. -Slight optimization of the crescents, and mask variations, if the compiler wasn't doing it already. -Unique file naming was broken because it was looking for _# and the default names ended with -#. -Disallow renaming of an ember in the library tree to an empty string. -Severe bug that prevented some variations from being read correctly from params generated outside this program. -Severe OpenCL randomization bug. The first x coordinates of the first points in the first kernel call of the first ember of a render since the OpenCL renderer object was created were not random and were mostly -1. -Severe bug when populating xform selection distributions that could sometimes cause a crash due to roundoff error. Fix by using double. -Limit the max number of variations in a random ember to MAX_CL_VARS, which is 8. This ensures they'll look the same on CPU and GPU. -Prevent user from saving stylesheet to default.qss, it's a special reserved filename. --Code changes -Generalize using the running sum output point inside of a variation for all cases: pre, reg and post. -Allow for array variables in variations where the address of each element is stored in m_Params. -Qualify all math functions with std:: -No longer use our own Clamp() in OpenCL, instead use the standard clamp(). -Redesign how functions are used in the variations OpenCL code. -Add tests to EmberTester to verify some of the new functionality. -Place more const and override qualifiers on functions where appropriate. -Add a global rand with a lock to be used very sparingly. -Use a map instead of a vector for bad param names in Xml parsing. -Prefix affine interpolation mode defines with "AFFINE_" to make their purpose more clear. -Allow for variations that change state during iteration by sending a separate copy of the ember to each rendering thread. -Implement this same functionality with a local struct in OpenCL. It's members are the total of all variables that need to change state within an ember. -Add Contains() function to Utils.h. -EmberRender: print names of kernels being printed with --dump_kernel option. -Clean up EmberTester to handle some of the recent changes. -Fix various casts. -Replace % 2 with & 1, even though the compiler was likely doing this already. -Add new file Variations06.h to accommodate new variations. -General cleanup.
2015-11-22 17:15:07 -05:00
m_RzSin = std::sin(m_Weight * T(M_PI_2));
m_RzCos = std::cos(m_Weight * T(M_PI_2));
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
}
protected:
void Init()
{
string prefix = Prefix();
m_Params.clear();
m_Params.push_back(ParamWithName<T>(true, &m_RzSin, prefix + "rotate_z_sin"));//Precalcs only, no params.
m_Params.push_back(ParamWithName<T>(true, &m_RzCos, prefix + "rotate_z_cos"));
}
private:
T m_RzSin;
T m_RzCos;
};
/// <summary>
/// MirrorX.
Numerous fixes 0.4.0.5 Beta 07/18/2014 --User Changes Allow for vibrancy values > 1. Add flatten and unflatten menu items. Automatically flatten like Apophysis does. Add plugin and new_linear tags to Xml to be compatible with Apophysis. --Bug Fixes Fix blur, blur3d, bubble, cropn, cross, curl, curl3d, epispiral, ho, julia3d, julia3dz, loonie, mirror_x, mirror_y, mirror_z, rotate_x, sinusoidal, spherical, spherical3d, stripes. Unique filename on final render was completely broken. Two severe OpenCL bugs. Random seeds were biased and fusing was being reset too often leading to results that differ from the CPU. Subtle, but sometimes severe bug in the setup of the xaos weights. Use properly defined epsilon by getting the value from std::numeric_limits, rather than hard coding 1e-6 or 1e-10. Omit incorrect usage of epsilon everywhere. It should not be automatically added to denominators. Rather, it should only be used if the denominator is zero. Force final render progress bars to 100 on completion. Sometimes they didn't seem to make it there. Make variation name and params comparisons be case insensitive. --Code Changes Make ForEach and FindIf wrappers around std::for_each and std::find_if.
2014-07-19 02:33:18 -04:00
/// This uses in/out in a rare and different way.
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
/// </summary>
template <typename T>
--Code changes -Split VariationList into .h/cpp files. --Make VariationList.cpp the only place where the variations files are included from. This speeds up the build, but variations can no longer be directly instantiated. --Variations are no longer exported.
2016-02-18 21:58:24 -05:00
class MirrorXVariation : public Variation<T>
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
public:
--Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums.
2016-01-04 19:50:15 -05:00
MirrorXVariation(T weight = 1.0) : Variation<T>("mirror_x", eVariationId::VAR_MIRROR_X, weight) { }
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
VARCOPY(MirrorXVariation)
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
--Bug fixes -mirror_x, y and z variations were totally wrong. -Reading color curves from older version 2 .chaos files which did not have the "knots" and "values" names was broken.
2019-05-22 23:33:19 -04:00
helper.Out.x = std::abs(helper.In.x);
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
--Bug fixes -mirror_x, y and z variations were totally wrong. -Reading color curves from older version 2 .chaos files which did not have the "knots" and "values" names was broken.
2019-05-22 23:33:19 -04:00
if (rand.RandBit())
helper.Out.x = -helper.Out.x;
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
--Bug fixes -mirror_x, y and z variations were totally wrong. -Reading color curves from older version 2 .chaos files which did not have the "knots" and "values" names was broken.
2019-05-22 23:33:19 -04:00
helper.Out.y = helper.In.y;
helper.Out.z = helper.In.z;
Numerous fixes 0.4.0.5 Beta 07/18/2014 --User Changes Allow for vibrancy values > 1. Add flatten and unflatten menu items. Automatically flatten like Apophysis does. Add plugin and new_linear tags to Xml to be compatible with Apophysis. --Bug Fixes Fix blur, blur3d, bubble, cropn, cross, curl, curl3d, epispiral, ho, julia3d, julia3dz, loonie, mirror_x, mirror_y, mirror_z, rotate_x, sinusoidal, spherical, spherical3d, stripes. Unique filename on final render was completely broken. Two severe OpenCL bugs. Random seeds were biased and fusing was being reset too often leading to results that differ from the CPU. Subtle, but sometimes severe bug in the setup of the xaos weights. Use properly defined epsilon by getting the value from std::numeric_limits, rather than hard coding 1e-6 or 1e-10. Omit incorrect usage of epsilon everywhere. It should not be automatically added to denominators. Rather, it should only be used if the denominator is zero. Force final render progress bars to 100 on completion. Sometimes they didn't seem to make it there. Make variation name and params comparisons be case insensitive. --Code Changes Make ForEach and FindIf wrappers around std::for_each and std::find_if.
2014-07-19 02:33:18 -04:00
--Bug fixes -mirror_x, y and z variations were totally wrong. -Reading color curves from older version 2 .chaos files which did not have the "knots" and "values" names was broken.
2019-05-22 23:33:19 -04:00
if (m_VarType == eVariationType::VARTYPE_REG)
{
outPoint.m_X = 0;//All will be added.
outPoint.m_Y = 0;
outPoint.m_Z = 0;
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
}
}
--Code changes -Better use of const functions. -Ember.pro should not define BUILDING_EMBERCL, instead define BUILDING_EMBER.
2015-06-29 23:13:53 -04:00
virtual string OpenCLString() const override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
ostringstream ss;
--Bug fixes -mirror_x, y and z variations were totally wrong. -Reading color curves from older version 2 .chaos files which did not have the "knots" and "values" names was broken.
2019-05-22 23:33:19 -04:00
ss << "\t{\n"
"\t\tvOut.x = fabs(vIn.x);\n"
"\n"
"\t\tif (MwcNext(mwc) & 1)\n"
"\t\t vOut.x = -vOut.x;\n"
"\n"
"\t\tvOut.y = vIn.y;\n"
"\t\tvOut.z = vIn.z;\n";
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
--Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums.
2016-01-04 19:50:15 -05:00
if (m_VarType == eVariationType::VARTYPE_REG)
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
ss <<
--Bug fixes -mirror_x, y and z variations were totally wrong. -Reading color curves from older version 2 .chaos files which did not have the "knots" and "values" names was broken.
2019-05-22 23:33:19 -04:00
"\t\toutPoint->m_X = 0;\n"
"\t\toutPoint->m_Y = 0;\n"
"\t\toutPoint->m_Z = 0;\n";
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
}
ss << "\t}\n";
return ss.str();
}
};
/// <summary>
/// MirrorY.
Numerous fixes 0.4.0.5 Beta 07/18/2014 --User Changes Allow for vibrancy values > 1. Add flatten and unflatten menu items. Automatically flatten like Apophysis does. Add plugin and new_linear tags to Xml to be compatible with Apophysis. --Bug Fixes Fix blur, blur3d, bubble, cropn, cross, curl, curl3d, epispiral, ho, julia3d, julia3dz, loonie, mirror_x, mirror_y, mirror_z, rotate_x, sinusoidal, spherical, spherical3d, stripes. Unique filename on final render was completely broken. Two severe OpenCL bugs. Random seeds were biased and fusing was being reset too often leading to results that differ from the CPU. Subtle, but sometimes severe bug in the setup of the xaos weights. Use properly defined epsilon by getting the value from std::numeric_limits, rather than hard coding 1e-6 or 1e-10. Omit incorrect usage of epsilon everywhere. It should not be automatically added to denominators. Rather, it should only be used if the denominator is zero. Force final render progress bars to 100 on completion. Sometimes they didn't seem to make it there. Make variation name and params comparisons be case insensitive. --Code Changes Make ForEach and FindIf wrappers around std::for_each and std::find_if.
2014-07-19 02:33:18 -04:00
/// This uses in/out in a rare and different way.
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
/// </summary>
template <typename T>
--Code changes -Split VariationList into .h/cpp files. --Make VariationList.cpp the only place where the variations files are included from. This speeds up the build, but variations can no longer be directly instantiated. --Variations are no longer exported.
2016-02-18 21:58:24 -05:00
class MirrorYVariation : public Variation<T>
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
public:
--Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums.
2016-01-04 19:50:15 -05:00
MirrorYVariation(T weight = 1.0) : Variation<T>("mirror_y", eVariationId::VAR_MIRROR_Y, weight) { }
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
VARCOPY(MirrorYVariation)
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
--Bug fixes -mirror_x, y and z variations were totally wrong. -Reading color curves from older version 2 .chaos files which did not have the "knots" and "values" names was broken.
2019-05-22 23:33:19 -04:00
helper.Out.y = std::abs(helper.In.y);
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
--Bug fixes -mirror_x, y and z variations were totally wrong. -Reading color curves from older version 2 .chaos files which did not have the "knots" and "values" names was broken.
2019-05-22 23:33:19 -04:00
if (rand.RandBit())
helper.Out.y = -helper.Out.y;
Numerous fixes 0.4.0.5 Beta 07/18/2014 --User Changes Allow for vibrancy values > 1. Add flatten and unflatten menu items. Automatically flatten like Apophysis does. Add plugin and new_linear tags to Xml to be compatible with Apophysis. --Bug Fixes Fix blur, blur3d, bubble, cropn, cross, curl, curl3d, epispiral, ho, julia3d, julia3dz, loonie, mirror_x, mirror_y, mirror_z, rotate_x, sinusoidal, spherical, spherical3d, stripes. Unique filename on final render was completely broken. Two severe OpenCL bugs. Random seeds were biased and fusing was being reset too often leading to results that differ from the CPU. Subtle, but sometimes severe bug in the setup of the xaos weights. Use properly defined epsilon by getting the value from std::numeric_limits, rather than hard coding 1e-6 or 1e-10. Omit incorrect usage of epsilon everywhere. It should not be automatically added to denominators. Rather, it should only be used if the denominator is zero. Force final render progress bars to 100 on completion. Sometimes they didn't seem to make it there. Make variation name and params comparisons be case insensitive. --Code Changes Make ForEach and FindIf wrappers around std::for_each and std::find_if.
2014-07-19 02:33:18 -04:00
--Bug fixes -mirror_x, y and z variations were totally wrong. -Reading color curves from older version 2 .chaos files which did not have the "knots" and "values" names was broken.
2019-05-22 23:33:19 -04:00
helper.Out.x = helper.In.x;
helper.Out.z = helper.In.z;
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
--Bug fixes -mirror_x, y and z variations were totally wrong. -Reading color curves from older version 2 .chaos files which did not have the "knots" and "values" names was broken.
2019-05-22 23:33:19 -04:00
if (m_VarType == eVariationType::VARTYPE_REG)
{
outPoint.m_X = 0;//All will be added.
outPoint.m_Y = 0;
outPoint.m_Z = 0;
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
}
}
--Code changes -Better use of const functions. -Ember.pro should not define BUILDING_EMBERCL, instead define BUILDING_EMBER.
2015-06-29 23:13:53 -04:00
virtual string OpenCLString() const override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
ostringstream ss;
--Bug fixes -mirror_x, y and z variations were totally wrong. -Reading color curves from older version 2 .chaos files which did not have the "knots" and "values" names was broken.
2019-05-22 23:33:19 -04:00
ss << "\t{\n"
"\t\tvOut.y = fabs(vIn.y);\n"
"\n"
"\t\tif (MwcNext(mwc) & 1)\n"
"\t\t vOut.y = -vOut.y;\n"
"\n"
"\t\tvOut.x = vIn.x;\n"
"\t\tvOut.z = vIn.z;\n";
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
--Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums.
2016-01-04 19:50:15 -05:00
if (m_VarType == eVariationType::VARTYPE_REG)
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
ss <<
--Bug fixes -mirror_x, y and z variations were totally wrong. -Reading color curves from older version 2 .chaos files which did not have the "knots" and "values" names was broken.
2019-05-22 23:33:19 -04:00
"\t\toutPoint->m_X = 0;\n"
"\t\toutPoint->m_Y = 0;\n"
"\t\toutPoint->m_Z = 0;\n";
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
}
ss << "\t}\n";
return ss.str();
}
};
/// <summary>
/// MirrorZ.
Numerous fixes 0.4.0.5 Beta 07/18/2014 --User Changes Allow for vibrancy values > 1. Add flatten and unflatten menu items. Automatically flatten like Apophysis does. Add plugin and new_linear tags to Xml to be compatible with Apophysis. --Bug Fixes Fix blur, blur3d, bubble, cropn, cross, curl, curl3d, epispiral, ho, julia3d, julia3dz, loonie, mirror_x, mirror_y, mirror_z, rotate_x, sinusoidal, spherical, spherical3d, stripes. Unique filename on final render was completely broken. Two severe OpenCL bugs. Random seeds were biased and fusing was being reset too often leading to results that differ from the CPU. Subtle, but sometimes severe bug in the setup of the xaos weights. Use properly defined epsilon by getting the value from std::numeric_limits, rather than hard coding 1e-6 or 1e-10. Omit incorrect usage of epsilon everywhere. It should not be automatically added to denominators. Rather, it should only be used if the denominator is zero. Force final render progress bars to 100 on completion. Sometimes they didn't seem to make it there. Make variation name and params comparisons be case insensitive. --Code Changes Make ForEach and FindIf wrappers around std::for_each and std::find_if.
2014-07-19 02:33:18 -04:00
/// This uses in/out in a rare and different way.
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
/// </summary>
template <typename T>
--Code changes -Split VariationList into .h/cpp files. --Make VariationList.cpp the only place where the variations files are included from. This speeds up the build, but variations can no longer be directly instantiated. --Variations are no longer exported.
2016-02-18 21:58:24 -05:00
class MirrorZVariation : public Variation<T>
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
public:
--Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums.
2016-01-04 19:50:15 -05:00
MirrorZVariation(T weight = 1.0) : Variation<T>("mirror_z", eVariationId::VAR_MIRROR_Z, weight) { }
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
VARCOPY(MirrorZVariation)
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
--Bug fixes -mirror_x, y and z variations were totally wrong. -Reading color curves from older version 2 .chaos files which did not have the "knots" and "values" names was broken.
2019-05-22 23:33:19 -04:00
helper.Out.z = std::abs(helper.In.z);
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
--Bug fixes -mirror_x, y and z variations were totally wrong. -Reading color curves from older version 2 .chaos files which did not have the "knots" and "values" names was broken.
2019-05-22 23:33:19 -04:00
if (rand.RandBit())
helper.Out.z = -helper.Out.z;
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
--Bug fixes -mirror_x, y and z variations were totally wrong. -Reading color curves from older version 2 .chaos files which did not have the "knots" and "values" names was broken.
2019-05-22 23:33:19 -04:00
helper.Out.x = helper.In.x;
helper.Out.y = helper.In.y;
Numerous fixes 0.4.0.5 Beta 07/18/2014 --User Changes Allow for vibrancy values > 1. Add flatten and unflatten menu items. Automatically flatten like Apophysis does. Add plugin and new_linear tags to Xml to be compatible with Apophysis. --Bug Fixes Fix blur, blur3d, bubble, cropn, cross, curl, curl3d, epispiral, ho, julia3d, julia3dz, loonie, mirror_x, mirror_y, mirror_z, rotate_x, sinusoidal, spherical, spherical3d, stripes. Unique filename on final render was completely broken. Two severe OpenCL bugs. Random seeds were biased and fusing was being reset too often leading to results that differ from the CPU. Subtle, but sometimes severe bug in the setup of the xaos weights. Use properly defined epsilon by getting the value from std::numeric_limits, rather than hard coding 1e-6 or 1e-10. Omit incorrect usage of epsilon everywhere. It should not be automatically added to denominators. Rather, it should only be used if the denominator is zero. Force final render progress bars to 100 on completion. Sometimes they didn't seem to make it there. Make variation name and params comparisons be case insensitive. --Code Changes Make ForEach and FindIf wrappers around std::for_each and std::find_if.
2014-07-19 02:33:18 -04:00
--Bug fixes -mirror_x, y and z variations were totally wrong. -Reading color curves from older version 2 .chaos files which did not have the "knots" and "values" names was broken.
2019-05-22 23:33:19 -04:00
if (m_VarType == eVariationType::VARTYPE_REG)
{
outPoint.m_X = 0;//All will be added.
outPoint.m_Y = 0;
outPoint.m_Z = 0;
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
}
}
--Code changes -Better use of const functions. -Ember.pro should not define BUILDING_EMBERCL, instead define BUILDING_EMBER.
2015-06-29 23:13:53 -04:00
virtual string OpenCLString() const override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
ostringstream ss;
--Bug fixes -mirror_x, y and z variations were totally wrong. -Reading color curves from older version 2 .chaos files which did not have the "knots" and "values" names was broken.
2019-05-22 23:33:19 -04:00
ss << "\t{\n"
"\t\tvOut.z = fabs(vIn.z);\n"
"\n"
"\t\tif (MwcNext(mwc) & 1)\n"
"\t\t vOut.z = -vOut.z;\n"
"\n"
"\t\tvOut.x = vIn.x;\n"
"\t\tvOut.y = vIn.y;\n";
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
--Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums.
2016-01-04 19:50:15 -05:00
if (m_VarType == eVariationType::VARTYPE_REG)
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
ss <<
--Bug fixes -mirror_x, y and z variations were totally wrong. -Reading color curves from older version 2 .chaos files which did not have the "knots" and "values" names was broken.
2019-05-22 23:33:19 -04:00
"\t\toutPoint->m_X = 0;\n"
"\t\toutPoint->m_Y = 0;\n"
--User changes -Add post_smartcrop. --Bug fixes -Fix bug in crackle. -Wrong point assignment in hexaplay3D, hexnix3D. -Improper Z assignment in rblur. -Fix inconsistency with original in circlecrop. -Put EMBER_ROOT bakc to ./../../../ in default.pri. This is TBD. --Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums. -Optimize crob, nBlur. -Fix weird assignment statement in falloff3. -Cleanup in VarFuncs::SimplexNoise3D(). -Replace fabs() with std::abs(). -General cleanup.
2016-01-12 23:42:12 -05:00
"\t\toutPoint->m_Z = 0;\n";
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
}
Numerous fixes 0.4.0.5 Beta 07/18/2014 --User Changes Allow for vibrancy values > 1. Add flatten and unflatten menu items. Automatically flatten like Apophysis does. Add plugin and new_linear tags to Xml to be compatible with Apophysis. --Bug Fixes Fix blur, blur3d, bubble, cropn, cross, curl, curl3d, epispiral, ho, julia3d, julia3dz, loonie, mirror_x, mirror_y, mirror_z, rotate_x, sinusoidal, spherical, spherical3d, stripes. Unique filename on final render was completely broken. Two severe OpenCL bugs. Random seeds were biased and fusing was being reset too often leading to results that differ from the CPU. Subtle, but sometimes severe bug in the setup of the xaos weights. Use properly defined epsilon by getting the value from std::numeric_limits, rather than hard coding 1e-6 or 1e-10. Omit incorrect usage of epsilon everywhere. It should not be automatically added to denominators. Rather, it should only be used if the denominator is zero. Force final render progress bars to 100 on completion. Sometimes they didn't seem to make it there. Make variation name and params comparisons be case insensitive. --Code Changes Make ForEach and FindIf wrappers around std::for_each and std::find_if.
2014-07-19 02:33:18 -04:00
ss << "\t}\n";
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
return ss.str();
}
};
/// <summary>
/// RBlur.
/// </summary>
template <typename T>
--Code changes -Split VariationList into .h/cpp files. --Make VariationList.cpp the only place where the variations files are included from. This speeds up the build, but variations can no longer be directly instantiated. --Variations are no longer exported.
2016-02-18 21:58:24 -05:00
class RBlurVariation : public ParametricVariation<T>
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
public:
--Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums.
2016-01-04 19:50:15 -05:00
RBlurVariation(T weight = 1.0) : ParametricVariation<T>("rblur", eVariationId::VAR_RBLUR, weight)
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
Init();
}
PARVARCOPY(RBlurVariation)
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
T sx = helper.In.x - m_CenterX;
T sy = helper.In.y - m_CenterY;
Explicitly namespace calls to sqrt and isnan osx/linux have alternative top-level implementations of these methods, so the calls are ambiguous given the - so these should be used explicitly
2015-04-13 07:32:58 -04:00
T r = std::sqrt(SQR(sx) + SQR(sy)) - m_Offset;
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
r = r < 0 ? 0 : r;
r *= m_S2;
helper.Out.x = m_Weight * (helper.In.x + (rand.Frand01<T>() - T(0.5)) * r);
helper.Out.y = m_Weight * (helper.In.y + (rand.Frand01<T>() - T(0.5)) * r);
--User changes -Allow for pausing the renderer in the main window. This makes is more efficient when entering many parameters, such as when following a tutorial. -Add support for new variations: erf, gamma, jac_cn, jac_dn, jac_sn, logDB, pressure_wave, pRose3D, splits3D, w, waves2b, x, xerf, y, z. -Inform user of the start and stop of file parsing in EmberAnimate because the files could potentially be very large. -Move the follwing fields to a new table called Animation: Interpolation, Affine Interpolation, Temporal Samples, Temporal Filter Width, Temporal Filter Type. -These currently have no effect on the interactive renderer and instead are used when running flames through EmberGenome to generate sequences, and then animating them in Fractorium or EmberAnimate. -Add new parameter overrides for EmberRender and EmberAnimate which directly assign values to all flames being rendered, rather than scale: --quality --demin --demax --Bug fixes -Left pad instead of right pad names of sequence outputs from EmberGenome. -Unique file naming was broken for files which already had an underscore in them. -Properly report that png is the default format of EmberRender and EmberAnimate output instead of erroneously claiming it was jpg. -Make command line programs search these folders in this order for the palette file: ./ ~/.fractorium ~/.config/fractorium /usr/share/fractorium /usr/local/share/fractorium -Fix possible bad values in hexes. -Better assignment of Z variables. -Fix boarders due to use of poorly implemented rint() function from flam3. Use std::rint() now. -wedge_sph was completely wrong due to having accidentally swapped the mapping of two parameters. -Make juliascope juliascope_power parameter be of type REAL_NONZERO since it's used as a denominator. -Make Z assignment compatible with the originals in: -arch, bcircle, bCollide, bent, bent2, bisplit, blob, blur_linear, blur_square, bMod, boarders, boarders2, bSwirl, bTransform, butterfly, cardioid, cell, circleblur, circlize, circlize2, circus, collideoscope, cos, cosine, cosh, coth, cpow, cpow2, crescents, cropn, csc, csch, curl, curve, dc_gridout, deltaa, diamond, disc2, eclipse, eCollide, edisc, eJulia, elliptic, eMod, eMotion, ennepers, epispiral, ePush, eRotate, eScale, eSwirl, ex, exp, expo, exponential, fan, fdisc, fibonacci, fibonacci2, fisheye, flipcircle, flipy, flower, flux, funnel, glynnia, GlynnSim1, GlynnSim2, GlynnSim3, gridout, handkerchief, heart, hole, idisc, julia, julian2, juliaNab, kaleidoscope, lazyTravis, Lissajous, mask, MobiusN, mobius_strip, modulus, murl, murl2, npolar, ortho, oscilloscope, parabola, perspective, petal, phoenix_julia, pie (was also inconsistent between cpu and gpu), poincare, popcorn, popcorn2, power, pow_block, rational3, rays, rblur, rings, rippled, roundspher, sec, secant2, sigmoid, sin, sineblur, sinh, sinusgrid, sphericaln, spiralwing, spirograph, split, squarize, squirrel, squish, sschecks, starblur, stripes, stwin, super_shape, tan, tancos, tangent, tanh, TwinTrian, twoface, unpolar, waves, wavesn, wedge_julia, whorl, xheart, zblur, zscale. --Code changes -Generalize Variation::PrecalcHelper() and rename to PrePostPrecalcHelper(). --Do the same for the OpenCL version and rename it PrePostPrecalcOpenCLString(). -Rename Variation::m_AssignType to m_PrePostAssignType since it's only relevant to pre/post variations.
2016-01-29 20:02:15 -05:00
helper.Out.z = DefaultZ(helper);
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
}
--Code changes -Better use of const functions. -Ember.pro should not define BUILDING_EMBERCL, instead define BUILDING_EMBER.
2015-06-29 23:13:53 -04:00
virtual string OpenCLString() const override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
ostringstream ss, ss2;
0.4.1.3 Beta 10/14/2014 --User Changes Size is no longer fixed to the window size. Size scaling is done differently in the final render dialog. This fixes several bugs. Remove Xml saving size from settings and options dialog, it no longer applies. Final render can be broken into strips. Set default save path to the desktop if none is found in the settings file. Set default output size to 1920x1080 if none is found in the settings file. --Bug Fixes Better memory size reporting in final render dialog. --Code Changes Migrate to C++11, Qt 5.3.1, and Visual Studio 2013. Change most instances of unsigned int to size_t, and int to intmax_t. Add m_OrigPixPerUnit and m_ScaleType to Ember for scaling purposes. Replace some sprintf_s() calls in XmlToEmber with ostringstream. Move more non-templated members into RendererBase. Add CopyVec() overload that takes a per element function pointer. Add vector Memset(). Replace '&' with '+' instead of "&amp;" in XmlToEmber for much faster parsing. Break strips rendering out into EmberCommon and call from EmberRender and Fractorium. Make AddAndWriteBuffer() just call WriteBuffer(). Make AddAndWriteImage() delete the existing image first before replacing it. Add SetOutputTexture() to RendererCL to support making new textures in response to resize events. Remove multiple return statements in RendererCL, and replace with a bool that tracks results. Add ToDouble(), MakeEnd(), ToString() and Exists() wrappers in Fractorium. Add Size() wrapper in EmberFile. Make QString function arguments const QString&, and string with const string&. Make ShowCritical() wrapper for invoking a message box from another thread. Add combo box to TwoButtonWidget and rename.
2014-10-14 11:53:15 -04:00
intmax_t i = 0, varIndex = IndexInXform();
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
ss2 << "_" << XformIndexInEmber() << "]";
string index = ss2.str();
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
string weight = WeightDefineString();
--User changes -Add new options to EmberGenome to better help assess which variations have certain characteristics. The options are: --sumvars: Display the names of all regular variations which have the default behavior of summing their outputs and exit. --assignvars: Display the names of all regular variations which have the non-standard behavior of assigning their outputs and exit. --ppsumvars: Display the names of all pre/post variations which have the non-standard behavior of summing their outputs and exit. --ppassignvars: Display the names of all pre/post variations which have the default behavior of assigning their outputs and exit. --dcvars: Display the names of all variations which use direct coloring and exit. --parvars: Display the names of all variations which have parameters and exit. --nonparvars: Display the names of all variations which do not have parameters (weight only) and exit. --Code changes -Make VariationList vectors have const elements since no callers should ever change them. -Add new function to VariationList to retrieve a const ref to the parametric variations. -Move some search functions out of EmberTester and into EmberCommon.h. -General code cleanup.
2016-05-15 02:33:08 -04:00
string strength = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
string offset = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string centerX = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string centerY = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string s2 = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
ss << "\t{\n"
<< "\t\treal_t sx = vIn.x - " << centerX << ";\n"
<< "\t\treal_t sy = vIn.y - " << centerY << ";\n"
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
<< "\t\treal_t r = sqrt(fma(sx, sx, SQR(sy))) - " << offset << ";\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\n"
<< "\t\tr = r < 0 ? 0 : r;\n"
<< "\t\tr *= " << s2 << ";\n"
<< "\n"
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
<< "\t\tvOut.x = " << weight << " * fma(MwcNext01(mwc) - (real_t)(0.5), r, vIn.x);\n"
<< "\t\tvOut.y = " << weight << " * fma(MwcNext01(mwc) - (real_t)(0.5), r, vIn.y);\n"
--User changes -Allow for pausing the renderer in the main window. This makes is more efficient when entering many parameters, such as when following a tutorial. -Add support for new variations: erf, gamma, jac_cn, jac_dn, jac_sn, logDB, pressure_wave, pRose3D, splits3D, w, waves2b, x, xerf, y, z. -Inform user of the start and stop of file parsing in EmberAnimate because the files could potentially be very large. -Move the follwing fields to a new table called Animation: Interpolation, Affine Interpolation, Temporal Samples, Temporal Filter Width, Temporal Filter Type. -These currently have no effect on the interactive renderer and instead are used when running flames through EmberGenome to generate sequences, and then animating them in Fractorium or EmberAnimate. -Add new parameter overrides for EmberRender and EmberAnimate which directly assign values to all flames being rendered, rather than scale: --quality --demin --demax --Bug fixes -Left pad instead of right pad names of sequence outputs from EmberGenome. -Unique file naming was broken for files which already had an underscore in them. -Properly report that png is the default format of EmberRender and EmberAnimate output instead of erroneously claiming it was jpg. -Make command line programs search these folders in this order for the palette file: ./ ~/.fractorium ~/.config/fractorium /usr/share/fractorium /usr/local/share/fractorium -Fix possible bad values in hexes. -Better assignment of Z variables. -Fix boarders due to use of poorly implemented rint() function from flam3. Use std::rint() now. -wedge_sph was completely wrong due to having accidentally swapped the mapping of two parameters. -Make juliascope juliascope_power parameter be of type REAL_NONZERO since it's used as a denominator. -Make Z assignment compatible with the originals in: -arch, bcircle, bCollide, bent, bent2, bisplit, blob, blur_linear, blur_square, bMod, boarders, boarders2, bSwirl, bTransform, butterfly, cardioid, cell, circleblur, circlize, circlize2, circus, collideoscope, cos, cosine, cosh, coth, cpow, cpow2, crescents, cropn, csc, csch, curl, curve, dc_gridout, deltaa, diamond, disc2, eclipse, eCollide, edisc, eJulia, elliptic, eMod, eMotion, ennepers, epispiral, ePush, eRotate, eScale, eSwirl, ex, exp, expo, exponential, fan, fdisc, fibonacci, fibonacci2, fisheye, flipcircle, flipy, flower, flux, funnel, glynnia, GlynnSim1, GlynnSim2, GlynnSim3, gridout, handkerchief, heart, hole, idisc, julia, julian2, juliaNab, kaleidoscope, lazyTravis, Lissajous, mask, MobiusN, mobius_strip, modulus, murl, murl2, npolar, ortho, oscilloscope, parabola, perspective, petal, phoenix_julia, pie (was also inconsistent between cpu and gpu), poincare, popcorn, popcorn2, power, pow_block, rational3, rays, rblur, rings, rippled, roundspher, sec, secant2, sigmoid, sin, sineblur, sinh, sinusgrid, sphericaln, spiralwing, spirograph, split, squarize, squirrel, squish, sschecks, starblur, stripes, stwin, super_shape, tan, tancos, tangent, tanh, TwinTrian, twoface, unpolar, waves, wavesn, wedge_julia, whorl, xheart, zblur, zscale. --Code changes -Generalize Variation::PrecalcHelper() and rename to PrePostPrecalcHelper(). --Do the same for the OpenCL version and rename it PrePostPrecalcOpenCLString(). -Rename Variation::m_AssignType to m_PrePostAssignType since it's only relevant to pre/post variations.
2016-01-29 20:02:15 -05:00
<< "\t\tvOut.z = " << DefaultZCl()
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t}\n";
return ss.str();
}
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
virtual void Precalc() override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
m_S2 = 2 * m_Strength;
}
protected:
void Init()
{
string prefix = Prefix();
m_Params.clear();
m_Params.push_back(ParamWithName<T>(&m_Strength, prefix + "rblur_strength", 1));
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
m_Params.push_back(ParamWithName<T>(&m_Offset, prefix + "rblur_offset", 1));
m_Params.push_back(ParamWithName<T>(&m_CenterX, prefix + "rblur_center_x"));
m_Params.push_back(ParamWithName<T>(&m_CenterY, prefix + "rblur_center_y"));
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
m_Params.push_back(ParamWithName<T>(true, &m_S2, prefix + "rblur_s2"));//Precalc.
}
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
private:
T m_Strength;
T m_Offset;
T m_CenterX;
T m_CenterY;
T m_S2;//Precalc.
};
/// <summary>
/// JuliaNab.
/// </summary>
template <typename T>
--Code changes -Split VariationList into .h/cpp files. --Make VariationList.cpp the only place where the variations files are included from. This speeds up the build, but variations can no longer be directly instantiated. --Variations are no longer exported.
2016-02-18 21:58:24 -05:00
class JuliaNabVariation : public ParametricVariation<T>
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
public:
--Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums.
2016-01-04 19:50:15 -05:00
JuliaNabVariation(T weight = 1.0) : ParametricVariation<T>("juliaNab", eVariationId::VAR_JULIANAB, weight, true)
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
Init();
}
PARVARCOPY(JuliaNabVariation)
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
--User changes -Add post_smartcrop. --Bug fixes -Fix bug in crackle. -Wrong point assignment in hexaplay3D, hexnix3D. -Improper Z assignment in rblur. -Fix inconsistency with original in circlecrop. -Put EMBER_ROOT bakc to ./../../../ in default.pri. This is TBD. --Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums. -Optimize crob, nBlur. -Fix weird assignment statement in falloff3. -Cleanup in VarFuncs::SimplexNoise3D(). -Replace fabs() with std::abs(). -General cleanup.
2016-01-12 23:42:12 -05:00
T jun = Zeps(std::abs(m_N));
T a = (std::atan2(helper.In.y, std::pow(std::abs(helper.In.x), m_Sep)) + M_2PI * Floor<T>(rand.Frand01<T>() * m_AbsN)) / jun;
--User changes -Add new variations: bubbleT3D, crob, hexaplay3D, hexcrop, hexes, hexnix3D, loonie2, loonie3, nBlur, octapol and synth. -Allow for pre/post versions of dc_bubble, dc_cylinder and dc_linear whereas before they were omitted. -When saving a file with multiple embers in it, detect if time values are all the same and if so, start them at zero and increment by 1 for each ember. -Allow for numerous quality increases to be coalesced into one. It will pick up at the end of the current render. -Show selection highlight on variations tree in response to mouse hover. This makes it easier to see for which variation or param the current mouse wheel action will apply. -Make default temporal samples be 100, whereas before it was 1000 which was overkill. -Require the shift key to be held with delete for deleting an ember to prevent it from triggering when the user enters delete in the edit box. -This wasn't otherwise fixable without writing a lot more code. --Bug fixes -EmberGenome was crashing when generating a sequence from a source file with more than 2 embers in it. -EmberGenome was improperly handling the first frame of a merge after the last frame of the loop. -These bugs were due to a previous commit. Revert parts of that commit. -Prevent a zoom value of less than 0 when reading from xml. -Slight optimization of the crescents, and mask variations, if the compiler wasn't doing it already. -Unique file naming was broken because it was looking for _# and the default names ended with -#. -Disallow renaming of an ember in the library tree to an empty string. -Severe bug that prevented some variations from being read correctly from params generated outside this program. -Severe OpenCL randomization bug. The first x coordinates of the first points in the first kernel call of the first ember of a render since the OpenCL renderer object was created were not random and were mostly -1. -Severe bug when populating xform selection distributions that could sometimes cause a crash due to roundoff error. Fix by using double. -Limit the max number of variations in a random ember to MAX_CL_VARS, which is 8. This ensures they'll look the same on CPU and GPU. -Prevent user from saving stylesheet to default.qss, it's a special reserved filename. --Code changes -Generalize using the running sum output point inside of a variation for all cases: pre, reg and post. -Allow for array variables in variations where the address of each element is stored in m_Params. -Qualify all math functions with std:: -No longer use our own Clamp() in OpenCL, instead use the standard clamp(). -Redesign how functions are used in the variations OpenCL code. -Add tests to EmberTester to verify some of the new functionality. -Place more const and override qualifiers on functions where appropriate. -Add a global rand with a lock to be used very sparingly. -Use a map instead of a vector for bad param names in Xml parsing. -Prefix affine interpolation mode defines with "AFFINE_" to make their purpose more clear. -Allow for variations that change state during iteration by sending a separate copy of the ember to each rendering thread. -Implement this same functionality with a local struct in OpenCL. It's members are the total of all variables that need to change state within an ember. -Add Contains() function to Utils.h. -EmberRender: print names of kernels being printed with --dump_kernel option. -Clean up EmberTester to handle some of the recent changes. -Fix various casts. -Replace % 2 with & 1, even though the compiler was likely doing this already. -Add new file Variations06.h to accommodate new variations. -General cleanup.
2015-11-22 17:15:07 -05:00
T r = m_Weight * std::pow(helper.m_PrecalcSumSquares, m_Cn * m_A);
helper.Out.x = r * std::cos(a) + m_B;
helper.Out.y = r * std::sin(a) + m_B;
--User changes -Allow for pausing the renderer in the main window. This makes is more efficient when entering many parameters, such as when following a tutorial. -Add support for new variations: erf, gamma, jac_cn, jac_dn, jac_sn, logDB, pressure_wave, pRose3D, splits3D, w, waves2b, x, xerf, y, z. -Inform user of the start and stop of file parsing in EmberAnimate because the files could potentially be very large. -Move the follwing fields to a new table called Animation: Interpolation, Affine Interpolation, Temporal Samples, Temporal Filter Width, Temporal Filter Type. -These currently have no effect on the interactive renderer and instead are used when running flames through EmberGenome to generate sequences, and then animating them in Fractorium or EmberAnimate. -Add new parameter overrides for EmberRender and EmberAnimate which directly assign values to all flames being rendered, rather than scale: --quality --demin --demax --Bug fixes -Left pad instead of right pad names of sequence outputs from EmberGenome. -Unique file naming was broken for files which already had an underscore in them. -Properly report that png is the default format of EmberRender and EmberAnimate output instead of erroneously claiming it was jpg. -Make command line programs search these folders in this order for the palette file: ./ ~/.fractorium ~/.config/fractorium /usr/share/fractorium /usr/local/share/fractorium -Fix possible bad values in hexes. -Better assignment of Z variables. -Fix boarders due to use of poorly implemented rint() function from flam3. Use std::rint() now. -wedge_sph was completely wrong due to having accidentally swapped the mapping of two parameters. -Make juliascope juliascope_power parameter be of type REAL_NONZERO since it's used as a denominator. -Make Z assignment compatible with the originals in: -arch, bcircle, bCollide, bent, bent2, bisplit, blob, blur_linear, blur_square, bMod, boarders, boarders2, bSwirl, bTransform, butterfly, cardioid, cell, circleblur, circlize, circlize2, circus, collideoscope, cos, cosine, cosh, coth, cpow, cpow2, crescents, cropn, csc, csch, curl, curve, dc_gridout, deltaa, diamond, disc2, eclipse, eCollide, edisc, eJulia, elliptic, eMod, eMotion, ennepers, epispiral, ePush, eRotate, eScale, eSwirl, ex, exp, expo, exponential, fan, fdisc, fibonacci, fibonacci2, fisheye, flipcircle, flipy, flower, flux, funnel, glynnia, GlynnSim1, GlynnSim2, GlynnSim3, gridout, handkerchief, heart, hole, idisc, julia, julian2, juliaNab, kaleidoscope, lazyTravis, Lissajous, mask, MobiusN, mobius_strip, modulus, murl, murl2, npolar, ortho, oscilloscope, parabola, perspective, petal, phoenix_julia, pie (was also inconsistent between cpu and gpu), poincare, popcorn, popcorn2, power, pow_block, rational3, rays, rblur, rings, rippled, roundspher, sec, secant2, sigmoid, sin, sineblur, sinh, sinusgrid, sphericaln, spiralwing, spirograph, split, squarize, squirrel, squish, sschecks, starblur, stripes, stwin, super_shape, tan, tancos, tangent, tanh, TwinTrian, twoface, unpolar, waves, wavesn, wedge_julia, whorl, xheart, zblur, zscale. --Code changes -Generalize Variation::PrecalcHelper() and rename to PrePostPrecalcHelper(). --Do the same for the OpenCL version and rename it PrePostPrecalcOpenCLString(). -Rename Variation::m_AssignType to m_PrePostAssignType since it's only relevant to pre/post variations.
2016-01-29 20:02:15 -05:00
helper.Out.z = helper.In.z;
if (m_VarType == eVariationType::VARTYPE_REG)
outPoint.m_Z = 0;
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
}
--Code changes -Better use of const functions. -Ember.pro should not define BUILDING_EMBERCL, instead define BUILDING_EMBER.
2015-06-29 23:13:53 -04:00
virtual string OpenCLString() const override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
ostringstream ss, ss2;
0.4.1.3 Beta 10/14/2014 --User Changes Size is no longer fixed to the window size. Size scaling is done differently in the final render dialog. This fixes several bugs. Remove Xml saving size from settings and options dialog, it no longer applies. Final render can be broken into strips. Set default save path to the desktop if none is found in the settings file. Set default output size to 1920x1080 if none is found in the settings file. --Bug Fixes Better memory size reporting in final render dialog. --Code Changes Migrate to C++11, Qt 5.3.1, and Visual Studio 2013. Change most instances of unsigned int to size_t, and int to intmax_t. Add m_OrigPixPerUnit and m_ScaleType to Ember for scaling purposes. Replace some sprintf_s() calls in XmlToEmber with ostringstream. Move more non-templated members into RendererBase. Add CopyVec() overload that takes a per element function pointer. Add vector Memset(). Replace '&' with '+' instead of "&amp;" in XmlToEmber for much faster parsing. Break strips rendering out into EmberCommon and call from EmberRender and Fractorium. Make AddAndWriteBuffer() just call WriteBuffer(). Make AddAndWriteImage() delete the existing image first before replacing it. Add SetOutputTexture() to RendererCL to support making new textures in response to resize events. Remove multiple return statements in RendererCL, and replace with a bool that tracks results. Add ToDouble(), MakeEnd(), ToString() and Exists() wrappers in Fractorium. Add Size() wrapper in EmberFile. Make QString function arguments const QString&, and string with const string&. Make ShowCritical() wrapper for invoking a message box from another thread. Add combo box to TwoButtonWidget and rename.
2014-10-14 11:53:15 -04:00
intmax_t i = 0, varIndex = IndexInXform();
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
ss2 << "_" << XformIndexInEmber() << "]";
string index = ss2.str();
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
string weight = WeightDefineString();
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
string n = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string a = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string b = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string sep = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
--User changes -Add new options to EmberGenome to better help assess which variations have certain characteristics. The options are: --sumvars: Display the names of all regular variations which have the default behavior of summing their outputs and exit. --assignvars: Display the names of all regular variations which have the non-standard behavior of assigning their outputs and exit. --ppsumvars: Display the names of all pre/post variations which have the non-standard behavior of summing their outputs and exit. --ppassignvars: Display the names of all pre/post variations which have the default behavior of assigning their outputs and exit. --dcvars: Display the names of all variations which use direct coloring and exit. --parvars: Display the names of all variations which have parameters and exit. --nonparvars: Display the names of all variations which do not have parameters (weight only) and exit. --Code changes -Make VariationList vectors have const elements since no callers should ever change them. -Add new function to VariationList to retrieve a const ref to the parametric variations. -Move some search functions out of EmberTester and into EmberCommon.h. -General code cleanup.
2016-05-15 02:33:08 -04:00
string absN = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
string cn = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
ss << "\t{\n"
Numerous fixes 0.4.0.5 Beta 07/18/2014 --User Changes Allow for vibrancy values > 1. Add flatten and unflatten menu items. Automatically flatten like Apophysis does. Add plugin and new_linear tags to Xml to be compatible with Apophysis. --Bug Fixes Fix blur, blur3d, bubble, cropn, cross, curl, curl3d, epispiral, ho, julia3d, julia3dz, loonie, mirror_x, mirror_y, mirror_z, rotate_x, sinusoidal, spherical, spherical3d, stripes. Unique filename on final render was completely broken. Two severe OpenCL bugs. Random seeds were biased and fusing was being reset too often leading to results that differ from the CPU. Subtle, but sometimes severe bug in the setup of the xaos weights. Use properly defined epsilon by getting the value from std::numeric_limits, rather than hard coding 1e-6 or 1e-10. Omit incorrect usage of epsilon everywhere. It should not be automatically added to denominators. Rather, it should only be used if the denominator is zero. Force final render progress bars to 100 on completion. Sometimes they didn't seem to make it there. Make variation name and params comparisons be case insensitive. --Code Changes Make ForEach and FindIf wrappers around std::for_each and std::find_if.
2014-07-19 02:33:18 -04:00
<< "\t\treal_t jun = Zeps(fabs(" << n << "));\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\n"
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
<< "\t\treal_t a = fma(M_2PI, floor(MwcNext01(mwc) * " << absN << "), atan2(vIn.y, pow(fabs(vIn.x), " << sep << "))) / jun;\n"
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
<< "\t\treal_t r = " << weight << " * pow(precalcSumSquares, " << cn << " * " << a << ");\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\n"
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
<< "\t\tvOut.x = fma(r, cos(a), " << b << ");\n"
<< "\t\tvOut.y = fma(r, sin(a), " << b << ");\n"
--User changes -Allow for pausing the renderer in the main window. This makes is more efficient when entering many parameters, such as when following a tutorial. -Add support for new variations: erf, gamma, jac_cn, jac_dn, jac_sn, logDB, pressure_wave, pRose3D, splits3D, w, waves2b, x, xerf, y, z. -Inform user of the start and stop of file parsing in EmberAnimate because the files could potentially be very large. -Move the follwing fields to a new table called Animation: Interpolation, Affine Interpolation, Temporal Samples, Temporal Filter Width, Temporal Filter Type. -These currently have no effect on the interactive renderer and instead are used when running flames through EmberGenome to generate sequences, and then animating them in Fractorium or EmberAnimate. -Add new parameter overrides for EmberRender and EmberAnimate which directly assign values to all flames being rendered, rather than scale: --quality --demin --demax --Bug fixes -Left pad instead of right pad names of sequence outputs from EmberGenome. -Unique file naming was broken for files which already had an underscore in them. -Properly report that png is the default format of EmberRender and EmberAnimate output instead of erroneously claiming it was jpg. -Make command line programs search these folders in this order for the palette file: ./ ~/.fractorium ~/.config/fractorium /usr/share/fractorium /usr/local/share/fractorium -Fix possible bad values in hexes. -Better assignment of Z variables. -Fix boarders due to use of poorly implemented rint() function from flam3. Use std::rint() now. -wedge_sph was completely wrong due to having accidentally swapped the mapping of two parameters. -Make juliascope juliascope_power parameter be of type REAL_NONZERO since it's used as a denominator. -Make Z assignment compatible with the originals in: -arch, bcircle, bCollide, bent, bent2, bisplit, blob, blur_linear, blur_square, bMod, boarders, boarders2, bSwirl, bTransform, butterfly, cardioid, cell, circleblur, circlize, circlize2, circus, collideoscope, cos, cosine, cosh, coth, cpow, cpow2, crescents, cropn, csc, csch, curl, curve, dc_gridout, deltaa, diamond, disc2, eclipse, eCollide, edisc, eJulia, elliptic, eMod, eMotion, ennepers, epispiral, ePush, eRotate, eScale, eSwirl, ex, exp, expo, exponential, fan, fdisc, fibonacci, fibonacci2, fisheye, flipcircle, flipy, flower, flux, funnel, glynnia, GlynnSim1, GlynnSim2, GlynnSim3, gridout, handkerchief, heart, hole, idisc, julia, julian2, juliaNab, kaleidoscope, lazyTravis, Lissajous, mask, MobiusN, mobius_strip, modulus, murl, murl2, npolar, ortho, oscilloscope, parabola, perspective, petal, phoenix_julia, pie (was also inconsistent between cpu and gpu), poincare, popcorn, popcorn2, power, pow_block, rational3, rays, rblur, rings, rippled, roundspher, sec, secant2, sigmoid, sin, sineblur, sinh, sinusgrid, sphericaln, spiralwing, spirograph, split, squarize, squirrel, squish, sschecks, starblur, stripes, stwin, super_shape, tan, tancos, tangent, tanh, TwinTrian, twoface, unpolar, waves, wavesn, wedge_julia, whorl, xheart, zblur, zscale. --Code changes -Generalize Variation::PrecalcHelper() and rename to PrePostPrecalcHelper(). --Do the same for the OpenCL version and rename it PrePostPrecalcOpenCLString(). -Rename Variation::m_AssignType to m_PrePostAssignType since it's only relevant to pre/post variations.
2016-01-29 20:02:15 -05:00
<< "\t\tvOut.z = vIn.z;\n";
if (m_VarType == eVariationType::VARTYPE_REG)
ss << "\t\toutPoint->m_Z = 0;\n";
ss
<< "\t}\n";
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
return ss.str();
}
--User changes -Add new variations: bubbleT3D, crob, hexaplay3D, hexcrop, hexes, hexnix3D, loonie2, loonie3, nBlur, octapol and synth. -Allow for pre/post versions of dc_bubble, dc_cylinder and dc_linear whereas before they were omitted. -When saving a file with multiple embers in it, detect if time values are all the same and if so, start them at zero and increment by 1 for each ember. -Allow for numerous quality increases to be coalesced into one. It will pick up at the end of the current render. -Show selection highlight on variations tree in response to mouse hover. This makes it easier to see for which variation or param the current mouse wheel action will apply. -Make default temporal samples be 100, whereas before it was 1000 which was overkill. -Require the shift key to be held with delete for deleting an ember to prevent it from triggering when the user enters delete in the edit box. -This wasn't otherwise fixable without writing a lot more code. --Bug fixes -EmberGenome was crashing when generating a sequence from a source file with more than 2 embers in it. -EmberGenome was improperly handling the first frame of a merge after the last frame of the loop. -These bugs were due to a previous commit. Revert parts of that commit. -Prevent a zoom value of less than 0 when reading from xml. -Slight optimization of the crescents, and mask variations, if the compiler wasn't doing it already. -Unique file naming was broken because it was looking for _# and the default names ended with -#. -Disallow renaming of an ember in the library tree to an empty string. -Severe bug that prevented some variations from being read correctly from params generated outside this program. -Severe OpenCL randomization bug. The first x coordinates of the first points in the first kernel call of the first ember of a render since the OpenCL renderer object was created were not random and were mostly -1. -Severe bug when populating xform selection distributions that could sometimes cause a crash due to roundoff error. Fix by using double. -Limit the max number of variations in a random ember to MAX_CL_VARS, which is 8. This ensures they'll look the same on CPU and GPU. -Prevent user from saving stylesheet to default.qss, it's a special reserved filename. --Code changes -Generalize using the running sum output point inside of a variation for all cases: pre, reg and post. -Allow for array variables in variations where the address of each element is stored in m_Params. -Qualify all math functions with std:: -No longer use our own Clamp() in OpenCL, instead use the standard clamp(). -Redesign how functions are used in the variations OpenCL code. -Add tests to EmberTester to verify some of the new functionality. -Place more const and override qualifiers on functions where appropriate. -Add a global rand with a lock to be used very sparingly. -Use a map instead of a vector for bad param names in Xml parsing. -Prefix affine interpolation mode defines with "AFFINE_" to make their purpose more clear. -Allow for variations that change state during iteration by sending a separate copy of the ember to each rendering thread. -Implement this same functionality with a local struct in OpenCL. It's members are the total of all variables that need to change state within an ember. -Add Contains() function to Utils.h. -EmberRender: print names of kernels being printed with --dump_kernel option. -Clean up EmberTester to handle some of the recent changes. -Fix various casts. -Replace % 2 with & 1, even though the compiler was likely doing this already. -Add new file Variations06.h to accommodate new variations. -General cleanup.
2015-11-22 17:15:07 -05:00
virtual vector<string> OpenCLGlobalFuncNames() const override
{
return vector<string> { "Zeps" };
}
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
virtual void Precalc() override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
--User changes -Add post_smartcrop. --Bug fixes -Fix bug in crackle. -Wrong point assignment in hexaplay3D, hexnix3D. -Improper Z assignment in rblur. -Fix inconsistency with original in circlecrop. -Put EMBER_ROOT bakc to ./../../../ in default.pri. This is TBD. --Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums. -Optimize crob, nBlur. -Fix weird assignment statement in falloff3. -Cleanup in VarFuncs::SimplexNoise3D(). -Replace fabs() with std::abs(). -General cleanup.
2016-01-12 23:42:12 -05:00
T jun = Zeps(std::abs(m_N));
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
m_AbsN = abs(m_N);
m_Cn = 1 / jun / 2;
}
protected:
void Init()
{
string prefix = Prefix();
m_Params.clear();
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
m_Params.push_back(ParamWithName<T>(&m_N, prefix + "juliaNab_n", 1));
m_Params.push_back(ParamWithName<T>(&m_A, prefix + "juliaNab_a", 1));
m_Params.push_back(ParamWithName<T>(&m_B, prefix + "juliaNab_b", 1));
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
m_Params.push_back(ParamWithName<T>(&m_Sep, prefix + "juliaNab_separ", 1));
m_Params.push_back(ParamWithName<T>(true, &m_AbsN, prefix + "juliaNab_absn"));//Precalc.
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
m_Params.push_back(ParamWithName<T>(true, &m_Cn, prefix + "juliaNab_cn"));
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
}
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
private:
T m_N;
T m_A;
T m_B;
T m_Sep;
T m_AbsN;//Precalc.
T m_Cn;
};
/// <summary>
/// Sintrange.
/// </summary>
template <typename T>
--Code changes -Split VariationList into .h/cpp files. --Make VariationList.cpp the only place where the variations files are included from. This speeds up the build, but variations can no longer be directly instantiated. --Variations are no longer exported.
2016-02-18 21:58:24 -05:00
class SintrangeVariation : public ParametricVariation<T>
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
public:
--Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums.
2016-01-04 19:50:15 -05:00
SintrangeVariation(T weight = 1.0) : ParametricVariation<T>("sintrange", eVariationId::VAR_SINTRANGE, weight)
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
Init();
}
PARVARCOPY(SintrangeVariation)
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
T sqX = SQR(helper.In.x);
T sqY = SQR(helper.In.y);
T v = (sqX + sqY) * m_W;//Do not use precalcSumSquares here because its components are needed below.
--User changes -Add new variations: bubbleT3D, crob, hexaplay3D, hexcrop, hexes, hexnix3D, loonie2, loonie3, nBlur, octapol and synth. -Allow for pre/post versions of dc_bubble, dc_cylinder and dc_linear whereas before they were omitted. -When saving a file with multiple embers in it, detect if time values are all the same and if so, start them at zero and increment by 1 for each ember. -Allow for numerous quality increases to be coalesced into one. It will pick up at the end of the current render. -Show selection highlight on variations tree in response to mouse hover. This makes it easier to see for which variation or param the current mouse wheel action will apply. -Make default temporal samples be 100, whereas before it was 1000 which was overkill. -Require the shift key to be held with delete for deleting an ember to prevent it from triggering when the user enters delete in the edit box. -This wasn't otherwise fixable without writing a lot more code. --Bug fixes -EmberGenome was crashing when generating a sequence from a source file with more than 2 embers in it. -EmberGenome was improperly handling the first frame of a merge after the last frame of the loop. -These bugs were due to a previous commit. Revert parts of that commit. -Prevent a zoom value of less than 0 when reading from xml. -Slight optimization of the crescents, and mask variations, if the compiler wasn't doing it already. -Unique file naming was broken because it was looking for _# and the default names ended with -#. -Disallow renaming of an ember in the library tree to an empty string. -Severe bug that prevented some variations from being read correctly from params generated outside this program. -Severe OpenCL randomization bug. The first x coordinates of the first points in the first kernel call of the first ember of a render since the OpenCL renderer object was created were not random and were mostly -1. -Severe bug when populating xform selection distributions that could sometimes cause a crash due to roundoff error. Fix by using double. -Limit the max number of variations in a random ember to MAX_CL_VARS, which is 8. This ensures they'll look the same on CPU and GPU. -Prevent user from saving stylesheet to default.qss, it's a special reserved filename. --Code changes -Generalize using the running sum output point inside of a variation for all cases: pre, reg and post. -Allow for array variables in variations where the address of each element is stored in m_Params. -Qualify all math functions with std:: -No longer use our own Clamp() in OpenCL, instead use the standard clamp(). -Redesign how functions are used in the variations OpenCL code. -Add tests to EmberTester to verify some of the new functionality. -Place more const and override qualifiers on functions where appropriate. -Add a global rand with a lock to be used very sparingly. -Use a map instead of a vector for bad param names in Xml parsing. -Prefix affine interpolation mode defines with "AFFINE_" to make their purpose more clear. -Allow for variations that change state during iteration by sending a separate copy of the ember to each rendering thread. -Implement this same functionality with a local struct in OpenCL. It's members are the total of all variables that need to change state within an ember. -Add Contains() function to Utils.h. -EmberRender: print names of kernels being printed with --dump_kernel option. -Clean up EmberTester to handle some of the recent changes. -Fix various casts. -Replace % 2 with & 1, even though the compiler was likely doing this already. -Add new file Variations06.h to accommodate new variations. -General cleanup.
2015-11-22 17:15:07 -05:00
helper.Out.x = m_Weight * std::sin(helper.In.x) * (sqX + m_W - v);
helper.Out.y = m_Weight * std::sin(helper.In.y) * (sqY + m_W - v);
--User changes -Allow for pausing the renderer in the main window. This makes is more efficient when entering many parameters, such as when following a tutorial. -Add support for new variations: erf, gamma, jac_cn, jac_dn, jac_sn, logDB, pressure_wave, pRose3D, splits3D, w, waves2b, x, xerf, y, z. -Inform user of the start and stop of file parsing in EmberAnimate because the files could potentially be very large. -Move the follwing fields to a new table called Animation: Interpolation, Affine Interpolation, Temporal Samples, Temporal Filter Width, Temporal Filter Type. -These currently have no effect on the interactive renderer and instead are used when running flames through EmberGenome to generate sequences, and then animating them in Fractorium or EmberAnimate. -Add new parameter overrides for EmberRender and EmberAnimate which directly assign values to all flames being rendered, rather than scale: --quality --demin --demax --Bug fixes -Left pad instead of right pad names of sequence outputs from EmberGenome. -Unique file naming was broken for files which already had an underscore in them. -Properly report that png is the default format of EmberRender and EmberAnimate output instead of erroneously claiming it was jpg. -Make command line programs search these folders in this order for the palette file: ./ ~/.fractorium ~/.config/fractorium /usr/share/fractorium /usr/local/share/fractorium -Fix possible bad values in hexes. -Better assignment of Z variables. -Fix boarders due to use of poorly implemented rint() function from flam3. Use std::rint() now. -wedge_sph was completely wrong due to having accidentally swapped the mapping of two parameters. -Make juliascope juliascope_power parameter be of type REAL_NONZERO since it's used as a denominator. -Make Z assignment compatible with the originals in: -arch, bcircle, bCollide, bent, bent2, bisplit, blob, blur_linear, blur_square, bMod, boarders, boarders2, bSwirl, bTransform, butterfly, cardioid, cell, circleblur, circlize, circlize2, circus, collideoscope, cos, cosine, cosh, coth, cpow, cpow2, crescents, cropn, csc, csch, curl, curve, dc_gridout, deltaa, diamond, disc2, eclipse, eCollide, edisc, eJulia, elliptic, eMod, eMotion, ennepers, epispiral, ePush, eRotate, eScale, eSwirl, ex, exp, expo, exponential, fan, fdisc, fibonacci, fibonacci2, fisheye, flipcircle, flipy, flower, flux, funnel, glynnia, GlynnSim1, GlynnSim2, GlynnSim3, gridout, handkerchief, heart, hole, idisc, julia, julian2, juliaNab, kaleidoscope, lazyTravis, Lissajous, mask, MobiusN, mobius_strip, modulus, murl, murl2, npolar, ortho, oscilloscope, parabola, perspective, petal, phoenix_julia, pie (was also inconsistent between cpu and gpu), poincare, popcorn, popcorn2, power, pow_block, rational3, rays, rblur, rings, rippled, roundspher, sec, secant2, sigmoid, sin, sineblur, sinh, sinusgrid, sphericaln, spiralwing, spirograph, split, squarize, squirrel, squish, sschecks, starblur, stripes, stwin, super_shape, tan, tancos, tangent, tanh, TwinTrian, twoface, unpolar, waves, wavesn, wedge_julia, whorl, xheart, zblur, zscale. --Code changes -Generalize Variation::PrecalcHelper() and rename to PrePostPrecalcHelper(). --Do the same for the OpenCL version and rename it PrePostPrecalcOpenCLString(). -Rename Variation::m_AssignType to m_PrePostAssignType since it's only relevant to pre/post variations.
2016-01-29 20:02:15 -05:00
helper.Out.z = DefaultZ(helper);
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
}
--Code changes -Better use of const functions. -Ember.pro should not define BUILDING_EMBERCL, instead define BUILDING_EMBER.
2015-06-29 23:13:53 -04:00
virtual string OpenCLString() const override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
ostringstream ss, ss2;
0.4.1.3 Beta 10/14/2014 --User Changes Size is no longer fixed to the window size. Size scaling is done differently in the final render dialog. This fixes several bugs. Remove Xml saving size from settings and options dialog, it no longer applies. Final render can be broken into strips. Set default save path to the desktop if none is found in the settings file. Set default output size to 1920x1080 if none is found in the settings file. --Bug Fixes Better memory size reporting in final render dialog. --Code Changes Migrate to C++11, Qt 5.3.1, and Visual Studio 2013. Change most instances of unsigned int to size_t, and int to intmax_t. Add m_OrigPixPerUnit and m_ScaleType to Ember for scaling purposes. Replace some sprintf_s() calls in XmlToEmber with ostringstream. Move more non-templated members into RendererBase. Add CopyVec() overload that takes a per element function pointer. Add vector Memset(). Replace '&' with '+' instead of "&amp;" in XmlToEmber for much faster parsing. Break strips rendering out into EmberCommon and call from EmberRender and Fractorium. Make AddAndWriteBuffer() just call WriteBuffer(). Make AddAndWriteImage() delete the existing image first before replacing it. Add SetOutputTexture() to RendererCL to support making new textures in response to resize events. Remove multiple return statements in RendererCL, and replace with a bool that tracks results. Add ToDouble(), MakeEnd(), ToString() and Exists() wrappers in Fractorium. Add Size() wrapper in EmberFile. Make QString function arguments const QString&, and string with const string&. Make ShowCritical() wrapper for invoking a message box from another thread. Add combo box to TwoButtonWidget and rename.
2014-10-14 11:53:15 -04:00
intmax_t i = 0, varIndex = IndexInXform();
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
ss2 << "_" << XformIndexInEmber() << "]";
string index = ss2.str();
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
string weight = WeightDefineString();
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
string w = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
ss << "\t{\n"
<< "\t\treal_t sqX = SQR(vIn.x);\n"
<< "\t\treal_t sqY = SQR(vIn.y);\n"
<< "\t\treal_t v = (sqX + sqY) * " << w << ";\n"
<< "\n"
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
<< "\t\tvOut.x = " << weight << " * sin(vIn.x) * (sqX + " << w << " - v);\n"
<< "\t\tvOut.y = " << weight << " * sin(vIn.y) * (sqY + " << w << " - v);\n"
--User changes -Allow for pausing the renderer in the main window. This makes is more efficient when entering many parameters, such as when following a tutorial. -Add support for new variations: erf, gamma, jac_cn, jac_dn, jac_sn, logDB, pressure_wave, pRose3D, splits3D, w, waves2b, x, xerf, y, z. -Inform user of the start and stop of file parsing in EmberAnimate because the files could potentially be very large. -Move the follwing fields to a new table called Animation: Interpolation, Affine Interpolation, Temporal Samples, Temporal Filter Width, Temporal Filter Type. -These currently have no effect on the interactive renderer and instead are used when running flames through EmberGenome to generate sequences, and then animating them in Fractorium or EmberAnimate. -Add new parameter overrides for EmberRender and EmberAnimate which directly assign values to all flames being rendered, rather than scale: --quality --demin --demax --Bug fixes -Left pad instead of right pad names of sequence outputs from EmberGenome. -Unique file naming was broken for files which already had an underscore in them. -Properly report that png is the default format of EmberRender and EmberAnimate output instead of erroneously claiming it was jpg. -Make command line programs search these folders in this order for the palette file: ./ ~/.fractorium ~/.config/fractorium /usr/share/fractorium /usr/local/share/fractorium -Fix possible bad values in hexes. -Better assignment of Z variables. -Fix boarders due to use of poorly implemented rint() function from flam3. Use std::rint() now. -wedge_sph was completely wrong due to having accidentally swapped the mapping of two parameters. -Make juliascope juliascope_power parameter be of type REAL_NONZERO since it's used as a denominator. -Make Z assignment compatible with the originals in: -arch, bcircle, bCollide, bent, bent2, bisplit, blob, blur_linear, blur_square, bMod, boarders, boarders2, bSwirl, bTransform, butterfly, cardioid, cell, circleblur, circlize, circlize2, circus, collideoscope, cos, cosine, cosh, coth, cpow, cpow2, crescents, cropn, csc, csch, curl, curve, dc_gridout, deltaa, diamond, disc2, eclipse, eCollide, edisc, eJulia, elliptic, eMod, eMotion, ennepers, epispiral, ePush, eRotate, eScale, eSwirl, ex, exp, expo, exponential, fan, fdisc, fibonacci, fibonacci2, fisheye, flipcircle, flipy, flower, flux, funnel, glynnia, GlynnSim1, GlynnSim2, GlynnSim3, gridout, handkerchief, heart, hole, idisc, julia, julian2, juliaNab, kaleidoscope, lazyTravis, Lissajous, mask, MobiusN, mobius_strip, modulus, murl, murl2, npolar, ortho, oscilloscope, parabola, perspective, petal, phoenix_julia, pie (was also inconsistent between cpu and gpu), poincare, popcorn, popcorn2, power, pow_block, rational3, rays, rblur, rings, rippled, roundspher, sec, secant2, sigmoid, sin, sineblur, sinh, sinusgrid, sphericaln, spiralwing, spirograph, split, squarize, squirrel, squish, sschecks, starblur, stripes, stwin, super_shape, tan, tancos, tangent, tanh, TwinTrian, twoface, unpolar, waves, wavesn, wedge_julia, whorl, xheart, zblur, zscale. --Code changes -Generalize Variation::PrecalcHelper() and rename to PrePostPrecalcHelper(). --Do the same for the OpenCL version and rename it PrePostPrecalcOpenCLString(). -Rename Variation::m_AssignType to m_PrePostAssignType since it's only relevant to pre/post variations.
2016-01-29 20:02:15 -05:00
<< "\t\tvOut.z = " << DefaultZCl()
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t}\n";
return ss.str();
}
protected:
void Init()
{
string prefix = Prefix();
m_Params.clear();
m_Params.push_back(ParamWithName<T>(&m_W, prefix + "sintrange_w", 1));
}
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
private:
T m_W;
};
/// <summary>
/// Voron.
/// </summary>
template <typename T>
--Code changes -Split VariationList into .h/cpp files. --Make VariationList.cpp the only place where the variations files are included from. This speeds up the build, but variations can no longer be directly instantiated. --Variations are no longer exported.
2016-02-18 21:58:24 -05:00
class VoronVariation : public ParametricVariation<T>
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
public:
--Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums.
2016-01-04 19:50:15 -05:00
VoronVariation(T weight = 1.0) : ParametricVariation<T>("Voron", eVariationId::VAR_VORON, weight)
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
Init();
}
PARVARCOPY(VoronVariation)
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
--User changes -Add support for multiple GPU devices. --These options are present in the command line and in Fractorium. -Change scheme of specifying devices from platform,device to just total device index. --Single number on the command line. --Change from combo boxes for device selection to a table of all devices in Fractorium. -Temporal samples defaults to 100 instead of 1000 which was needless overkill. --Bug fixes -EmberAnimate, EmberRender, FractoriumSettings, FinalRenderDialog: Fix wrong order of arguments to Clamp() when assigning thread priority. -VariationsDC.h: Fix NVidia OpenCL compilation error in DCTriangleVariation. -FractoriumXformsColor.cpp: Checking for null pixmap pointer is not enough, must also check if the underlying buffer is null via call to QPixmap::isNull(). --Code changes -Ember.h: Add case for FLAME_MOTION_NONE and default in ApplyFlameMotion(). -EmberMotion.h: Call base constructor. -EmberPch.h: #pragma once only on Windows. -EmberToXml.h: --Handle different types of exceptions. --Add default cases to ToString(). -Isaac.h: Remove unused variable in constructor. -Point.h: Call base constructor in Color(). -Renderer.h/cpp: --Add bool to Alloc() to only allocate memory for the histogram. Needed for multi-GPU. --Make CoordMap() return a const ref, not a pointer. -SheepTools.h: --Use 64-bit types like the rest of the code already does. --Fix some comment misspellings. -Timing.h: Make BeginTime(), EndTime(), ElapsedTime() and Format() be const functions. -Utils.h: --Add new functions Equal() and Split(). --Handle more exception types in ReadFile(). --Get rid of most legacy blending of C and C++ argument parsing. -XmlToEmber.h: --Get rid of most legacy blending of C and C++ code from flam3. --Remove some unused variables. -EmberAnimate: --Support multi-GPU processing that alternates full frames between devices. --Use OpenCLInfo instead of OpenCLWrapper for --openclinfo option. --Remove bucketT template parameter, and hard code float in its place. --If a render fails, exit since there is no point in continuing an animation with a missing frame. --Pass variables to threaded save better, which most likely fixes a very subtle bug that existed before. --Remove some unused variables. -EmberGenome, EmberRender: --Support multi-GPU processing that alternates full frames between devices. --Use OpenCLInfo instead of OpenCLWrapper for --openclinfo option. --Remove bucketT template parameter, and hard code float in its place. -EmberRender: --Support multi-GPU processing that alternates full frames between devices. --Use OpenCLInfo instead of OpenCLWrapper for --openclinfo option. --Remove bucketT template parameter, and hard code float in its place. --Only print values when not rendering with OpenCL, since they're always 0 in that case. -EmberCLPch.h: --#pragma once only on Windows. --#include <atomic>. -IterOpenCLKernelCreator.h: Add new kernel for summing two histograms. This is needed for multi-GPU. -OpenCLWrapper.h: --Move all OpenCL info related code into its own class OpenCLInfo. --Add members to cache the values of global memory size and max allocation size. -RendererCL.h/cpp: --Redesign to accomodate multi-GPU. --Constructor now takes a vector of devices. --Remove DumpErrorReport() function, it's handled in the base. --ClearBuffer(), ReadPoints(), WritePoints(), ReadHist() and WriteHist() now optionally take a device index as a parameter. --MakeDmap() override and m_DmapCL member removed because it no longer applies since the histogram is always float since the last commit. --Add new function SumDeviceHist() to sum histograms from two devices by first copying to a temporary on the host, then a temporary on the device, then summing. --m_Calls member removed, as it's now per-device. --OpenCLWrapper removed. --m_Seeds member is now a vector of vector of seeds, to accomodate a separate and different array of seeds for each device. --Added member m_Devices, a vector of unique_ptr of RendererCLDevice. -EmberCommon.h --Added Devices() function to convert from a vector of device indices to a vector of platform,device indices. --Changed CreateRenderer() to accept a vector of devices to create a single RendererCL which will split work across multiple devices. --Added CreateRenderers() function to accept a vector of devices to create multiple RendererCL, each which will render on a single device. --Add more comments to some existing functions. -EmberCommonPch.h: #pragma once only on Windows. -EmberOptions.h: --Remove --platform option, it's just sequential device number now with the --device option. --Make --out be OPT_USE_RENDER instead of OPT_RENDER_ANIM since it's an error condition when animating. It makes no sense to write all frames to a single image. --Add Devices() function to parse comma separated --device option string and return a vector of device indices. --Make int and uint types be 64-bit, so intmax_t and size_t. --Make better use of macros. -JpegUtils.h: Make string parameters to WriteJpeg() and WritePng() be const ref. -All project files: Turn off buffer security check option in Visual Studio (/Gs-) -deployment.pri: Remove the line OTHER_FILES +=, it's pointless and was causing problems. -Ember.pro, EmberCL.pro: Add CONFIG += plugin, otherwise it wouldn't link. -EmberCL.pro: Add new files for multi-GPU support. -build_all.sh: use -j4 and QMAKE=${QMAKE:/usr/bin/qmake} -shared_settings.pri: -Add version string. -Remove old DESTDIR definitions. -Add the following lines or else nothing would build: CONFIG(release, debug|release) { CONFIG += warn_off DESTDIR = ../../../Bin/release } CONFIG(debug, debug|release) { DESTDIR = ../../../Bin/debug } QMAKE_POST_LINK += $$quote(cp --update ../../../Data/flam3-palettes.xml $${DESTDIR}$$escape_expand(\n\t)) LIBS += -L/usr/lib -lpthread -AboutDialog.ui: Another futile attempt to make it look correct on Linux. -FinalRenderDialog.h/cpp: --Add support for multi-GPU. --Change from combo boxes for device selection to a table of all devices. --Ensure device selection makes sense. --Remove "FinalRender" prefix of various function names, it's implied given the context. -FinalRenderEmberController.h/cpp: --Add support for multi-GPU. --Change m_FinishedImageCount to be atomic. --Move CancelRender() from the base to FinalRenderEmberController<T>. --Refactor RenderComplete() to omit any progress related functionality or image saving since it can be potentially ran in a thread. --Consolidate setting various renderer fields into SyncGuiToRenderer(). -Fractorium.cpp: Allow for resizing of the options dialog to show the entire device table. -FractoriumCommon.h: Add various functions to handle a table showing the available OpenCL devices on the system. -FractoriumEmberController.h/cpp: Remove m_FinalImageIndex, it's no longer needed. -FractoriumRender.cpp: Scale the interactive sub batch count and quality by the number of devices used. -FractoriumSettings.h/cpp: --Temporal samples defaults to 100 instead of 1000 which was needless overkill. --Add multi-GPU support, remove old device,platform pair. -FractoriumToolbar.cpp: Disable OpenCL toolbar button if there are no devices present on the system. -FractoriumOptionsDialog.h/cpp: --Add support for multi-GPU. --Consolidate more assignments in DataToGui(). --Enable/disable CPU/OpenCL items in response to OpenCL checkbox event. -Misc: Convert almost everything to size_t for unsigned, intmax_t for signed.
2015-09-12 21:33:45 -04:00
intmax_t l, k;
int i, j, m, m1, n, n1;
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
T r, rMin, offsetX, offsetY, x0 = 0, y0 = 0, x, y;
rMin = 20;
--User changes -Add support for multiple GPU devices. --These options are present in the command line and in Fractorium. -Change scheme of specifying devices from platform,device to just total device index. --Single number on the command line. --Change from combo boxes for device selection to a table of all devices in Fractorium. -Temporal samples defaults to 100 instead of 1000 which was needless overkill. --Bug fixes -EmberAnimate, EmberRender, FractoriumSettings, FinalRenderDialog: Fix wrong order of arguments to Clamp() when assigning thread priority. -VariationsDC.h: Fix NVidia OpenCL compilation error in DCTriangleVariation. -FractoriumXformsColor.cpp: Checking for null pixmap pointer is not enough, must also check if the underlying buffer is null via call to QPixmap::isNull(). --Code changes -Ember.h: Add case for FLAME_MOTION_NONE and default in ApplyFlameMotion(). -EmberMotion.h: Call base constructor. -EmberPch.h: #pragma once only on Windows. -EmberToXml.h: --Handle different types of exceptions. --Add default cases to ToString(). -Isaac.h: Remove unused variable in constructor. -Point.h: Call base constructor in Color(). -Renderer.h/cpp: --Add bool to Alloc() to only allocate memory for the histogram. Needed for multi-GPU. --Make CoordMap() return a const ref, not a pointer. -SheepTools.h: --Use 64-bit types like the rest of the code already does. --Fix some comment misspellings. -Timing.h: Make BeginTime(), EndTime(), ElapsedTime() and Format() be const functions. -Utils.h: --Add new functions Equal() and Split(). --Handle more exception types in ReadFile(). --Get rid of most legacy blending of C and C++ argument parsing. -XmlToEmber.h: --Get rid of most legacy blending of C and C++ code from flam3. --Remove some unused variables. -EmberAnimate: --Support multi-GPU processing that alternates full frames between devices. --Use OpenCLInfo instead of OpenCLWrapper for --openclinfo option. --Remove bucketT template parameter, and hard code float in its place. --If a render fails, exit since there is no point in continuing an animation with a missing frame. --Pass variables to threaded save better, which most likely fixes a very subtle bug that existed before. --Remove some unused variables. -EmberGenome, EmberRender: --Support multi-GPU processing that alternates full frames between devices. --Use OpenCLInfo instead of OpenCLWrapper for --openclinfo option. --Remove bucketT template parameter, and hard code float in its place. -EmberRender: --Support multi-GPU processing that alternates full frames between devices. --Use OpenCLInfo instead of OpenCLWrapper for --openclinfo option. --Remove bucketT template parameter, and hard code float in its place. --Only print values when not rendering with OpenCL, since they're always 0 in that case. -EmberCLPch.h: --#pragma once only on Windows. --#include <atomic>. -IterOpenCLKernelCreator.h: Add new kernel for summing two histograms. This is needed for multi-GPU. -OpenCLWrapper.h: --Move all OpenCL info related code into its own class OpenCLInfo. --Add members to cache the values of global memory size and max allocation size. -RendererCL.h/cpp: --Redesign to accomodate multi-GPU. --Constructor now takes a vector of devices. --Remove DumpErrorReport() function, it's handled in the base. --ClearBuffer(), ReadPoints(), WritePoints(), ReadHist() and WriteHist() now optionally take a device index as a parameter. --MakeDmap() override and m_DmapCL member removed because it no longer applies since the histogram is always float since the last commit. --Add new function SumDeviceHist() to sum histograms from two devices by first copying to a temporary on the host, then a temporary on the device, then summing. --m_Calls member removed, as it's now per-device. --OpenCLWrapper removed. --m_Seeds member is now a vector of vector of seeds, to accomodate a separate and different array of seeds for each device. --Added member m_Devices, a vector of unique_ptr of RendererCLDevice. -EmberCommon.h --Added Devices() function to convert from a vector of device indices to a vector of platform,device indices. --Changed CreateRenderer() to accept a vector of devices to create a single RendererCL which will split work across multiple devices. --Added CreateRenderers() function to accept a vector of devices to create multiple RendererCL, each which will render on a single device. --Add more comments to some existing functions. -EmberCommonPch.h: #pragma once only on Windows. -EmberOptions.h: --Remove --platform option, it's just sequential device number now with the --device option. --Make --out be OPT_USE_RENDER instead of OPT_RENDER_ANIM since it's an error condition when animating. It makes no sense to write all frames to a single image. --Add Devices() function to parse comma separated --device option string and return a vector of device indices. --Make int and uint types be 64-bit, so intmax_t and size_t. --Make better use of macros. -JpegUtils.h: Make string parameters to WriteJpeg() and WritePng() be const ref. -All project files: Turn off buffer security check option in Visual Studio (/Gs-) -deployment.pri: Remove the line OTHER_FILES +=, it's pointless and was causing problems. -Ember.pro, EmberCL.pro: Add CONFIG += plugin, otherwise it wouldn't link. -EmberCL.pro: Add new files for multi-GPU support. -build_all.sh: use -j4 and QMAKE=${QMAKE:/usr/bin/qmake} -shared_settings.pri: -Add version string. -Remove old DESTDIR definitions. -Add the following lines or else nothing would build: CONFIG(release, debug|release) { CONFIG += warn_off DESTDIR = ../../../Bin/release } CONFIG(debug, debug|release) { DESTDIR = ../../../Bin/debug } QMAKE_POST_LINK += $$quote(cp --update ../../../Data/flam3-palettes.xml $${DESTDIR}$$escape_expand(\n\t)) LIBS += -L/usr/lib -lpthread -AboutDialog.ui: Another futile attempt to make it look correct on Linux. -FinalRenderDialog.h/cpp: --Add support for multi-GPU. --Change from combo boxes for device selection to a table of all devices. --Ensure device selection makes sense. --Remove "FinalRender" prefix of various function names, it's implied given the context. -FinalRenderEmberController.h/cpp: --Add support for multi-GPU. --Change m_FinishedImageCount to be atomic. --Move CancelRender() from the base to FinalRenderEmberController<T>. --Refactor RenderComplete() to omit any progress related functionality or image saving since it can be potentially ran in a thread. --Consolidate setting various renderer fields into SyncGuiToRenderer(). -Fractorium.cpp: Allow for resizing of the options dialog to show the entire device table. -FractoriumCommon.h: Add various functions to handle a table showing the available OpenCL devices on the system. -FractoriumEmberController.h/cpp: Remove m_FinalImageIndex, it's no longer needed. -FractoriumRender.cpp: Scale the interactive sub batch count and quality by the number of devices used. -FractoriumSettings.h/cpp: --Temporal samples defaults to 100 instead of 1000 which was needless overkill. --Add multi-GPU support, remove old device,platform pair. -FractoriumToolbar.cpp: Disable OpenCL toolbar button if there are no devices present on the system. -FractoriumOptionsDialog.h/cpp: --Add support for multi-GPU. --Consolidate more assignments in DataToGui(). --Enable/disable CPU/OpenCL items in response to OpenCL checkbox event. -Misc: Convert almost everything to size_t for unsigned, intmax_t for signed.
2015-09-12 21:33:45 -04:00
m = int(Floor<T>(helper.In.x / m_Step));
n = int(Floor<T>(helper.In.y / m_Step));
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
for (i = -1; i < 2; i++)
{
m1 = m + i;
for (j = -1; j < 2; j++)
{
n1 = n + j;
k = 1 + Floor<T>(m_Num * DiscreteNoise(int(19 * m1 + 257 * n1 + m_XSeed)));
for (l = 0; l < k; l++)
{
x = T(DiscreteNoise(int(l + 64 * m1 + 15 * n1 + m_XSeed)) + m1) * m_Step;
y = T(DiscreteNoise(int(l + 21 * m1 + 33 * n1 + m_YSeed)) + n1) * m_Step;
offsetX = helper.In.x - x;
offsetY = helper.In.y - y;
Explicitly namespace calls to sqrt and isnan osx/linux have alternative top-level implementations of these methods, so the calls are ambiguous given the - so these should be used explicitly
2015-04-13 07:32:58 -04:00
r = std::sqrt(SQR(offsetX) + SQR(offsetY));
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
if (r < rMin)
{
rMin = r;
x0 = x;
y0 = y;
}
}
}
}
helper.Out.x = m_Weight * (m_K * (helper.In.x - x0) + x0);
helper.Out.y = m_Weight * (m_K * (helper.In.y - y0) + y0);
--User changes -Allow for pausing the renderer in the main window. This makes is more efficient when entering many parameters, such as when following a tutorial. -Add support for new variations: erf, gamma, jac_cn, jac_dn, jac_sn, logDB, pressure_wave, pRose3D, splits3D, w, waves2b, x, xerf, y, z. -Inform user of the start and stop of file parsing in EmberAnimate because the files could potentially be very large. -Move the follwing fields to a new table called Animation: Interpolation, Affine Interpolation, Temporal Samples, Temporal Filter Width, Temporal Filter Type. -These currently have no effect on the interactive renderer and instead are used when running flames through EmberGenome to generate sequences, and then animating them in Fractorium or EmberAnimate. -Add new parameter overrides for EmberRender and EmberAnimate which directly assign values to all flames being rendered, rather than scale: --quality --demin --demax --Bug fixes -Left pad instead of right pad names of sequence outputs from EmberGenome. -Unique file naming was broken for files which already had an underscore in them. -Properly report that png is the default format of EmberRender and EmberAnimate output instead of erroneously claiming it was jpg. -Make command line programs search these folders in this order for the palette file: ./ ~/.fractorium ~/.config/fractorium /usr/share/fractorium /usr/local/share/fractorium -Fix possible bad values in hexes. -Better assignment of Z variables. -Fix boarders due to use of poorly implemented rint() function from flam3. Use std::rint() now. -wedge_sph was completely wrong due to having accidentally swapped the mapping of two parameters. -Make juliascope juliascope_power parameter be of type REAL_NONZERO since it's used as a denominator. -Make Z assignment compatible with the originals in: -arch, bcircle, bCollide, bent, bent2, bisplit, blob, blur_linear, blur_square, bMod, boarders, boarders2, bSwirl, bTransform, butterfly, cardioid, cell, circleblur, circlize, circlize2, circus, collideoscope, cos, cosine, cosh, coth, cpow, cpow2, crescents, cropn, csc, csch, curl, curve, dc_gridout, deltaa, diamond, disc2, eclipse, eCollide, edisc, eJulia, elliptic, eMod, eMotion, ennepers, epispiral, ePush, eRotate, eScale, eSwirl, ex, exp, expo, exponential, fan, fdisc, fibonacci, fibonacci2, fisheye, flipcircle, flipy, flower, flux, funnel, glynnia, GlynnSim1, GlynnSim2, GlynnSim3, gridout, handkerchief, heart, hole, idisc, julia, julian2, juliaNab, kaleidoscope, lazyTravis, Lissajous, mask, MobiusN, mobius_strip, modulus, murl, murl2, npolar, ortho, oscilloscope, parabola, perspective, petal, phoenix_julia, pie (was also inconsistent between cpu and gpu), poincare, popcorn, popcorn2, power, pow_block, rational3, rays, rblur, rings, rippled, roundspher, sec, secant2, sigmoid, sin, sineblur, sinh, sinusgrid, sphericaln, spiralwing, spirograph, split, squarize, squirrel, squish, sschecks, starblur, stripes, stwin, super_shape, tan, tancos, tangent, tanh, TwinTrian, twoface, unpolar, waves, wavesn, wedge_julia, whorl, xheart, zblur, zscale. --Code changes -Generalize Variation::PrecalcHelper() and rename to PrePostPrecalcHelper(). --Do the same for the OpenCL version and rename it PrePostPrecalcOpenCLString(). -Rename Variation::m_AssignType to m_PrePostAssignType since it's only relevant to pre/post variations.
2016-01-29 20:02:15 -05:00
helper.Out.z = DefaultZ(helper);
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
}
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
--Code changes -Better use of const functions. -Ember.pro should not define BUILDING_EMBERCL, instead define BUILDING_EMBER.
2015-06-29 23:13:53 -04:00
virtual string OpenCLString() const override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
ostringstream ss, ss2;
0.4.1.3 Beta 10/14/2014 --User Changes Size is no longer fixed to the window size. Size scaling is done differently in the final render dialog. This fixes several bugs. Remove Xml saving size from settings and options dialog, it no longer applies. Final render can be broken into strips. Set default save path to the desktop if none is found in the settings file. Set default output size to 1920x1080 if none is found in the settings file. --Bug Fixes Better memory size reporting in final render dialog. --Code Changes Migrate to C++11, Qt 5.3.1, and Visual Studio 2013. Change most instances of unsigned int to size_t, and int to intmax_t. Add m_OrigPixPerUnit and m_ScaleType to Ember for scaling purposes. Replace some sprintf_s() calls in XmlToEmber with ostringstream. Move more non-templated members into RendererBase. Add CopyVec() overload that takes a per element function pointer. Add vector Memset(). Replace '&' with '+' instead of "&amp;" in XmlToEmber for much faster parsing. Break strips rendering out into EmberCommon and call from EmberRender and Fractorium. Make AddAndWriteBuffer() just call WriteBuffer(). Make AddAndWriteImage() delete the existing image first before replacing it. Add SetOutputTexture() to RendererCL to support making new textures in response to resize events. Remove multiple return statements in RendererCL, and replace with a bool that tracks results. Add ToDouble(), MakeEnd(), ToString() and Exists() wrappers in Fractorium. Add Size() wrapper in EmberFile. Make QString function arguments const QString&, and string with const string&. Make ShowCritical() wrapper for invoking a message box from another thread. Add combo box to TwoButtonWidget and rename.
2014-10-14 11:53:15 -04:00
intmax_t i = 0, varIndex = IndexInXform();
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
ss2 << "_" << XformIndexInEmber() << "]";
string index = ss2.str();
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
string weight = WeightDefineString();
string m_k = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string step = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string num = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
string xSeed = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string ySeed = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
ss << "\t{\n"
<< "\t\tint i, j, l, k, m, m1, n, n1;\n"
Ditch T(f) define in opencl programs in preference to already-defined real_t. Also removed some erroneously-added double-'T(' statements
2015-02-26 16:18:55 -05:00
<< "\t\treal_t r, rMin, offsetX, offsetY, x0 = (real_t)(0.0), y0 = (real_t)(0.0), x, y;\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\n"
<< "\t\trMin = 20;\n"
<< "\t\tm = (int)floor(vIn.x / " << step << ");\n"
<< "\t\tn = (int)floor(vIn.y / " << step << ");\n"
<< "\n"
<< "\t\tfor (i = -1; i < 2; i++)\n"
<< "\t\t{\n"
<< "\t\t m1 = m + i;\n"
<< "\n"
<< "\t\t for (j = -1; j < 2; j++)\n"
<< "\t\t {\n"
<< "\t\t n1 = n + j;\n"
<< "\t\t k = 1 + (int)floor(" << num << " * VoronDiscreteNoise((int)(19 * m1 + 257 * n1 + " << xSeed << ")));\n"
<< "\n"
<< "\t\t for (l = 0; l < k; l++)\n"
<< "\t\t {\n"
<< "\t\t x = (real_t)(VoronDiscreteNoise((int)(l + 64 * m1 + 15 * n1 + " << xSeed << ")) + m1) * " << step << ";\n"
<< "\t\t y = (real_t)(VoronDiscreteNoise((int)(l + 21 * m1 + 33 * n1 + " << ySeed << ")) + n1) * " << step << ";\n"
<< "\t\t offsetX = vIn.x - x;\n"
<< "\t\t offsetY = vIn.y - y;\n"
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
<< "\t\t r = sqrt(fma(offsetX, offsetX, SQR(offsetY)));\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\n"
<< "\t\t if (r < rMin)\n"
<< "\t\t {\n"
<< "\t\t rMin = r;\n"
<< "\t\t x0 = x;\n"
<< "\t\t y0 = y;\n"
<< "\t\t }\n"
<< "\t\t }\n"
<< "\t\t }\n"
<< "\t\t}\n"
<< "\n"
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
<< "\t\tvOut.x = " << weight << " * fma(" << m_k << ", (vIn.x - x0), x0);\n"
<< "\t\tvOut.y = " << weight << " * fma(" << m_k << ", (vIn.y - y0), y0);\n"
--User changes -Allow for pausing the renderer in the main window. This makes is more efficient when entering many parameters, such as when following a tutorial. -Add support for new variations: erf, gamma, jac_cn, jac_dn, jac_sn, logDB, pressure_wave, pRose3D, splits3D, w, waves2b, x, xerf, y, z. -Inform user of the start and stop of file parsing in EmberAnimate because the files could potentially be very large. -Move the follwing fields to a new table called Animation: Interpolation, Affine Interpolation, Temporal Samples, Temporal Filter Width, Temporal Filter Type. -These currently have no effect on the interactive renderer and instead are used when running flames through EmberGenome to generate sequences, and then animating them in Fractorium or EmberAnimate. -Add new parameter overrides for EmberRender and EmberAnimate which directly assign values to all flames being rendered, rather than scale: --quality --demin --demax --Bug fixes -Left pad instead of right pad names of sequence outputs from EmberGenome. -Unique file naming was broken for files which already had an underscore in them. -Properly report that png is the default format of EmberRender and EmberAnimate output instead of erroneously claiming it was jpg. -Make command line programs search these folders in this order for the palette file: ./ ~/.fractorium ~/.config/fractorium /usr/share/fractorium /usr/local/share/fractorium -Fix possible bad values in hexes. -Better assignment of Z variables. -Fix boarders due to use of poorly implemented rint() function from flam3. Use std::rint() now. -wedge_sph was completely wrong due to having accidentally swapped the mapping of two parameters. -Make juliascope juliascope_power parameter be of type REAL_NONZERO since it's used as a denominator. -Make Z assignment compatible with the originals in: -arch, bcircle, bCollide, bent, bent2, bisplit, blob, blur_linear, blur_square, bMod, boarders, boarders2, bSwirl, bTransform, butterfly, cardioid, cell, circleblur, circlize, circlize2, circus, collideoscope, cos, cosine, cosh, coth, cpow, cpow2, crescents, cropn, csc, csch, curl, curve, dc_gridout, deltaa, diamond, disc2, eclipse, eCollide, edisc, eJulia, elliptic, eMod, eMotion, ennepers, epispiral, ePush, eRotate, eScale, eSwirl, ex, exp, expo, exponential, fan, fdisc, fibonacci, fibonacci2, fisheye, flipcircle, flipy, flower, flux, funnel, glynnia, GlynnSim1, GlynnSim2, GlynnSim3, gridout, handkerchief, heart, hole, idisc, julia, julian2, juliaNab, kaleidoscope, lazyTravis, Lissajous, mask, MobiusN, mobius_strip, modulus, murl, murl2, npolar, ortho, oscilloscope, parabola, perspective, petal, phoenix_julia, pie (was also inconsistent between cpu and gpu), poincare, popcorn, popcorn2, power, pow_block, rational3, rays, rblur, rings, rippled, roundspher, sec, secant2, sigmoid, sin, sineblur, sinh, sinusgrid, sphericaln, spiralwing, spirograph, split, squarize, squirrel, squish, sschecks, starblur, stripes, stwin, super_shape, tan, tancos, tangent, tanh, TwinTrian, twoface, unpolar, waves, wavesn, wedge_julia, whorl, xheart, zblur, zscale. --Code changes -Generalize Variation::PrecalcHelper() and rename to PrePostPrecalcHelper(). --Do the same for the OpenCL version and rename it PrePostPrecalcOpenCLString(). -Rename Variation::m_AssignType to m_PrePostAssignType since it's only relevant to pre/post variations.
2016-01-29 20:02:15 -05:00
<< "\t\tvOut.z = " << DefaultZCl()
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t}\n";
return ss.str();
}
--User changes -Add new variations: bubbleT3D, crob, hexaplay3D, hexcrop, hexes, hexnix3D, loonie2, loonie3, nBlur, octapol and synth. -Allow for pre/post versions of dc_bubble, dc_cylinder and dc_linear whereas before they were omitted. -When saving a file with multiple embers in it, detect if time values are all the same and if so, start them at zero and increment by 1 for each ember. -Allow for numerous quality increases to be coalesced into one. It will pick up at the end of the current render. -Show selection highlight on variations tree in response to mouse hover. This makes it easier to see for which variation or param the current mouse wheel action will apply. -Make default temporal samples be 100, whereas before it was 1000 which was overkill. -Require the shift key to be held with delete for deleting an ember to prevent it from triggering when the user enters delete in the edit box. -This wasn't otherwise fixable without writing a lot more code. --Bug fixes -EmberGenome was crashing when generating a sequence from a source file with more than 2 embers in it. -EmberGenome was improperly handling the first frame of a merge after the last frame of the loop. -These bugs were due to a previous commit. Revert parts of that commit. -Prevent a zoom value of less than 0 when reading from xml. -Slight optimization of the crescents, and mask variations, if the compiler wasn't doing it already. -Unique file naming was broken because it was looking for _# and the default names ended with -#. -Disallow renaming of an ember in the library tree to an empty string. -Severe bug that prevented some variations from being read correctly from params generated outside this program. -Severe OpenCL randomization bug. The first x coordinates of the first points in the first kernel call of the first ember of a render since the OpenCL renderer object was created were not random and were mostly -1. -Severe bug when populating xform selection distributions that could sometimes cause a crash due to roundoff error. Fix by using double. -Limit the max number of variations in a random ember to MAX_CL_VARS, which is 8. This ensures they'll look the same on CPU and GPU. -Prevent user from saving stylesheet to default.qss, it's a special reserved filename. --Code changes -Generalize using the running sum output point inside of a variation for all cases: pre, reg and post. -Allow for array variables in variations where the address of each element is stored in m_Params. -Qualify all math functions with std:: -No longer use our own Clamp() in OpenCL, instead use the standard clamp(). -Redesign how functions are used in the variations OpenCL code. -Add tests to EmberTester to verify some of the new functionality. -Place more const and override qualifiers on functions where appropriate. -Add a global rand with a lock to be used very sparingly. -Use a map instead of a vector for bad param names in Xml parsing. -Prefix affine interpolation mode defines with "AFFINE_" to make their purpose more clear. -Allow for variations that change state during iteration by sending a separate copy of the ember to each rendering thread. -Implement this same functionality with a local struct in OpenCL. It's members are the total of all variables that need to change state within an ember. -Add Contains() function to Utils.h. -EmberRender: print names of kernels being printed with --dump_kernel option. -Clean up EmberTester to handle some of the recent changes. -Fix various casts. -Replace % 2 with & 1, even though the compiler was likely doing this already. -Add new file Variations06.h to accommodate new variations. -General cleanup.
2015-11-22 17:15:07 -05:00
virtual string OpenCLFuncsString() const override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
return
"real_t VoronDiscreteNoise(int x)\n"
"{\n"
" const real_t im = 2147483647;\n"
" const real_t am = (1 / im);\n"
"\n"
" int n = x;\n"
" n = (n << 13) ^ n;\n"
" return ((n * (n * n * 15731 + 789221) + 1376312589) & 0x7fffffff) * am;\n"
"}\n"
"\n";
}
protected:
void Init()
{
string prefix = Prefix();
m_Params.clear();
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
m_Params.push_back(ParamWithName<T>(&m_K, prefix + "Voron_K", T(0.99)));
m_Params.push_back(ParamWithName<T>(&m_Step, prefix + "Voron_Step", T(0.25), eParamType::REAL_NONZERO));
m_Params.push_back(ParamWithName<T>(&m_Num, prefix + "Voron_Num", 1, eParamType::INTEGER, 1, 25));
--Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums.
2016-01-04 19:50:15 -05:00
m_Params.push_back(ParamWithName<T>(&m_XSeed, prefix + "Voron_XSeed", 3, eParamType::INTEGER));
m_Params.push_back(ParamWithName<T>(&m_YSeed, prefix + "Voron_YSeed", 7, eParamType::INTEGER));
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
}
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
private:
T DiscreteNoise(int x)
{
const T im = T(2147483647);
const T am = (1 / im);
int n = x;
n = (n << 13) ^ n;
return ((n * (n * n * 15731 + 789221) + 1376312589) & 0x7fffffff) * am;
}
T m_K;//Params.
T m_Step;
T m_Num;
T m_XSeed;
T m_YSeed;
};
/// <summary>
/// Waffle.
/// </summary>
template <typename T>
--Code changes -Split VariationList into .h/cpp files. --Make VariationList.cpp the only place where the variations files are included from. This speeds up the build, but variations can no longer be directly instantiated. --Variations are no longer exported.
2016-02-18 21:58:24 -05:00
class WaffleVariation : public ParametricVariation<T>
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
public:
--Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums.
2016-01-04 19:50:15 -05:00
WaffleVariation(T weight = 1.0) : ParametricVariation<T>("waffle", eVariationId::VAR_WAFFLE, weight)
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
Init();
}
PARVARCOPY(WaffleVariation)
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
T a = 0, r = 0;
switch (rand.Rand(5))
{
case 0:
More Linux work. Convert all casts to new style, away from legacy.
2014-12-07 02:51:44 -05:00
a = (rand.Rand(ISAAC_INT(m_Slices)) + rand.Frand01<T>() * m_XThickness) / m_Slices;
r = (rand.Rand(ISAAC_INT(m_Slices)) + rand.Frand01<T>() * m_YThickness) / m_Slices;
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
break;
--User changes -Add post_smartcrop. --Bug fixes -Fix bug in crackle. -Wrong point assignment in hexaplay3D, hexnix3D. -Improper Z assignment in rblur. -Fix inconsistency with original in circlecrop. -Put EMBER_ROOT bakc to ./../../../ in default.pri. This is TBD. --Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums. -Optimize crob, nBlur. -Fix weird assignment statement in falloff3. -Cleanup in VarFuncs::SimplexNoise3D(). -Replace fabs() with std::abs(). -General cleanup.
2016-01-12 23:42:12 -05:00
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
case 1:
More Linux work. Convert all casts to new style, away from legacy.
2014-12-07 02:51:44 -05:00
a = (rand.Rand(ISAAC_INT(m_Slices)) + rand.Frand01<T>()) / m_Slices;
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
r = (rand.Rand(ISAAC_INT(m_Slices)) + m_YThickness) / m_Slices;
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
break;
--User changes -Add post_smartcrop. --Bug fixes -Fix bug in crackle. -Wrong point assignment in hexaplay3D, hexnix3D. -Improper Z assignment in rblur. -Fix inconsistency with original in circlecrop. -Put EMBER_ROOT bakc to ./../../../ in default.pri. This is TBD. --Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums. -Optimize crob, nBlur. -Fix weird assignment statement in falloff3. -Cleanup in VarFuncs::SimplexNoise3D(). -Replace fabs() with std::abs(). -General cleanup.
2016-01-12 23:42:12 -05:00
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
case 2:
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
a = (rand.Rand(ISAAC_INT(m_Slices)) + m_XThickness) / m_Slices;
More Linux work. Convert all casts to new style, away from legacy.
2014-12-07 02:51:44 -05:00
r = (rand.Rand(ISAAC_INT(m_Slices)) + rand.Frand01<T>()) / m_Slices;
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
break;
--User changes -Add post_smartcrop. --Bug fixes -Fix bug in crackle. -Wrong point assignment in hexaplay3D, hexnix3D. -Improper Z assignment in rblur. -Fix inconsistency with original in circlecrop. -Put EMBER_ROOT bakc to ./../../../ in default.pri. This is TBD. --Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums. -Optimize crob, nBlur. -Fix weird assignment statement in falloff3. -Cleanup in VarFuncs::SimplexNoise3D(). -Replace fabs() with std::abs(). -General cleanup.
2016-01-12 23:42:12 -05:00
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
case 3:
a = rand.Frand01<T>();
More Linux work. Convert all casts to new style, away from legacy.
2014-12-07 02:51:44 -05:00
r = (rand.Rand(ISAAC_INT(m_Slices)) + m_YThickness + rand.Frand01<T>() * (1 - m_YThickness)) / m_Slices;
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
break;
--User changes -Add post_smartcrop. --Bug fixes -Fix bug in crackle. -Wrong point assignment in hexaplay3D, hexnix3D. -Improper Z assignment in rblur. -Fix inconsistency with original in circlecrop. -Put EMBER_ROOT bakc to ./../../../ in default.pri. This is TBD. --Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums. -Optimize crob, nBlur. -Fix weird assignment statement in falloff3. -Cleanup in VarFuncs::SimplexNoise3D(). -Replace fabs() with std::abs(). -General cleanup.
2016-01-12 23:42:12 -05:00
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
case 4:
More C++11
2014-09-10 01:41:26 -04:00
default:
More Linux work. Convert all casts to new style, away from legacy.
2014-12-07 02:51:44 -05:00
a = (rand.Rand(ISAAC_INT(m_Slices)) + m_XThickness + rand.Frand01<T>() * (1 - m_XThickness)) / m_Slices;
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
r = rand.Frand01<T>();
break;
}
helper.Out.x = m_CosR * a + m_SinR * r;
helper.Out.y = -m_SinR * a + m_CosR * r;
--User changes -Allow for pausing the renderer in the main window. This makes is more efficient when entering many parameters, such as when following a tutorial. -Add support for new variations: erf, gamma, jac_cn, jac_dn, jac_sn, logDB, pressure_wave, pRose3D, splits3D, w, waves2b, x, xerf, y, z. -Inform user of the start and stop of file parsing in EmberAnimate because the files could potentially be very large. -Move the follwing fields to a new table called Animation: Interpolation, Affine Interpolation, Temporal Samples, Temporal Filter Width, Temporal Filter Type. -These currently have no effect on the interactive renderer and instead are used when running flames through EmberGenome to generate sequences, and then animating them in Fractorium or EmberAnimate. -Add new parameter overrides for EmberRender and EmberAnimate which directly assign values to all flames being rendered, rather than scale: --quality --demin --demax --Bug fixes -Left pad instead of right pad names of sequence outputs from EmberGenome. -Unique file naming was broken for files which already had an underscore in them. -Properly report that png is the default format of EmberRender and EmberAnimate output instead of erroneously claiming it was jpg. -Make command line programs search these folders in this order for the palette file: ./ ~/.fractorium ~/.config/fractorium /usr/share/fractorium /usr/local/share/fractorium -Fix possible bad values in hexes. -Better assignment of Z variables. -Fix boarders due to use of poorly implemented rint() function from flam3. Use std::rint() now. -wedge_sph was completely wrong due to having accidentally swapped the mapping of two parameters. -Make juliascope juliascope_power parameter be of type REAL_NONZERO since it's used as a denominator. -Make Z assignment compatible with the originals in: -arch, bcircle, bCollide, bent, bent2, bisplit, blob, blur_linear, blur_square, bMod, boarders, boarders2, bSwirl, bTransform, butterfly, cardioid, cell, circleblur, circlize, circlize2, circus, collideoscope, cos, cosine, cosh, coth, cpow, cpow2, crescents, cropn, csc, csch, curl, curve, dc_gridout, deltaa, diamond, disc2, eclipse, eCollide, edisc, eJulia, elliptic, eMod, eMotion, ennepers, epispiral, ePush, eRotate, eScale, eSwirl, ex, exp, expo, exponential, fan, fdisc, fibonacci, fibonacci2, fisheye, flipcircle, flipy, flower, flux, funnel, glynnia, GlynnSim1, GlynnSim2, GlynnSim3, gridout, handkerchief, heart, hole, idisc, julia, julian2, juliaNab, kaleidoscope, lazyTravis, Lissajous, mask, MobiusN, mobius_strip, modulus, murl, murl2, npolar, ortho, oscilloscope, parabola, perspective, petal, phoenix_julia, pie (was also inconsistent between cpu and gpu), poincare, popcorn, popcorn2, power, pow_block, rational3, rays, rblur, rings, rippled, roundspher, sec, secant2, sigmoid, sin, sineblur, sinh, sinusgrid, sphericaln, spiralwing, spirograph, split, squarize, squirrel, squish, sschecks, starblur, stripes, stwin, super_shape, tan, tancos, tangent, tanh, TwinTrian, twoface, unpolar, waves, wavesn, wedge_julia, whorl, xheart, zblur, zscale. --Code changes -Generalize Variation::PrecalcHelper() and rename to PrePostPrecalcHelper(). --Do the same for the OpenCL version and rename it PrePostPrecalcOpenCLString(). -Rename Variation::m_AssignType to m_PrePostAssignType since it's only relevant to pre/post variations.
2016-01-29 20:02:15 -05:00
helper.Out.z = DefaultZ(helper);
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
}
--Code changes -Better use of const functions. -Ember.pro should not define BUILDING_EMBERCL, instead define BUILDING_EMBER.
2015-06-29 23:13:53 -04:00
virtual string OpenCLString() const override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
ostringstream ss, ss2;
More C++11
2014-09-10 01:41:26 -04:00
int i = 0;
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
ss2 << "_" << XformIndexInEmber() << "]";
string index = ss2.str();
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
string weight = WeightDefineString();
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
string slices = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
string xThickness = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string yThickness = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
string rotation = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string sinr = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string cosr = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
ss << "\t{\n"
<< "\t\treal_t a = 0, r = 0;\n"
<< "\n"
<< "\t\tswitch (MwcNextRange(mwc, 5))\n"
<< "\t\t{\n"
<< "\t\t case 0:\n"
<< "\t\t a = (MwcNextRange(mwc, (int)" << slices << ") + MwcNext01(mwc) * " << xThickness << ") / " << slices << ";\n"
<< "\t\t r = (MwcNextRange(mwc, (int)" << slices << ") + MwcNext01(mwc) * " << yThickness << ") / " << slices << ";\n"
<< "\t\t break;\n"
<< "\t\t case 1:\n"
<< "\t\t a = (MwcNextRange(mwc, (int)" << slices << ") + MwcNext01(mwc)) / " << slices << ";\n"
<< "\t\t r = (MwcNextRange(mwc, (int)" << slices << ") + " << yThickness << ") / " << slices << ";\n"
<< "\t\t break;\n"
<< "\t\t case 2:\n"
<< "\t\t a = (MwcNextRange(mwc, (int)" << slices << ") + " << xThickness << ") / " << slices << ";\n"
<< "\t\t r = (MwcNextRange(mwc, (int)" << slices << ") + MwcNext01(mwc)) / " << slices << ";\n"
<< "\t\t break;\n"
<< "\t\t case 3:\n"
<< "\t\t a = MwcNext01(mwc);\n"
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
<< "\t\t r = fma(MwcNext01(mwc), 1 - " << yThickness << ", MwcNextRange(mwc, (int)" << slices << ") + " << yThickness << ") / " << slices << ";\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t\t break;\n"
<< "\t\t case 4:\n"
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
<< "\t\t a = fma(MwcNext01(mwc), (1 - " << xThickness << "), MwcNextRange(mwc, (int)" << slices << ") + " << xThickness << ") / " << slices << ";\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t\t r = MwcNext01(mwc);\n"
<< "\t\t break;\n"
<< "\t\t}\n"
<< "\n"
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
<< "\t\tvOut.x = fma(" << cosr << ", a, " << sinr << " * r);\n"
<< "\t\tvOut.y = -fma(" << sinr << ", a, " << cosr << " * r);\n"
--User changes -Allow for pausing the renderer in the main window. This makes is more efficient when entering many parameters, such as when following a tutorial. -Add support for new variations: erf, gamma, jac_cn, jac_dn, jac_sn, logDB, pressure_wave, pRose3D, splits3D, w, waves2b, x, xerf, y, z. -Inform user of the start and stop of file parsing in EmberAnimate because the files could potentially be very large. -Move the follwing fields to a new table called Animation: Interpolation, Affine Interpolation, Temporal Samples, Temporal Filter Width, Temporal Filter Type. -These currently have no effect on the interactive renderer and instead are used when running flames through EmberGenome to generate sequences, and then animating them in Fractorium or EmberAnimate. -Add new parameter overrides for EmberRender and EmberAnimate which directly assign values to all flames being rendered, rather than scale: --quality --demin --demax --Bug fixes -Left pad instead of right pad names of sequence outputs from EmberGenome. -Unique file naming was broken for files which already had an underscore in them. -Properly report that png is the default format of EmberRender and EmberAnimate output instead of erroneously claiming it was jpg. -Make command line programs search these folders in this order for the palette file: ./ ~/.fractorium ~/.config/fractorium /usr/share/fractorium /usr/local/share/fractorium -Fix possible bad values in hexes. -Better assignment of Z variables. -Fix boarders due to use of poorly implemented rint() function from flam3. Use std::rint() now. -wedge_sph was completely wrong due to having accidentally swapped the mapping of two parameters. -Make juliascope juliascope_power parameter be of type REAL_NONZERO since it's used as a denominator. -Make Z assignment compatible with the originals in: -arch, bcircle, bCollide, bent, bent2, bisplit, blob, blur_linear, blur_square, bMod, boarders, boarders2, bSwirl, bTransform, butterfly, cardioid, cell, circleblur, circlize, circlize2, circus, collideoscope, cos, cosine, cosh, coth, cpow, cpow2, crescents, cropn, csc, csch, curl, curve, dc_gridout, deltaa, diamond, disc2, eclipse, eCollide, edisc, eJulia, elliptic, eMod, eMotion, ennepers, epispiral, ePush, eRotate, eScale, eSwirl, ex, exp, expo, exponential, fan, fdisc, fibonacci, fibonacci2, fisheye, flipcircle, flipy, flower, flux, funnel, glynnia, GlynnSim1, GlynnSim2, GlynnSim3, gridout, handkerchief, heart, hole, idisc, julia, julian2, juliaNab, kaleidoscope, lazyTravis, Lissajous, mask, MobiusN, mobius_strip, modulus, murl, murl2, npolar, ortho, oscilloscope, parabola, perspective, petal, phoenix_julia, pie (was also inconsistent between cpu and gpu), poincare, popcorn, popcorn2, power, pow_block, rational3, rays, rblur, rings, rippled, roundspher, sec, secant2, sigmoid, sin, sineblur, sinh, sinusgrid, sphericaln, spiralwing, spirograph, split, squarize, squirrel, squish, sschecks, starblur, stripes, stwin, super_shape, tan, tancos, tangent, tanh, TwinTrian, twoface, unpolar, waves, wavesn, wedge_julia, whorl, xheart, zblur, zscale. --Code changes -Generalize Variation::PrecalcHelper() and rename to PrePostPrecalcHelper(). --Do the same for the OpenCL version and rename it PrePostPrecalcOpenCLString(). -Rename Variation::m_AssignType to m_PrePostAssignType since it's only relevant to pre/post variations.
2016-01-29 20:02:15 -05:00
<< "\t\tvOut.z = " << DefaultZCl()
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t}\n";
return ss.str();
}
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
virtual void Precalc() override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
--User changes -Add new variations: bubbleT3D, crob, hexaplay3D, hexcrop, hexes, hexnix3D, loonie2, loonie3, nBlur, octapol and synth. -Allow for pre/post versions of dc_bubble, dc_cylinder and dc_linear whereas before they were omitted. -When saving a file with multiple embers in it, detect if time values are all the same and if so, start them at zero and increment by 1 for each ember. -Allow for numerous quality increases to be coalesced into one. It will pick up at the end of the current render. -Show selection highlight on variations tree in response to mouse hover. This makes it easier to see for which variation or param the current mouse wheel action will apply. -Make default temporal samples be 100, whereas before it was 1000 which was overkill. -Require the shift key to be held with delete for deleting an ember to prevent it from triggering when the user enters delete in the edit box. -This wasn't otherwise fixable without writing a lot more code. --Bug fixes -EmberGenome was crashing when generating a sequence from a source file with more than 2 embers in it. -EmberGenome was improperly handling the first frame of a merge after the last frame of the loop. -These bugs were due to a previous commit. Revert parts of that commit. -Prevent a zoom value of less than 0 when reading from xml. -Slight optimization of the crescents, and mask variations, if the compiler wasn't doing it already. -Unique file naming was broken because it was looking for _# and the default names ended with -#. -Disallow renaming of an ember in the library tree to an empty string. -Severe bug that prevented some variations from being read correctly from params generated outside this program. -Severe OpenCL randomization bug. The first x coordinates of the first points in the first kernel call of the first ember of a render since the OpenCL renderer object was created were not random and were mostly -1. -Severe bug when populating xform selection distributions that could sometimes cause a crash due to roundoff error. Fix by using double. -Limit the max number of variations in a random ember to MAX_CL_VARS, which is 8. This ensures they'll look the same on CPU and GPU. -Prevent user from saving stylesheet to default.qss, it's a special reserved filename. --Code changes -Generalize using the running sum output point inside of a variation for all cases: pre, reg and post. -Allow for array variables in variations where the address of each element is stored in m_Params. -Qualify all math functions with std:: -No longer use our own Clamp() in OpenCL, instead use the standard clamp(). -Redesign how functions are used in the variations OpenCL code. -Add tests to EmberTester to verify some of the new functionality. -Place more const and override qualifiers on functions where appropriate. -Add a global rand with a lock to be used very sparingly. -Use a map instead of a vector for bad param names in Xml parsing. -Prefix affine interpolation mode defines with "AFFINE_" to make their purpose more clear. -Allow for variations that change state during iteration by sending a separate copy of the ember to each rendering thread. -Implement this same functionality with a local struct in OpenCL. It's members are the total of all variables that need to change state within an ember. -Add Contains() function to Utils.h. -EmberRender: print names of kernels being printed with --dump_kernel option. -Clean up EmberTester to handle some of the recent changes. -Fix various casts. -Replace % 2 with & 1, even though the compiler was likely doing this already. -Add new file Variations06.h to accommodate new variations. -General cleanup.
2015-11-22 17:15:07 -05:00
m_SinR = std::sin(m_Rotation);
m_CosR = std::cos(m_Rotation);
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
}
protected:
void Init()
{
string prefix = Prefix();
m_Params.clear();
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
m_Params.push_back(ParamWithName<T>(&m_Slices, prefix + "waffle_slices", 6, eParamType::INTEGER_NONZERO));
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
m_Params.push_back(ParamWithName<T>(&m_XThickness, prefix + "waffle_xthickness", T(0.5)));
m_Params.push_back(ParamWithName<T>(&m_YThickness, prefix + "waffle_ythickness", T(0.5)));
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
m_Params.push_back(ParamWithName<T>(&m_Rotation, prefix + "waffle_rotation"));
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
m_Params.push_back(ParamWithName<T>(true, &m_SinR, prefix + "waffle_sinr"));
m_Params.push_back(ParamWithName<T>(true, &m_CosR, prefix + "waffle_cosr"));
}
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
private:
T m_Slices;
T m_XThickness;
T m_YThickness;
T m_Rotation;
T m_SinR;//Precalc.
T m_CosR;
};
/// <summary>
/// Square3D.
/// </summary>
template <typename T>
--Code changes -Split VariationList into .h/cpp files. --Make VariationList.cpp the only place where the variations files are included from. This speeds up the build, but variations can no longer be directly instantiated. --Variations are no longer exported.
2016-02-18 21:58:24 -05:00
class Square3DVariation : public Variation<T>
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
public:
--Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums.
2016-01-04 19:50:15 -05:00
Square3DVariation(T weight = 1.0) : Variation<T>("square3D", eVariationId::VAR_SQUARE3D, weight) { }
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
VARCOPY(Square3DVariation)
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
helper.Out.x = m_Weight * (rand.Frand01<T>() - T(0.5));
helper.Out.y = m_Weight * (rand.Frand01<T>() - T(0.5));
helper.Out.z = m_Weight * (rand.Frand01<T>() - T(0.5));
}
--Code changes -Better use of const functions. -Ember.pro should not define BUILDING_EMBERCL, instead define BUILDING_EMBER.
2015-06-29 23:13:53 -04:00
virtual string OpenCLString() const override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
ostringstream ss;
0.4.1.3 Beta 10/14/2014 --User Changes Size is no longer fixed to the window size. Size scaling is done differently in the final render dialog. This fixes several bugs. Remove Xml saving size from settings and options dialog, it no longer applies. Final render can be broken into strips. Set default save path to the desktop if none is found in the settings file. Set default output size to 1920x1080 if none is found in the settings file. --Bug Fixes Better memory size reporting in final render dialog. --Code Changes Migrate to C++11, Qt 5.3.1, and Visual Studio 2013. Change most instances of unsigned int to size_t, and int to intmax_t. Add m_OrigPixPerUnit and m_ScaleType to Ember for scaling purposes. Replace some sprintf_s() calls in XmlToEmber with ostringstream. Move more non-templated members into RendererBase. Add CopyVec() overload that takes a per element function pointer. Add vector Memset(). Replace '&' with '+' instead of "&amp;" in XmlToEmber for much faster parsing. Break strips rendering out into EmberCommon and call from EmberRender and Fractorium. Make AddAndWriteBuffer() just call WriteBuffer(). Make AddAndWriteImage() delete the existing image first before replacing it. Add SetOutputTexture() to RendererCL to support making new textures in response to resize events. Remove multiple return statements in RendererCL, and replace with a bool that tracks results. Add ToDouble(), MakeEnd(), ToString() and Exists() wrappers in Fractorium. Add Size() wrapper in EmberFile. Make QString function arguments const QString&, and string with const string&. Make ShowCritical() wrapper for invoking a message box from another thread. Add combo box to TwoButtonWidget and rename.
2014-10-14 11:53:15 -04:00
intmax_t varIndex = IndexInXform();
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
string weight = WeightDefineString();
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
ss << "\t{\n"
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
<< "\t\tvOut.x = " << weight << " * (MwcNext01(mwc) - (real_t)(0.5));\n"
<< "\t\tvOut.y = " << weight << " * (MwcNext01(mwc) - (real_t)(0.5));\n"
<< "\t\tvOut.z = " << weight << " * (MwcNext01(mwc) - (real_t)(0.5));\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t}\n";
return ss.str();
}
};
/// <summary>
/// SuperShape3D.
/// </summary>
template <typename T>
--Code changes -Split VariationList into .h/cpp files. --Make VariationList.cpp the only place where the variations files are included from. This speeds up the build, but variations can no longer be directly instantiated. --Variations are no longer exported.
2016-02-18 21:58:24 -05:00
class SuperShape3DVariation : public ParametricVariation<T>
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
public:
--Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums.
2016-01-04 19:50:15 -05:00
SuperShape3DVariation(T weight = 1.0) : ParametricVariation<T>("SuperShape3D", eVariationId::VAR_SUPER_SHAPE3D, weight)
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
Init();
}
PARVARCOPY(SuperShape3DVariation)
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
T pr1, r1, pr2, r2, rho1, phi1, sinr, sinp, cosr, cosp, msinr, msinp, mcosr, mcosp, temp;
rho1 = rand.Frand01<T>() * m_Rho2Pi;
phi1 = rand.Frand01<T>() * m_Phi2Pi;
if (rand.RandBit())
phi1 = -phi1;
--User changes -Add new variations: bubbleT3D, crob, hexaplay3D, hexcrop, hexes, hexnix3D, loonie2, loonie3, nBlur, octapol and synth. -Allow for pre/post versions of dc_bubble, dc_cylinder and dc_linear whereas before they were omitted. -When saving a file with multiple embers in it, detect if time values are all the same and if so, start them at zero and increment by 1 for each ember. -Allow for numerous quality increases to be coalesced into one. It will pick up at the end of the current render. -Show selection highlight on variations tree in response to mouse hover. This makes it easier to see for which variation or param the current mouse wheel action will apply. -Make default temporal samples be 100, whereas before it was 1000 which was overkill. -Require the shift key to be held with delete for deleting an ember to prevent it from triggering when the user enters delete in the edit box. -This wasn't otherwise fixable without writing a lot more code. --Bug fixes -EmberGenome was crashing when generating a sequence from a source file with more than 2 embers in it. -EmberGenome was improperly handling the first frame of a merge after the last frame of the loop. -These bugs were due to a previous commit. Revert parts of that commit. -Prevent a zoom value of less than 0 when reading from xml. -Slight optimization of the crescents, and mask variations, if the compiler wasn't doing it already. -Unique file naming was broken because it was looking for _# and the default names ended with -#. -Disallow renaming of an ember in the library tree to an empty string. -Severe bug that prevented some variations from being read correctly from params generated outside this program. -Severe OpenCL randomization bug. The first x coordinates of the first points in the first kernel call of the first ember of a render since the OpenCL renderer object was created were not random and were mostly -1. -Severe bug when populating xform selection distributions that could sometimes cause a crash due to roundoff error. Fix by using double. -Limit the max number of variations in a random ember to MAX_CL_VARS, which is 8. This ensures they'll look the same on CPU and GPU. -Prevent user from saving stylesheet to default.qss, it's a special reserved filename. --Code changes -Generalize using the running sum output point inside of a variation for all cases: pre, reg and post. -Allow for array variables in variations where the address of each element is stored in m_Params. -Qualify all math functions with std:: -No longer use our own Clamp() in OpenCL, instead use the standard clamp(). -Redesign how functions are used in the variations OpenCL code. -Add tests to EmberTester to verify some of the new functionality. -Place more const and override qualifiers on functions where appropriate. -Add a global rand with a lock to be used very sparingly. -Use a map instead of a vector for bad param names in Xml parsing. -Prefix affine interpolation mode defines with "AFFINE_" to make their purpose more clear. -Allow for variations that change state during iteration by sending a separate copy of the ember to each rendering thread. -Implement this same functionality with a local struct in OpenCL. It's members are the total of all variables that need to change state within an ember. -Add Contains() function to Utils.h. -EmberRender: print names of kernels being printed with --dump_kernel option. -Clean up EmberTester to handle some of the recent changes. -Fix various casts. -Replace % 2 with & 1, even though the compiler was likely doing this already. -Add new file Variations06.h to accommodate new variations. -General cleanup.
2015-11-22 17:15:07 -05:00
sinr = std::sin(rho1);
cosr = std::cos(rho1);
sinp = std::sin(phi1);
cosp = std::cos(phi1);
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
temp = m_M4_1 * rho1;
--User changes -Add new variations: bubbleT3D, crob, hexaplay3D, hexcrop, hexes, hexnix3D, loonie2, loonie3, nBlur, octapol and synth. -Allow for pre/post versions of dc_bubble, dc_cylinder and dc_linear whereas before they were omitted. -When saving a file with multiple embers in it, detect if time values are all the same and if so, start them at zero and increment by 1 for each ember. -Allow for numerous quality increases to be coalesced into one. It will pick up at the end of the current render. -Show selection highlight on variations tree in response to mouse hover. This makes it easier to see for which variation or param the current mouse wheel action will apply. -Make default temporal samples be 100, whereas before it was 1000 which was overkill. -Require the shift key to be held with delete for deleting an ember to prevent it from triggering when the user enters delete in the edit box. -This wasn't otherwise fixable without writing a lot more code. --Bug fixes -EmberGenome was crashing when generating a sequence from a source file with more than 2 embers in it. -EmberGenome was improperly handling the first frame of a merge after the last frame of the loop. -These bugs were due to a previous commit. Revert parts of that commit. -Prevent a zoom value of less than 0 when reading from xml. -Slight optimization of the crescents, and mask variations, if the compiler wasn't doing it already. -Unique file naming was broken because it was looking for _# and the default names ended with -#. -Disallow renaming of an ember in the library tree to an empty string. -Severe bug that prevented some variations from being read correctly from params generated outside this program. -Severe OpenCL randomization bug. The first x coordinates of the first points in the first kernel call of the first ember of a render since the OpenCL renderer object was created were not random and were mostly -1. -Severe bug when populating xform selection distributions that could sometimes cause a crash due to roundoff error. Fix by using double. -Limit the max number of variations in a random ember to MAX_CL_VARS, which is 8. This ensures they'll look the same on CPU and GPU. -Prevent user from saving stylesheet to default.qss, it's a special reserved filename. --Code changes -Generalize using the running sum output point inside of a variation for all cases: pre, reg and post. -Allow for array variables in variations where the address of each element is stored in m_Params. -Qualify all math functions with std:: -No longer use our own Clamp() in OpenCL, instead use the standard clamp(). -Redesign how functions are used in the variations OpenCL code. -Add tests to EmberTester to verify some of the new functionality. -Place more const and override qualifiers on functions where appropriate. -Add a global rand with a lock to be used very sparingly. -Use a map instead of a vector for bad param names in Xml parsing. -Prefix affine interpolation mode defines with "AFFINE_" to make their purpose more clear. -Allow for variations that change state during iteration by sending a separate copy of the ember to each rendering thread. -Implement this same functionality with a local struct in OpenCL. It's members are the total of all variables that need to change state within an ember. -Add Contains() function to Utils.h. -EmberRender: print names of kernels being printed with --dump_kernel option. -Clean up EmberTester to handle some of the recent changes. -Fix various casts. -Replace % 2 with & 1, even though the compiler was likely doing this already. -Add new file Variations06.h to accommodate new variations. -General cleanup.
2015-11-22 17:15:07 -05:00
msinr = std::sin(temp);
mcosr = std::cos(temp);
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
temp = m_M4_2 * phi1;
--User changes -Add new variations: bubbleT3D, crob, hexaplay3D, hexcrop, hexes, hexnix3D, loonie2, loonie3, nBlur, octapol and synth. -Allow for pre/post versions of dc_bubble, dc_cylinder and dc_linear whereas before they were omitted. -When saving a file with multiple embers in it, detect if time values are all the same and if so, start them at zero and increment by 1 for each ember. -Allow for numerous quality increases to be coalesced into one. It will pick up at the end of the current render. -Show selection highlight on variations tree in response to mouse hover. This makes it easier to see for which variation or param the current mouse wheel action will apply. -Make default temporal samples be 100, whereas before it was 1000 which was overkill. -Require the shift key to be held with delete for deleting an ember to prevent it from triggering when the user enters delete in the edit box. -This wasn't otherwise fixable without writing a lot more code. --Bug fixes -EmberGenome was crashing when generating a sequence from a source file with more than 2 embers in it. -EmberGenome was improperly handling the first frame of a merge after the last frame of the loop. -These bugs were due to a previous commit. Revert parts of that commit. -Prevent a zoom value of less than 0 when reading from xml. -Slight optimization of the crescents, and mask variations, if the compiler wasn't doing it already. -Unique file naming was broken because it was looking for _# and the default names ended with -#. -Disallow renaming of an ember in the library tree to an empty string. -Severe bug that prevented some variations from being read correctly from params generated outside this program. -Severe OpenCL randomization bug. The first x coordinates of the first points in the first kernel call of the first ember of a render since the OpenCL renderer object was created were not random and were mostly -1. -Severe bug when populating xform selection distributions that could sometimes cause a crash due to roundoff error. Fix by using double. -Limit the max number of variations in a random ember to MAX_CL_VARS, which is 8. This ensures they'll look the same on CPU and GPU. -Prevent user from saving stylesheet to default.qss, it's a special reserved filename. --Code changes -Generalize using the running sum output point inside of a variation for all cases: pre, reg and post. -Allow for array variables in variations where the address of each element is stored in m_Params. -Qualify all math functions with std:: -No longer use our own Clamp() in OpenCL, instead use the standard clamp(). -Redesign how functions are used in the variations OpenCL code. -Add tests to EmberTester to verify some of the new functionality. -Place more const and override qualifiers on functions where appropriate. -Add a global rand with a lock to be used very sparingly. -Use a map instead of a vector for bad param names in Xml parsing. -Prefix affine interpolation mode defines with "AFFINE_" to make their purpose more clear. -Allow for variations that change state during iteration by sending a separate copy of the ember to each rendering thread. -Implement this same functionality with a local struct in OpenCL. It's members are the total of all variables that need to change state within an ember. -Add Contains() function to Utils.h. -EmberRender: print names of kernels being printed with --dump_kernel option. -Clean up EmberTester to handle some of the recent changes. -Fix various casts. -Replace % 2 with & 1, even though the compiler was likely doing this already. -Add new file Variations06.h to accommodate new variations. -General cleanup.
2015-11-22 17:15:07 -05:00
msinp = std::sin(temp);
mcosp = std::cos(temp);
--User changes -Add post_smartcrop. --Bug fixes -Fix bug in crackle. -Wrong point assignment in hexaplay3D, hexnix3D. -Improper Z assignment in rblur. -Fix inconsistency with original in circlecrop. -Put EMBER_ROOT bakc to ./../../../ in default.pri. This is TBD. --Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums. -Optimize crob, nBlur. -Fix weird assignment statement in falloff3. -Cleanup in VarFuncs::SimplexNoise3D(). -Replace fabs() with std::abs(). -General cleanup.
2016-01-12 23:42:12 -05:00
pr1 = m_An2_1 * std::pow(std::abs(mcosr), m_N2_1) + m_Bn3_1 * std::pow(std::abs(msinr), m_N3_1);
pr2 = m_An2_2 * std::pow(std::abs(mcosp), m_N2_2) + m_Bn3_2 * std::pow(std::abs(msinp), m_N3_2);
r1 = std::pow(std::abs(pr1), m_N1_1) + m_Spiral * rho1;
r2 = std::pow(std::abs(pr2), m_N1_2);
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
More Linux work. Convert all casts to new style, away from legacy.
2014-12-07 02:51:44 -05:00
if (int(m_Toroidmap) == 1)
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
helper.Out.x = m_Weight * cosr * (r1 + r2 * cosp);
helper.Out.y = m_Weight * sinr * (r1 + r2 * cosp);
helper.Out.z = m_Weight * r2 * sinp;
}
else
{
helper.Out.x = m_Weight * r1 * cosr * r2 * cosp;
helper.Out.y = m_Weight * r1 * sinr * r2 * cosp;
helper.Out.z = m_Weight * r2 * sinp;
}
}
--Code changes -Better use of const functions. -Ember.pro should not define BUILDING_EMBERCL, instead define BUILDING_EMBER.
2015-06-29 23:13:53 -04:00
virtual string OpenCLString() const override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
ostringstream ss, ss2;
0.4.1.3 Beta 10/14/2014 --User Changes Size is no longer fixed to the window size. Size scaling is done differently in the final render dialog. This fixes several bugs. Remove Xml saving size from settings and options dialog, it no longer applies. Final render can be broken into strips. Set default save path to the desktop if none is found in the settings file. Set default output size to 1920x1080 if none is found in the settings file. --Bug Fixes Better memory size reporting in final render dialog. --Code Changes Migrate to C++11, Qt 5.3.1, and Visual Studio 2013. Change most instances of unsigned int to size_t, and int to intmax_t. Add m_OrigPixPerUnit and m_ScaleType to Ember for scaling purposes. Replace some sprintf_s() calls in XmlToEmber with ostringstream. Move more non-templated members into RendererBase. Add CopyVec() overload that takes a per element function pointer. Add vector Memset(). Replace '&' with '+' instead of "&amp;" in XmlToEmber for much faster parsing. Break strips rendering out into EmberCommon and call from EmberRender and Fractorium. Make AddAndWriteBuffer() just call WriteBuffer(). Make AddAndWriteImage() delete the existing image first before replacing it. Add SetOutputTexture() to RendererCL to support making new textures in response to resize events. Remove multiple return statements in RendererCL, and replace with a bool that tracks results. Add ToDouble(), MakeEnd(), ToString() and Exists() wrappers in Fractorium. Add Size() wrapper in EmberFile. Make QString function arguments const QString&, and string with const string&. Make ShowCritical() wrapper for invoking a message box from another thread. Add combo box to TwoButtonWidget and rename.
2014-10-14 11:53:15 -04:00
intmax_t i = 0, varIndex = IndexInXform();
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
ss2 << "_" << XformIndexInEmber() << "]";
string index = ss2.str();
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
string weight = WeightDefineString();
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
string rho = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string phi = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string m1 = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string m2 = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string a1 = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string a2 = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string b1 = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string b2 = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string n1_1 = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string n1_2 = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string n2_1 = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string n2_2 = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string n3_1 = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string n3_2 = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
string spiral = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string toroid = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
string n1n_1 = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string n1n_2 = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string an2_1 = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string an2_2 = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string bn3_1 = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string bn3_2 = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string m4_1 = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string m4_2 = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
string rho2pi = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string phi2pi = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
ss << "\t{\n"
<< "\t\treal_t pr1, r1, pr2, r2, rho1, phi1, sinr, sinp, cosr, cosp, msinr, msinp, mcosr, mcosp, temp;\n"
<< "\n"
<< "\t\trho1 = MwcNext01(mwc) * " << rho2pi << ";\n"
--User changes -Add post_smartcrop. --Bug fixes -Fix bug in crackle. -Wrong point assignment in hexaplay3D, hexnix3D. -Improper Z assignment in rblur. -Fix inconsistency with original in circlecrop. -Put EMBER_ROOT bakc to ./../../../ in default.pri. This is TBD. --Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums. -Optimize crob, nBlur. -Fix weird assignment statement in falloff3. -Cleanup in VarFuncs::SimplexNoise3D(). -Replace fabs() with std::abs(). -General cleanup.
2016-01-12 23:42:12 -05:00
<< "\t\tphi1 = MwcNext01(mwc) * " << phi2pi << ";\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\n"
<< "\t\tif (MwcNext(mwc) & 1)\n"
<< "\t\t phi1 = -phi1;\n"
<< "\n"
<< "\t\tsinr = sin(rho1);\n"
<< "\t\tcosr = cos(rho1);\n"
<< "\n"
<< "\t\tsinp = sin(phi1);\n"
<< "\t\tcosp = cos(phi1);\n"
<< "\n"
--User changes -Add post_smartcrop. --Bug fixes -Fix bug in crackle. -Wrong point assignment in hexaplay3D, hexnix3D. -Improper Z assignment in rblur. -Fix inconsistency with original in circlecrop. -Put EMBER_ROOT bakc to ./../../../ in default.pri. This is TBD. --Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums. -Optimize crob, nBlur. -Fix weird assignment statement in falloff3. -Cleanup in VarFuncs::SimplexNoise3D(). -Replace fabs() with std::abs(). -General cleanup.
2016-01-12 23:42:12 -05:00
<< "\t\ttemp = " << m4_1 << " * rho1;\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t\tmsinr = sin(temp);\n"
<< "\t\tmcosr = cos(temp);\n"
<< "\n"
<< "\t\ttemp = " << m4_2 << " * phi1;\n"
<< "\t\tmsinp = sin(temp);\n"
<< "\t\tmcosp = cos(temp);\n"
<< "\n"
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
<< "\t\tpr1 = fma(" << an2_1 << ", pow(fabs(mcosr), " << n2_1 << "), " << bn3_1 << " * pow(fabs(msinr), " << n3_1 << "));\n"
<< "\t\tpr2 = fma(" << an2_2 << ", pow(fabs(mcosp), " << n2_2 << "), " << bn3_2 << " * pow(fabs(msinp), " << n3_2 << "));\n"
<< "\t\tr1 = fma(" << spiral << ", rho1, pow(fabs(pr1), " << n1_1 << "));\n"
0.4.0.9 Beta 07/27/2014 0.4.0.9 Beta 07/27/2014 --User Changes Properly set tab order on all controls. Calculate and report iters/second in the final render dialog. Immediately draw yellow dot on xform mouse down on previously unselected xform. --Bug Fixes Fix GlynnSim1, GlynnSim2, GlynnSim3 and juliaNab by ensuring the first argument to pow() is >= 0. Ensure OpenCL platform and device combo boxes in the final render dialog expand as needed. --Code Changes Make VariationTreeSpinbox take its parent VariationTreeWidgetItem as a constructor argument. This makes SetupVariationTree() and VariationSpinBoxValueChanged() more efficient. Make Interference2 and ho use fabs().
2014-07-28 01:25:38 -04:00
<< "\t\tr2 = pow(fabs(pr2), " << n1_2 << ");\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\n"
<< "\t\tif ((int)" << toroid << " == 1)\n"
<< "\t\t{\n"
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
<< "\t\t vOut.x = " << weight << " * cosr * fma(r2, cosp, r1);\n"
<< "\t\t vOut.y = " << weight << " * sinr * fma(r2, cosp, r1);\n"
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
<< "\t\t vOut.z = " << weight << " * r2 * sinp;\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t\t}\n"
<< "\t\telse\n"
<< "\t\t{\n"
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
<< "\t\t vOut.x = " << weight << " * r1 * cosr * r2 * cosp;\n"
<< "\t\t vOut.y = " << weight << " * r1 * sinr * r2 * cosp;\n"
<< "\t\t vOut.z = " << weight << " * r2 * sinp;\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t\t}\n"
<< "\t}\n";
return ss.str();
}
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
virtual void Precalc() override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
m_N1n_1 = (-1 / m_N1_1);
m_N1n_2 = (-1 / m_N1_2);
--User changes -Add post_smartcrop. --Bug fixes -Fix bug in crackle. -Wrong point assignment in hexaplay3D, hexnix3D. -Improper Z assignment in rblur. -Fix inconsistency with original in circlecrop. -Put EMBER_ROOT bakc to ./../../../ in default.pri. This is TBD. --Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums. -Optimize crob, nBlur. -Fix weird assignment statement in falloff3. -Cleanup in VarFuncs::SimplexNoise3D(). -Replace fabs() with std::abs(). -General cleanup.
2016-01-12 23:42:12 -05:00
m_An2_1 = std::pow(std::abs(1 / m_A1), m_N2_1);
m_An2_2 = std::pow(std::abs(1 / m_A2), m_N2_2);
m_Bn3_1 = std::pow(std::abs(1 / m_B1), m_N3_1);
m_Bn3_2 = std::pow(std::abs(1 / m_B2), m_N3_2);
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
m_M4_1 = m_M1 / 4;
m_M4_2 = m_M2 / 4;
m_Rho2Pi = m_Rho * T(M_2_PI);
m_Phi2Pi = m_Phi * T(M_2_PI);
}
protected:
void Init()
{
string prefix = Prefix();
m_Params.clear();
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
m_Params.push_back(ParamWithName<T>(&m_Rho, prefix + "SuperShape3D_rho", T(9.9)));
m_Params.push_back(ParamWithName<T>(&m_Phi, prefix + "SuperShape3D_phi", T(2.5)));
m_Params.push_back(ParamWithName<T>(&m_M1, prefix + "SuperShape3D_m1", 6));
m_Params.push_back(ParamWithName<T>(&m_M2, prefix + "SuperShape3D_m2", 3));
m_Params.push_back(ParamWithName<T>(&m_A1, prefix + "SuperShape3D_a1", 1));
m_Params.push_back(ParamWithName<T>(&m_A2, prefix + "SuperShape3D_a2", 1));
m_Params.push_back(ParamWithName<T>(&m_B1, prefix + "SuperShape3D_b1", 1));
m_Params.push_back(ParamWithName<T>(&m_B2, prefix + "SuperShape3D_b2", 1));
m_Params.push_back(ParamWithName<T>(&m_N1_1, prefix + "SuperShape3D_n1_1", 1));
m_Params.push_back(ParamWithName<T>(&m_N1_2, prefix + "SuperShape3D_n1_2", 1));
m_Params.push_back(ParamWithName<T>(&m_N2_1, prefix + "SuperShape3D_n2_1", 1));
m_Params.push_back(ParamWithName<T>(&m_N2_2, prefix + "SuperShape3D_n2_2", 1));
m_Params.push_back(ParamWithName<T>(&m_N3_1, prefix + "SuperShape3D_n3_1", 1));
m_Params.push_back(ParamWithName<T>(&m_N3_2, prefix + "SuperShape3D_n3_2", 1));
m_Params.push_back(ParamWithName<T>(&m_Spiral, prefix + "SuperShape3D_spiral"));
--Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums.
2016-01-04 19:50:15 -05:00
m_Params.push_back(ParamWithName<T>(&m_Toroidmap, prefix + "SuperShape3D_toroidmap", 0, eParamType::INTEGER, 0, 1));
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
m_Params.push_back(ParamWithName<T>(true, &m_N1n_1, prefix + "SuperShape3D_n1n1"));//Precalc.
m_Params.push_back(ParamWithName<T>(true, &m_N1n_2, prefix + "SuperShape3D_n1n2"));
m_Params.push_back(ParamWithName<T>(true, &m_An2_1, prefix + "SuperShape3D_an21"));
m_Params.push_back(ParamWithName<T>(true, &m_An2_2, prefix + "SuperShape3D_an22"));
m_Params.push_back(ParamWithName<T>(true, &m_Bn3_1, prefix + "SuperShape3D_bn31"));
m_Params.push_back(ParamWithName<T>(true, &m_Bn3_2, prefix + "SuperShape3D_bn32"));
m_Params.push_back(ParamWithName<T>(true, &m_M4_1, prefix + "SuperShape3D_m41"));
m_Params.push_back(ParamWithName<T>(true, &m_M4_2, prefix + "SuperShape3D_m42"));
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
m_Params.push_back(ParamWithName<T>(true, &m_Rho2Pi, prefix + "SuperShape3D_rho2pi"));
m_Params.push_back(ParamWithName<T>(true, &m_Phi2Pi, prefix + "SuperShape3D_phi2pi"));
}
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
private:
T m_Rho;
T m_Phi;
T m_M1;
T m_M2;
T m_A1;
T m_A2;
T m_B1;
T m_B2;
T m_N1_1;
T m_N1_2;
T m_N2_1;
T m_N2_2;
T m_N3_1;
T m_N3_2;
T m_Spiral;
T m_Toroidmap;
T m_N1n_1;//Precalc.
T m_N1n_2;
T m_An2_1;
T m_An2_2;
T m_Bn3_1;
T m_Bn3_2;
T m_M4_1;
T m_M4_2;
T m_Rho2Pi;
T m_Phi2Pi;
};
/// <summary>
/// sphyp3D.
/// </summary>
template <typename T>
--Code changes -Split VariationList into .h/cpp files. --Make VariationList.cpp the only place where the variations files are included from. This speeds up the build, but variations can no longer be directly instantiated. --Variations are no longer exported.
2016-02-18 21:58:24 -05:00
class Sphyp3DVariation : public ParametricVariation<T>
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
public:
--Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums.
2016-01-04 19:50:15 -05:00
Sphyp3DVariation(T weight = 1.0) : ParametricVariation<T>("sphyp3D", eVariationId::VAR_SPHYP3D, weight, true)
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
Init();
}
PARVARCOPY(Sphyp3DVariation)
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
T t, rX, rY, rZ;
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
t = Zeps(helper.m_PrecalcSumSquares + SQR(helper.In.z));
--User changes -Add new variations: bubbleT3D, crob, hexaplay3D, hexcrop, hexes, hexnix3D, loonie2, loonie3, nBlur, octapol and synth. -Allow for pre/post versions of dc_bubble, dc_cylinder and dc_linear whereas before they were omitted. -When saving a file with multiple embers in it, detect if time values are all the same and if so, start them at zero and increment by 1 for each ember. -Allow for numerous quality increases to be coalesced into one. It will pick up at the end of the current render. -Show selection highlight on variations tree in response to mouse hover. This makes it easier to see for which variation or param the current mouse wheel action will apply. -Make default temporal samples be 100, whereas before it was 1000 which was overkill. -Require the shift key to be held with delete for deleting an ember to prevent it from triggering when the user enters delete in the edit box. -This wasn't otherwise fixable without writing a lot more code. --Bug fixes -EmberGenome was crashing when generating a sequence from a source file with more than 2 embers in it. -EmberGenome was improperly handling the first frame of a merge after the last frame of the loop. -These bugs were due to a previous commit. Revert parts of that commit. -Prevent a zoom value of less than 0 when reading from xml. -Slight optimization of the crescents, and mask variations, if the compiler wasn't doing it already. -Unique file naming was broken because it was looking for _# and the default names ended with -#. -Disallow renaming of an ember in the library tree to an empty string. -Severe bug that prevented some variations from being read correctly from params generated outside this program. -Severe OpenCL randomization bug. The first x coordinates of the first points in the first kernel call of the first ember of a render since the OpenCL renderer object was created were not random and were mostly -1. -Severe bug when populating xform selection distributions that could sometimes cause a crash due to roundoff error. Fix by using double. -Limit the max number of variations in a random ember to MAX_CL_VARS, which is 8. This ensures they'll look the same on CPU and GPU. -Prevent user from saving stylesheet to default.qss, it's a special reserved filename. --Code changes -Generalize using the running sum output point inside of a variation for all cases: pre, reg and post. -Allow for array variables in variations where the address of each element is stored in m_Params. -Qualify all math functions with std:: -No longer use our own Clamp() in OpenCL, instead use the standard clamp(). -Redesign how functions are used in the variations OpenCL code. -Add tests to EmberTester to verify some of the new functionality. -Place more const and override qualifiers on functions where appropriate. -Add a global rand with a lock to be used very sparingly. -Use a map instead of a vector for bad param names in Xml parsing. -Prefix affine interpolation mode defines with "AFFINE_" to make their purpose more clear. -Allow for variations that change state during iteration by sending a separate copy of the ember to each rendering thread. -Implement this same functionality with a local struct in OpenCL. It's members are the total of all variables that need to change state within an ember. -Add Contains() function to Utils.h. -EmberRender: print names of kernels being printed with --dump_kernel option. -Clean up EmberTester to handle some of the recent changes. -Fix various casts. -Replace % 2 with & 1, even though the compiler was likely doing this already. -Add new file Variations06.h to accommodate new variations. -General cleanup.
2015-11-22 17:15:07 -05:00
rX = m_Weight / std::pow(t, m_StretchX);
rY = m_Weight / std::pow(t, m_StretchY);
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
helper.Out.x = helper.In.x * rX;
helper.Out.y = helper.In.y * rY;
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
//Optional 3D calculation.
More Linux work. Convert all casts to new style, away from legacy.
2014-12-07 02:51:44 -05:00
if (int(m_ZOn) == 1)
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
--User changes -Add new variations: bubbleT3D, crob, hexaplay3D, hexcrop, hexes, hexnix3D, loonie2, loonie3, nBlur, octapol and synth. -Allow for pre/post versions of dc_bubble, dc_cylinder and dc_linear whereas before they were omitted. -When saving a file with multiple embers in it, detect if time values are all the same and if so, start them at zero and increment by 1 for each ember. -Allow for numerous quality increases to be coalesced into one. It will pick up at the end of the current render. -Show selection highlight on variations tree in response to mouse hover. This makes it easier to see for which variation or param the current mouse wheel action will apply. -Make default temporal samples be 100, whereas before it was 1000 which was overkill. -Require the shift key to be held with delete for deleting an ember to prevent it from triggering when the user enters delete in the edit box. -This wasn't otherwise fixable without writing a lot more code. --Bug fixes -EmberGenome was crashing when generating a sequence from a source file with more than 2 embers in it. -EmberGenome was improperly handling the first frame of a merge after the last frame of the loop. -These bugs were due to a previous commit. Revert parts of that commit. -Prevent a zoom value of less than 0 when reading from xml. -Slight optimization of the crescents, and mask variations, if the compiler wasn't doing it already. -Unique file naming was broken because it was looking for _# and the default names ended with -#. -Disallow renaming of an ember in the library tree to an empty string. -Severe bug that prevented some variations from being read correctly from params generated outside this program. -Severe OpenCL randomization bug. The first x coordinates of the first points in the first kernel call of the first ember of a render since the OpenCL renderer object was created were not random and were mostly -1. -Severe bug when populating xform selection distributions that could sometimes cause a crash due to roundoff error. Fix by using double. -Limit the max number of variations in a random ember to MAX_CL_VARS, which is 8. This ensures they'll look the same on CPU and GPU. -Prevent user from saving stylesheet to default.qss, it's a special reserved filename. --Code changes -Generalize using the running sum output point inside of a variation for all cases: pre, reg and post. -Allow for array variables in variations where the address of each element is stored in m_Params. -Qualify all math functions with std:: -No longer use our own Clamp() in OpenCL, instead use the standard clamp(). -Redesign how functions are used in the variations OpenCL code. -Add tests to EmberTester to verify some of the new functionality. -Place more const and override qualifiers on functions where appropriate. -Add a global rand with a lock to be used very sparingly. -Use a map instead of a vector for bad param names in Xml parsing. -Prefix affine interpolation mode defines with "AFFINE_" to make their purpose more clear. -Allow for variations that change state during iteration by sending a separate copy of the ember to each rendering thread. -Implement this same functionality with a local struct in OpenCL. It's members are the total of all variables that need to change state within an ember. -Add Contains() function to Utils.h. -EmberRender: print names of kernels being printed with --dump_kernel option. -Clean up EmberTester to handle some of the recent changes. -Fix various casts. -Replace % 2 with & 1, even though the compiler was likely doing this already. -Add new file Variations06.h to accommodate new variations. -General cleanup.
2015-11-22 17:15:07 -05:00
rZ = m_Weight / std::pow(t, m_StretchZ);
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
helper.Out.z = helper.In.z * rZ;
}
}
--Code changes -Better use of const functions. -Ember.pro should not define BUILDING_EMBERCL, instead define BUILDING_EMBER.
2015-06-29 23:13:53 -04:00
virtual string OpenCLString() const override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
ostringstream ss, ss2;
0.4.1.3 Beta 10/14/2014 --User Changes Size is no longer fixed to the window size. Size scaling is done differently in the final render dialog. This fixes several bugs. Remove Xml saving size from settings and options dialog, it no longer applies. Final render can be broken into strips. Set default save path to the desktop if none is found in the settings file. Set default output size to 1920x1080 if none is found in the settings file. --Bug Fixes Better memory size reporting in final render dialog. --Code Changes Migrate to C++11, Qt 5.3.1, and Visual Studio 2013. Change most instances of unsigned int to size_t, and int to intmax_t. Add m_OrigPixPerUnit and m_ScaleType to Ember for scaling purposes. Replace some sprintf_s() calls in XmlToEmber with ostringstream. Move more non-templated members into RendererBase. Add CopyVec() overload that takes a per element function pointer. Add vector Memset(). Replace '&' with '+' instead of "&amp;" in XmlToEmber for much faster parsing. Break strips rendering out into EmberCommon and call from EmberRender and Fractorium. Make AddAndWriteBuffer() just call WriteBuffer(). Make AddAndWriteImage() delete the existing image first before replacing it. Add SetOutputTexture() to RendererCL to support making new textures in response to resize events. Remove multiple return statements in RendererCL, and replace with a bool that tracks results. Add ToDouble(), MakeEnd(), ToString() and Exists() wrappers in Fractorium. Add Size() wrapper in EmberFile. Make QString function arguments const QString&, and string with const string&. Make ShowCritical() wrapper for invoking a message box from another thread. Add combo box to TwoButtonWidget and rename.
2014-10-14 11:53:15 -04:00
intmax_t i = 0, varIndex = IndexInXform();
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
ss2 << "_" << XformIndexInEmber() << "]";
string index = ss2.str();
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
string weight = WeightDefineString();
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
string stretchX = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string stretchY = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string stretchZ = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
string zOn = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
ss << "\t{\n"
<< "\t\treal_t t, rX, rY, rZ;\n"
<< "\n"
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
<< "\t\tt = Zeps(fma(vIn.z, vIn.z, precalcSumSquares));\n"
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
<< "\t\trX = " << weight << " / pow(t, " << stretchX << ");\n"
<< "\t\trY = " << weight << " / pow(t, " << stretchY << ");\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\n"
<< "\t\tvOut.x = vIn.x * rX;\n"
<< "\t\tvOut.y = vIn.y * rY;\n"
<< "\n"
<< "\t\tif ((int)" << zOn << " == 1)\n"
<< "\t\t{\n"
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
<< "\t\trZ = " << weight << " / pow(t, " << stretchZ << ");\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\n"
<< "\t\tvOut.z = vIn.z * rZ;\n"
<< "\t\t}\n"
<< "\t}\n";
return ss.str();
}
--User changes -Add new variations: bubbleT3D, crob, hexaplay3D, hexcrop, hexes, hexnix3D, loonie2, loonie3, nBlur, octapol and synth. -Allow for pre/post versions of dc_bubble, dc_cylinder and dc_linear whereas before they were omitted. -When saving a file with multiple embers in it, detect if time values are all the same and if so, start them at zero and increment by 1 for each ember. -Allow for numerous quality increases to be coalesced into one. It will pick up at the end of the current render. -Show selection highlight on variations tree in response to mouse hover. This makes it easier to see for which variation or param the current mouse wheel action will apply. -Make default temporal samples be 100, whereas before it was 1000 which was overkill. -Require the shift key to be held with delete for deleting an ember to prevent it from triggering when the user enters delete in the edit box. -This wasn't otherwise fixable without writing a lot more code. --Bug fixes -EmberGenome was crashing when generating a sequence from a source file with more than 2 embers in it. -EmberGenome was improperly handling the first frame of a merge after the last frame of the loop. -These bugs were due to a previous commit. Revert parts of that commit. -Prevent a zoom value of less than 0 when reading from xml. -Slight optimization of the crescents, and mask variations, if the compiler wasn't doing it already. -Unique file naming was broken because it was looking for _# and the default names ended with -#. -Disallow renaming of an ember in the library tree to an empty string. -Severe bug that prevented some variations from being read correctly from params generated outside this program. -Severe OpenCL randomization bug. The first x coordinates of the first points in the first kernel call of the first ember of a render since the OpenCL renderer object was created were not random and were mostly -1. -Severe bug when populating xform selection distributions that could sometimes cause a crash due to roundoff error. Fix by using double. -Limit the max number of variations in a random ember to MAX_CL_VARS, which is 8. This ensures they'll look the same on CPU and GPU. -Prevent user from saving stylesheet to default.qss, it's a special reserved filename. --Code changes -Generalize using the running sum output point inside of a variation for all cases: pre, reg and post. -Allow for array variables in variations where the address of each element is stored in m_Params. -Qualify all math functions with std:: -No longer use our own Clamp() in OpenCL, instead use the standard clamp(). -Redesign how functions are used in the variations OpenCL code. -Add tests to EmberTester to verify some of the new functionality. -Place more const and override qualifiers on functions where appropriate. -Add a global rand with a lock to be used very sparingly. -Use a map instead of a vector for bad param names in Xml parsing. -Prefix affine interpolation mode defines with "AFFINE_" to make their purpose more clear. -Allow for variations that change state during iteration by sending a separate copy of the ember to each rendering thread. -Implement this same functionality with a local struct in OpenCL. It's members are the total of all variables that need to change state within an ember. -Add Contains() function to Utils.h. -EmberRender: print names of kernels being printed with --dump_kernel option. -Clean up EmberTester to handle some of the recent changes. -Fix various casts. -Replace % 2 with & 1, even though the compiler was likely doing this already. -Add new file Variations06.h to accommodate new variations. -General cleanup.
2015-11-22 17:15:07 -05:00
virtual vector<string> OpenCLGlobalFuncNames() const override
{
return vector<string> { "Zeps" };
}
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
protected:
void Init()
{
string prefix = Prefix();
m_Params.clear();
m_Params.push_back(ParamWithName<T>(&m_StretchX, prefix + "sphyp3D_stretchX", 1));
m_Params.push_back(ParamWithName<T>(&m_StretchY, prefix + "sphyp3D_stretchY", 1));
m_Params.push_back(ParamWithName<T>(&m_StretchZ, prefix + "sphyp3D_stretchZ", 1));
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
m_Params.push_back(ParamWithName<T>(&m_ZOn, prefix + "sphyp3D_zOn", 1, eParamType::INTEGER, 0, 1));
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
}
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
private:
T m_StretchX;
T m_StretchY;
T m_StretchZ;
T m_ZOn;
};
/// <summary>
/// circlecrop.
/// </summary>
template <typename T>
--Code changes -Split VariationList into .h/cpp files. --Make VariationList.cpp the only place where the variations files are included from. This speeds up the build, but variations can no longer be directly instantiated. --Variations are no longer exported.
2016-02-18 21:58:24 -05:00
class CirclecropVariation : public ParametricVariation<T>
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
public:
--Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums.
2016-01-04 19:50:15 -05:00
CirclecropVariation(T weight = 1.0) : ParametricVariation<T>("circlecrop", eVariationId::VAR_CIRCLECROP, weight)
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
Init();
}
PARVARCOPY(CirclecropVariation)
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
--User changes -Add post_smartcrop. --Bug fixes -Fix bug in crackle. -Wrong point assignment in hexaplay3D, hexnix3D. -Improper Z assignment in rblur. -Fix inconsistency with original in circlecrop. -Put EMBER_ROOT bakc to ./../../../ in default.pri. This is TBD. --Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums. -Optimize crob, nBlur. -Fix weird assignment statement in falloff3. -Cleanup in VarFuncs::SimplexNoise3D(). -Replace fabs() with std::abs(). -General cleanup.
2016-01-12 23:42:12 -05:00
T xi = helper.In.x - m_X;
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
T yi = helper.In.y - m_Y;
--User changes -Add post_smartcrop. --Bug fixes -Fix bug in crackle. -Wrong point assignment in hexaplay3D, hexnix3D. -Improper Z assignment in rblur. -Fix inconsistency with original in circlecrop. -Put EMBER_ROOT bakc to ./../../../ in default.pri. This is TBD. --Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums. -Optimize crob, nBlur. -Fix weird assignment statement in falloff3. -Cleanup in VarFuncs::SimplexNoise3D(). -Replace fabs() with std::abs(). -General cleanup.
2016-01-12 23:42:12 -05:00
if (m_VarType == eVariationType::VARTYPE_REG)//Original altered the input pointed to for reg.
{
helper.m_TransX -= m_X;
helper.m_TransY -= m_Y;
helper.Out.z = m_Weight * helper.In.z;//Original only assigned z for reg. Will be summed.
}
else
{
helper.Out.z = helper.In.z;//Original did nothing with z for pre/post, so passthrough direct assign.
}
Explicitly namespace calls to sqrt and isnan osx/linux have alternative top-level implementations of these methods, so the calls are ambiguous given the - so these should be used explicitly
2015-04-13 07:32:58 -04:00
const T rad = std::sqrt(SQR(xi) + SQR(yi));
--User changes -Add new variations: bubbleT3D, crob, hexaplay3D, hexcrop, hexes, hexnix3D, loonie2, loonie3, nBlur, octapol and synth. -Allow for pre/post versions of dc_bubble, dc_cylinder and dc_linear whereas before they were omitted. -When saving a file with multiple embers in it, detect if time values are all the same and if so, start them at zero and increment by 1 for each ember. -Allow for numerous quality increases to be coalesced into one. It will pick up at the end of the current render. -Show selection highlight on variations tree in response to mouse hover. This makes it easier to see for which variation or param the current mouse wheel action will apply. -Make default temporal samples be 100, whereas before it was 1000 which was overkill. -Require the shift key to be held with delete for deleting an ember to prevent it from triggering when the user enters delete in the edit box. -This wasn't otherwise fixable without writing a lot more code. --Bug fixes -EmberGenome was crashing when generating a sequence from a source file with more than 2 embers in it. -EmberGenome was improperly handling the first frame of a merge after the last frame of the loop. -These bugs were due to a previous commit. Revert parts of that commit. -Prevent a zoom value of less than 0 when reading from xml. -Slight optimization of the crescents, and mask variations, if the compiler wasn't doing it already. -Unique file naming was broken because it was looking for _# and the default names ended with -#. -Disallow renaming of an ember in the library tree to an empty string. -Severe bug that prevented some variations from being read correctly from params generated outside this program. -Severe OpenCL randomization bug. The first x coordinates of the first points in the first kernel call of the first ember of a render since the OpenCL renderer object was created were not random and were mostly -1. -Severe bug when populating xform selection distributions that could sometimes cause a crash due to roundoff error. Fix by using double. -Limit the max number of variations in a random ember to MAX_CL_VARS, which is 8. This ensures they'll look the same on CPU and GPU. -Prevent user from saving stylesheet to default.qss, it's a special reserved filename. --Code changes -Generalize using the running sum output point inside of a variation for all cases: pre, reg and post. -Allow for array variables in variations where the address of each element is stored in m_Params. -Qualify all math functions with std:: -No longer use our own Clamp() in OpenCL, instead use the standard clamp(). -Redesign how functions are used in the variations OpenCL code. -Add tests to EmberTester to verify some of the new functionality. -Place more const and override qualifiers on functions where appropriate. -Add a global rand with a lock to be used very sparingly. -Use a map instead of a vector for bad param names in Xml parsing. -Prefix affine interpolation mode defines with "AFFINE_" to make their purpose more clear. -Allow for variations that change state during iteration by sending a separate copy of the ember to each rendering thread. -Implement this same functionality with a local struct in OpenCL. It's members are the total of all variables that need to change state within an ember. -Add Contains() function to Utils.h. -EmberRender: print names of kernels being printed with --dump_kernel option. -Clean up EmberTester to handle some of the recent changes. -Fix various casts. -Replace % 2 with & 1, even though the compiler was likely doing this already. -Add new file Variations06.h to accommodate new variations. -General cleanup.
2015-11-22 17:15:07 -05:00
const T ang = std::atan2(yi, xi);
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
const T rdc = m_Radius + (rand.Frand01<T>() * T(0.5) * m_Ca);
--User changes -Add new variations: bubbleT3D, crob, hexaplay3D, hexcrop, hexes, hexnix3D, loonie2, loonie3, nBlur, octapol and synth. -Allow for pre/post versions of dc_bubble, dc_cylinder and dc_linear whereas before they were omitted. -When saving a file with multiple embers in it, detect if time values are all the same and if so, start them at zero and increment by 1 for each ember. -Allow for numerous quality increases to be coalesced into one. It will pick up at the end of the current render. -Show selection highlight on variations tree in response to mouse hover. This makes it easier to see for which variation or param the current mouse wheel action will apply. -Make default temporal samples be 100, whereas before it was 1000 which was overkill. -Require the shift key to be held with delete for deleting an ember to prevent it from triggering when the user enters delete in the edit box. -This wasn't otherwise fixable without writing a lot more code. --Bug fixes -EmberGenome was crashing when generating a sequence from a source file with more than 2 embers in it. -EmberGenome was improperly handling the first frame of a merge after the last frame of the loop. -These bugs were due to a previous commit. Revert parts of that commit. -Prevent a zoom value of less than 0 when reading from xml. -Slight optimization of the crescents, and mask variations, if the compiler wasn't doing it already. -Unique file naming was broken because it was looking for _# and the default names ended with -#. -Disallow renaming of an ember in the library tree to an empty string. -Severe bug that prevented some variations from being read correctly from params generated outside this program. -Severe OpenCL randomization bug. The first x coordinates of the first points in the first kernel call of the first ember of a render since the OpenCL renderer object was created were not random and were mostly -1. -Severe bug when populating xform selection distributions that could sometimes cause a crash due to roundoff error. Fix by using double. -Limit the max number of variations in a random ember to MAX_CL_VARS, which is 8. This ensures they'll look the same on CPU and GPU. -Prevent user from saving stylesheet to default.qss, it's a special reserved filename. --Code changes -Generalize using the running sum output point inside of a variation for all cases: pre, reg and post. -Allow for array variables in variations where the address of each element is stored in m_Params. -Qualify all math functions with std:: -No longer use our own Clamp() in OpenCL, instead use the standard clamp(). -Redesign how functions are used in the variations OpenCL code. -Add tests to EmberTester to verify some of the new functionality. -Place more const and override qualifiers on functions where appropriate. -Add a global rand with a lock to be used very sparingly. -Use a map instead of a vector for bad param names in Xml parsing. -Prefix affine interpolation mode defines with "AFFINE_" to make their purpose more clear. -Allow for variations that change state during iteration by sending a separate copy of the ember to each rendering thread. -Implement this same functionality with a local struct in OpenCL. It's members are the total of all variables that need to change state within an ember. -Add Contains() function to Utils.h. -EmberRender: print names of kernels being printed with --dump_kernel option. -Clean up EmberTester to handle some of the recent changes. -Fix various casts. -Replace % 2 with & 1, even though the compiler was likely doing this already. -Add new file Variations06.h to accommodate new variations. -General cleanup.
2015-11-22 17:15:07 -05:00
const T s = std::sin(ang);
const T c = std::cos(ang);
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
const int esc = rad > m_Radius;
More Linux work. Convert all casts to new style, away from legacy.
2014-12-07 02:51:44 -05:00
const int cr0 = int(m_Zero);
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
--User changes -Add post_smartcrop. --Bug fixes -Fix bug in crackle. -Wrong point assignment in hexaplay3D, hexnix3D. -Improper Z assignment in rblur. -Fix inconsistency with original in circlecrop. -Put EMBER_ROOT bakc to ./../../../ in default.pri. This is TBD. --Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums. -Optimize crob, nBlur. -Fix weird assignment statement in falloff3. -Cleanup in VarFuncs::SimplexNoise3D(). -Replace fabs() with std::abs(). -General cleanup.
2016-01-12 23:42:12 -05:00
if (cr0 && esc)
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
helper.Out.x = helper.Out.y = 0;
--User changes -Add post_smartcrop. --Bug fixes -Fix bug in crackle. -Wrong point assignment in hexaplay3D, hexnix3D. -Improper Z assignment in rblur. -Fix inconsistency with original in circlecrop. -Put EMBER_ROOT bakc to ./../../../ in default.pri. This is TBD. --Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums. -Optimize crob, nBlur. -Fix weird assignment statement in falloff3. -Cleanup in VarFuncs::SimplexNoise3D(). -Replace fabs() with std::abs(). -General cleanup.
2016-01-12 23:42:12 -05:00
if (m_VarType == eVariationType::VARTYPE_REG)
outPoint.m_X = outPoint.m_Y = 0;
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
}
else if (cr0 && !esc)
{
helper.Out.x = m_Weight * xi + m_X;
helper.Out.y = m_Weight * yi + m_Y;
}
--Bug fixes -Fix some potential OpenCL compilation bugs on circlecrop, circlecrop2, oscilloscope2, Truchet_hex_crop. -Prevent the minus key from changing the current xform when being pressed in a textbox. --Code changes -Cleanup some casting in depth_blur, depth_blur2.
2019-06-24 22:43:15 -04:00
else if (!cr0 && esc)
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
helper.Out.x = m_Weight * rdc * c + m_X;
helper.Out.y = m_Weight * rdc * s + m_Y;
}
else if (!cr0 && !esc)
{
helper.Out.x = m_Weight * xi + m_X;
helper.Out.y = m_Weight * yi + m_Y;
}
}
--Code changes -Better use of const functions. -Ember.pro should not define BUILDING_EMBERCL, instead define BUILDING_EMBER.
2015-06-29 23:13:53 -04:00
virtual string OpenCLString() const override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
ostringstream ss, ss2;
0.4.1.3 Beta 10/14/2014 --User Changes Size is no longer fixed to the window size. Size scaling is done differently in the final render dialog. This fixes several bugs. Remove Xml saving size from settings and options dialog, it no longer applies. Final render can be broken into strips. Set default save path to the desktop if none is found in the settings file. Set default output size to 1920x1080 if none is found in the settings file. --Bug Fixes Better memory size reporting in final render dialog. --Code Changes Migrate to C++11, Qt 5.3.1, and Visual Studio 2013. Change most instances of unsigned int to size_t, and int to intmax_t. Add m_OrigPixPerUnit and m_ScaleType to Ember for scaling purposes. Replace some sprintf_s() calls in XmlToEmber with ostringstream. Move more non-templated members into RendererBase. Add CopyVec() overload that takes a per element function pointer. Add vector Memset(). Replace '&' with '+' instead of "&amp;" in XmlToEmber for much faster parsing. Break strips rendering out into EmberCommon and call from EmberRender and Fractorium. Make AddAndWriteBuffer() just call WriteBuffer(). Make AddAndWriteImage() delete the existing image first before replacing it. Add SetOutputTexture() to RendererCL to support making new textures in response to resize events. Remove multiple return statements in RendererCL, and replace with a bool that tracks results. Add ToDouble(), MakeEnd(), ToString() and Exists() wrappers in Fractorium. Add Size() wrapper in EmberFile. Make QString function arguments const QString&, and string with const string&. Make ShowCritical() wrapper for invoking a message box from another thread. Add combo box to TwoButtonWidget and rename.
2014-10-14 11:53:15 -04:00
intmax_t i = 0, varIndex = IndexInXform();
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
ss2 << "_" << XformIndexInEmber() << "]";
string index = ss2.str();
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
string weight = WeightDefineString();
string radius = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string x = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string y = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
string scatterArea = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
string zero = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string ca = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
ss << "\t{\n"
--User changes -Add post_smartcrop. --Bug fixes -Fix bug in crackle. -Wrong point assignment in hexaplay3D, hexnix3D. -Improper Z assignment in rblur. -Fix inconsistency with original in circlecrop. -Put EMBER_ROOT bakc to ./../../../ in default.pri. This is TBD. --Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums. -Optimize crob, nBlur. -Fix weird assignment statement in falloff3. -Cleanup in VarFuncs::SimplexNoise3D(). -Replace fabs() with std::abs(). -General cleanup.
2016-01-12 23:42:12 -05:00
<< "\t\treal_t xi = vIn.x - " << x << ";\n"
<< "\t\treal_t yi = vIn.y - " << y << ";\n"
<< "\n";
if (m_VarType == eVariationType::VARTYPE_REG)//Original altered the input pointed to for reg.
{
ss
<< "\t\ttransX -= " << x << ";\n"
<< "\t\ttransY -= " << y << ";\n"
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
<< "\t\tvOut.z = " << weight << " * vIn.z;\n";
--User changes -Add post_smartcrop. --Bug fixes -Fix bug in crackle. -Wrong point assignment in hexaplay3D, hexnix3D. -Improper Z assignment in rblur. -Fix inconsistency with original in circlecrop. -Put EMBER_ROOT bakc to ./../../../ in default.pri. This is TBD. --Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums. -Optimize crob, nBlur. -Fix weird assignment statement in falloff3. -Cleanup in VarFuncs::SimplexNoise3D(). -Replace fabs() with std::abs(). -General cleanup.
2016-01-12 23:42:12 -05:00
}
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
else
--User changes -Add post_smartcrop. --Bug fixes -Fix bug in crackle. -Wrong point assignment in hexaplay3D, hexnix3D. -Improper Z assignment in rblur. -Fix inconsistency with original in circlecrop. -Put EMBER_ROOT bakc to ./../../../ in default.pri. This is TBD. --Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums. -Optimize crob, nBlur. -Fix weird assignment statement in falloff3. -Cleanup in VarFuncs::SimplexNoise3D(). -Replace fabs() with std::abs(). -General cleanup.
2016-01-12 23:42:12 -05:00
{
ss
<< "\t\tvOut.z = vIn.z;\n";
}
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
--User changes -Add post_smartcrop. --Bug fixes -Fix bug in crackle. -Wrong point assignment in hexaplay3D, hexnix3D. -Improper Z assignment in rblur. -Fix inconsistency with original in circlecrop. -Put EMBER_ROOT bakc to ./../../../ in default.pri. This is TBD. --Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums. -Optimize crob, nBlur. -Fix weird assignment statement in falloff3. -Cleanup in VarFuncs::SimplexNoise3D(). -Replace fabs() with std::abs(). -General cleanup.
2016-01-12 23:42:12 -05:00
ss
<< "\t\tconst real_t rad = sqrt(SQR(xi) + SQR(yi));\n"
<< "\t\tconst real_t ang = atan2(yi, xi);\n"
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
<< "\t\tconst real_t rdc = fma(MwcNext01(mwc) * (real_t)(0.5), " << ca << ", " << radius << ");\n"
--User changes -Add post_smartcrop. --Bug fixes -Fix bug in crackle. -Wrong point assignment in hexaplay3D, hexnix3D. -Improper Z assignment in rblur. -Fix inconsistency with original in circlecrop. -Put EMBER_ROOT bakc to ./../../../ in default.pri. This is TBD. --Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums. -Optimize crob, nBlur. -Fix weird assignment statement in falloff3. -Cleanup in VarFuncs::SimplexNoise3D(). -Replace fabs() with std::abs(). -General cleanup.
2016-01-12 23:42:12 -05:00
<< "\t\tconst real_t s = sin(ang);\n"
<< "\t\tconst real_t c = cos(ang);\n"
<< "\n"
<< "\t\tconst int esc = rad > " << radius << ";\n"
<< "\t\tconst int cr0 = (int)" << zero << ";\n"
<< "\n"
--User changes -Add new options to EmberGenome to better help assess which variations have certain characteristics. The options are: --sumvars: Display the names of all regular variations which have the default behavior of summing their outputs and exit. --assignvars: Display the names of all regular variations which have the non-standard behavior of assigning their outputs and exit. --ppsumvars: Display the names of all pre/post variations which have the non-standard behavior of summing their outputs and exit. --ppassignvars: Display the names of all pre/post variations which have the default behavior of assigning their outputs and exit. --dcvars: Display the names of all variations which use direct coloring and exit. --parvars: Display the names of all variations which have parameters and exit. --nonparvars: Display the names of all variations which do not have parameters (weight only) and exit. --Code changes -Make VariationList vectors have const elements since no callers should ever change them. -Add new function to VariationList to retrieve a const ref to the parametric variations. -Move some search functions out of EmberTester and into EmberCommon.h. -General code cleanup.
2016-05-15 02:33:08 -04:00
<< "\t\tif (cr0 && esc)\n"
--User changes -Add post_smartcrop. --Bug fixes -Fix bug in crackle. -Wrong point assignment in hexaplay3D, hexnix3D. -Improper Z assignment in rblur. -Fix inconsistency with original in circlecrop. -Put EMBER_ROOT bakc to ./../../../ in default.pri. This is TBD. --Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums. -Optimize crob, nBlur. -Fix weird assignment statement in falloff3. -Cleanup in VarFuncs::SimplexNoise3D(). -Replace fabs() with std::abs(). -General cleanup.
2016-01-12 23:42:12 -05:00
<< "\t\t{\n"
<< "\t\t vOut.x = vOut.y = 0;\n";
if (m_VarType == eVariationType::VARTYPE_REG)
ss << "\t\t outPoint->m_X = outPoint->m_Y = 0;\n";
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
--User changes -Add post_smartcrop. --Bug fixes -Fix bug in crackle. -Wrong point assignment in hexaplay3D, hexnix3D. -Improper Z assignment in rblur. -Fix inconsistency with original in circlecrop. -Put EMBER_ROOT bakc to ./../../../ in default.pri. This is TBD. --Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums. -Optimize crob, nBlur. -Fix weird assignment statement in falloff3. -Cleanup in VarFuncs::SimplexNoise3D(). -Replace fabs() with std::abs(). -General cleanup.
2016-01-12 23:42:12 -05:00
ss
<< "\t\t}\n"
<< "\t\telse if (cr0 && !esc)\n"
<< "\t\t{\n"
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
<< "\t\t vOut.x = fma(" << weight << ", xi, " << x << ");\n"
<< "\t\t vOut.y = fma(" << weight << ", yi, " << y << ");\n"
--User changes -Add post_smartcrop. --Bug fixes -Fix bug in crackle. -Wrong point assignment in hexaplay3D, hexnix3D. -Improper Z assignment in rblur. -Fix inconsistency with original in circlecrop. -Put EMBER_ROOT bakc to ./../../../ in default.pri. This is TBD. --Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums. -Optimize crob, nBlur. -Fix weird assignment statement in falloff3. -Cleanup in VarFuncs::SimplexNoise3D(). -Replace fabs() with std::abs(). -General cleanup.
2016-01-12 23:42:12 -05:00
<< "\t\t}\n"
<< "\t\telse if (!cr0 && esc)\n"
<< "\t\t{\n"
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
<< "\t\t vOut.x = fma(" << weight << ", rdc * c, " << x << ");\n"
<< "\t\t vOut.y = fma(" << weight << ", rdc * s, " << y << ");\n"
--User changes -Add post_smartcrop. --Bug fixes -Fix bug in crackle. -Wrong point assignment in hexaplay3D, hexnix3D. -Improper Z assignment in rblur. -Fix inconsistency with original in circlecrop. -Put EMBER_ROOT bakc to ./../../../ in default.pri. This is TBD. --Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums. -Optimize crob, nBlur. -Fix weird assignment statement in falloff3. -Cleanup in VarFuncs::SimplexNoise3D(). -Replace fabs() with std::abs(). -General cleanup.
2016-01-12 23:42:12 -05:00
<< "\t\t}\n"
<< "\t\telse if (!cr0 && !esc)\n"
<< "\t\t{\n"
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
<< "\t\t vOut.x = fma(" << weight << ", xi, " << x << ");\n"
<< "\t\t vOut.y = fma(" << weight << ", yi, " << y << ");\n"
--User changes -Add post_smartcrop. --Bug fixes -Fix bug in crackle. -Wrong point assignment in hexaplay3D, hexnix3D. -Improper Z assignment in rblur. -Fix inconsistency with original in circlecrop. -Put EMBER_ROOT bakc to ./../../../ in default.pri. This is TBD. --Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums. -Optimize crob, nBlur. -Fix weird assignment statement in falloff3. -Cleanup in VarFuncs::SimplexNoise3D(). -Replace fabs() with std::abs(). -General cleanup.
2016-01-12 23:42:12 -05:00
<< "\t\t}\n"
<< "\t}\n";
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
return ss.str();
}
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
virtual void Precalc() override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
m_Ca = Clamp<T>(m_ScatterArea, -1, 1);
}
protected:
void Init()
{
string prefix = Prefix();
m_Params.clear();
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
m_Params.push_back(ParamWithName<T>(&m_Radius, prefix + "circlecrop_radius", 1));
m_Params.push_back(ParamWithName<T>(&m_X, prefix + "circlecrop_x"));
m_Params.push_back(ParamWithName<T>(&m_Y, prefix + "circlecrop_y"));
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
m_Params.push_back(ParamWithName<T>(&m_ScatterArea, prefix + "circlecrop_scatter_area"));
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
m_Params.push_back(ParamWithName<T>(&m_Zero, prefix + "circlecrop_zero", 1, eParamType::INTEGER, 0, 1));
m_Params.push_back(ParamWithName<T>(true, &m_Ca, prefix + "circlecrop_ca"));
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
}
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
private:
T m_Radius;
T m_X;
T m_Y;
T m_ScatterArea;
T m_Zero;
T m_Ca;//Precalc.
};
--User changes: -Show common folder locations such as documents, downloads, pictures in the sidebar in all file dialogs. -Warning message about exceeding memory in final render dialog now suggests strips as the solution to the problem. -Strips now has a tooltip explaining what it does. -Allow more digits in the spinners on the color section the flame tab. -Add manually adjustable size spinners in the final render dialog. Percentage scale and absolute size are fully synced. -Default prefix in final render is now the filename when doing animations (coming from sequence section of the library tab). -Changed the elliptic variation back to using a less precise version for float, and a more precise version for double. The last release had it always using double. -New applied xaos table that shows a read-only view of actual weights by taking the base xform weights and multiplying them by the xaos values. -New table in the xaos tab that gives a graphical representation of the probability that each xform is chosen, with and without xaos. -Add button to transpose the xaos rows and columns. -Add support for importing .chaos files from Chaotica. --Pasting back to Chaotica will work for most, but not all, variations due to incompatible parameter names in some. -Curves are now splines instead of Bezier. This adds compatibility with Chaotica, but breaks it for Apophysis. Xmls are still pastable, but the color curves will look different. --The curve editor on the palette tab can now add points by clicking on the lines and remove points by clicking on the points themselves, just like Chaotica. --Splines are saved in four new xml fields: overall_curve, red_curve, green_curve and blue_curve. -Allow for specifying the percentage of a sub batch each thread should iterate through per kernel call when running with OpenCL. This gives a roughly 1% performance increase due to having to make less kernel calls while iterating. --This field is present for interactive editing (where it's not very useful) and in the final render dialog. --On the command line, this is specified as --sbpctth for EmberRender and EmberAnimate. -Allow double clicking to toggle the supersample field in the flame tab between 1 and 2 for easily checking the effect of the field. -When showing affine values as polar coordinates, show angles normalized to 360 to match Chaotica. -Fuse Count spinner now toggles between 15 and 100 when double clicking for easily checking the effect of the field. -Added field for limiting the range in the x and y direction that the initial points are chosen from. -Added a field called K2 which is an alternative way to set brightness, ignored when zero. --This has no effect for many variations, but hs a noticeable effect for some. -Added new variations: arcsech arcsech2 arcsinh arctanh asteria block bwraps_rand circlecrop2 coth_spiral crackle2 depth_blur depth_blur2 depth_gaussian depth_gaussian2 depth_ngon depth_ngon2 depth_sine depth_sine2 dragonfire dspherical dust excinis exp2 flipx flowerdb foci_p gaussian glynnia2 glynnsim4 glynnsim5 henon henon hex_rand hex_truchet hypershift lazyjess lens lozi lozi modulusx modulusy oscilloscope2 point_symmetry pointsymmetry projective pulse rotate scry2 shift smartshape spher squares starblur2 swirl3 swirl3r tanh_spiral target0 target2 tile_hlp truchet_glyph truchet_inv truchet_knot unicorngaloshen vibration vibration2 --hex_truchet, hex_rand should always use double. They are extremely sensitive. --Bug fixes: -Bounds sign was flipped for x coordinate of world space when center was not zero. -Right clicking and dragging spinner showed menu on mouse up, even if it was very far away. -Text boxes for size in final render dialog were hard to type in. Same bug as xform weight used to be so fix the same way. -Fix spelling to be plural in toggle color speed box. -Stop using the blank user palette to generate flames. Either put colored palettes in it, or exclude it from randoms. -Clicking the random palette button for a palette file with only one palette in it would freeze the program. -Clicking none scale in final render did not re-render the preview. -Use less precision on random xaos. No need for 12 decimal places. -The term sub batch is overloaded in the options dialog. Change the naming and tooltip of those settings for cpu and opencl. --Also made clear in the tooltip for the default opencl quality setting that the value is per device. -The arrows spinner in palette editor appears like a read-only label. Made it look like a spinner. -Fix border colors for various spin boxes and table headers in the style sheet. Requires reload. -Fix a bug in the bwraps variation which would produce different results than Chaotica and Apophysis. -Synth was allowed to be selected for random flame generation when using an Nvidia card but it shouldn't have been because Nvidia has a hard time compiling synth. -A casting bug in the OpenCL kernels for log scaling and density filtering was preventing successful compilations on Intel iGPUs. Fixed even though we don't support anything other than AMD and Nvidia. -Palette rotation (click and drag) position was not being reset when loading a new flame. -When the xform circles were hidden, opening and closing the options dialog would improperly reshow them. -Double click toggle was broken on integer spin boxes. -Fixed tab order of some controls. -Creating a palette from a jpg in the palette editor only produced a single color. --Needed to package imageformats/qjpeg.dll with the Windows installer. -The basic memory benchmark test flame was not really testing memory. Make it more spread out. -Remove the temporal samples field from the flame tab, it was never used because it's only an animation parameter which is specified in the final render dialog or on the command line with EmberAnimate. --Code changes: -Add IsEmpty() to Palette to determine if a palette is all black. -Attempt to avoid selecting a blank palette in PaletteList::GetRandomPalette(). -Add function ScanForChaosNodes() and some associated helper functions in XmlToEmber. -Make variation param name correction be case insensitive in XmlToEmber. -Report error when assigning a variation param value in XmlToEmber. -Add SubBatchPercentPerThread() method to RendererCL. -Override enterEvent() and leaveEvent() in DoubleSpinBox and SpinBox to prevent the context menu from showing up on right mouse up after already leaving the spinner. -Filtering the mouse wheel event in TableWidget no longer appears to be needed. It was probably an old Qt bug that has been fixed. -Gui/ember syncing code in the final render dialog needed to be reworked to accommodate absolute sizes.
2019-04-13 22:00:46 -04:00
/// <summary>
/// circlecrop2.
/// By tatasz.
/// </summary>
template <typename T>
class Circlecrop2Variation : public ParametricVariation<T>
{
public:
Circlecrop2Variation(T weight = 1.0) : ParametricVariation<T>("circlecrop2", eVariationId::VAR_CIRCLECROP2, weight, true, true, false, false, true)
{
Init();
}
PARVARCOPY(Circlecrop2Variation)
virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
{
T rad = helper.m_PrecalcSqrtSumSquares;
T ang = helper.m_PrecalcAtanyx;
T s = 0;
T c = 0;
if (rad > m_Out || rad < m_In)
{
if (!m_Zero)
{
s = std::sin(ang) * m_OutWeight;
c = std::cos(ang) * m_OutWeight;
}
}
else
{
s = helper.In.x * m_Weight;
c = helper.In.y * m_Weight;
}
helper.Out.x = s;
helper.Out.y = c;
helper.Out.z = DefaultZ(helper);
}
virtual string OpenCLString() const override
{
ostringstream ss, ss2;
intmax_t i = 0, varIndex = IndexInXform();
ss2 << "_" << XformIndexInEmber() << "]";
string index = ss2.str();
string weight = WeightDefineString();
string inner = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string outer = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string zero = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string in = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string out = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string outweight = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
ss << "\t{\n"
<< "\t\treal_t rad = precalcSqrtSumSquares;\n"
<< "\t\treal_t ang = precalcAtanyx;\n"
<< "\t\treal_t s = 0;\n"
<< "\t\treal_t c = 0;\n"
<< "\n"
<< "\t\tif (rad > " << out << " || rad < " << in << ")\n"
<< "\t\t{\n"
--Bug fixes -Fix some potential OpenCL compilation bugs on circlecrop, circlecrop2, oscilloscope2, Truchet_hex_crop. -Prevent the minus key from changing the current xform when being pressed in a textbox. --Code changes -Cleanup some casting in depth_blur, depth_blur2.
2019-06-24 22:43:15 -04:00
<< "\t\t if (" << zero << " == 0)\n"
--User changes: -Show common folder locations such as documents, downloads, pictures in the sidebar in all file dialogs. -Warning message about exceeding memory in final render dialog now suggests strips as the solution to the problem. -Strips now has a tooltip explaining what it does. -Allow more digits in the spinners on the color section the flame tab. -Add manually adjustable size spinners in the final render dialog. Percentage scale and absolute size are fully synced. -Default prefix in final render is now the filename when doing animations (coming from sequence section of the library tab). -Changed the elliptic variation back to using a less precise version for float, and a more precise version for double. The last release had it always using double. -New applied xaos table that shows a read-only view of actual weights by taking the base xform weights and multiplying them by the xaos values. -New table in the xaos tab that gives a graphical representation of the probability that each xform is chosen, with and without xaos. -Add button to transpose the xaos rows and columns. -Add support for importing .chaos files from Chaotica. --Pasting back to Chaotica will work for most, but not all, variations due to incompatible parameter names in some. -Curves are now splines instead of Bezier. This adds compatibility with Chaotica, but breaks it for Apophysis. Xmls are still pastable, but the color curves will look different. --The curve editor on the palette tab can now add points by clicking on the lines and remove points by clicking on the points themselves, just like Chaotica. --Splines are saved in four new xml fields: overall_curve, red_curve, green_curve and blue_curve. -Allow for specifying the percentage of a sub batch each thread should iterate through per kernel call when running with OpenCL. This gives a roughly 1% performance increase due to having to make less kernel calls while iterating. --This field is present for interactive editing (where it's not very useful) and in the final render dialog. --On the command line, this is specified as --sbpctth for EmberRender and EmberAnimate. -Allow double clicking to toggle the supersample field in the flame tab between 1 and 2 for easily checking the effect of the field. -When showing affine values as polar coordinates, show angles normalized to 360 to match Chaotica. -Fuse Count spinner now toggles between 15 and 100 when double clicking for easily checking the effect of the field. -Added field for limiting the range in the x and y direction that the initial points are chosen from. -Added a field called K2 which is an alternative way to set brightness, ignored when zero. --This has no effect for many variations, but hs a noticeable effect for some. -Added new variations: arcsech arcsech2 arcsinh arctanh asteria block bwraps_rand circlecrop2 coth_spiral crackle2 depth_blur depth_blur2 depth_gaussian depth_gaussian2 depth_ngon depth_ngon2 depth_sine depth_sine2 dragonfire dspherical dust excinis exp2 flipx flowerdb foci_p gaussian glynnia2 glynnsim4 glynnsim5 henon henon hex_rand hex_truchet hypershift lazyjess lens lozi lozi modulusx modulusy oscilloscope2 point_symmetry pointsymmetry projective pulse rotate scry2 shift smartshape spher squares starblur2 swirl3 swirl3r tanh_spiral target0 target2 tile_hlp truchet_glyph truchet_inv truchet_knot unicorngaloshen vibration vibration2 --hex_truchet, hex_rand should always use double. They are extremely sensitive. --Bug fixes: -Bounds sign was flipped for x coordinate of world space when center was not zero. -Right clicking and dragging spinner showed menu on mouse up, even if it was very far away. -Text boxes for size in final render dialog were hard to type in. Same bug as xform weight used to be so fix the same way. -Fix spelling to be plural in toggle color speed box. -Stop using the blank user palette to generate flames. Either put colored palettes in it, or exclude it from randoms. -Clicking the random palette button for a palette file with only one palette in it would freeze the program. -Clicking none scale in final render did not re-render the preview. -Use less precision on random xaos. No need for 12 decimal places. -The term sub batch is overloaded in the options dialog. Change the naming and tooltip of those settings for cpu and opencl. --Also made clear in the tooltip for the default opencl quality setting that the value is per device. -The arrows spinner in palette editor appears like a read-only label. Made it look like a spinner. -Fix border colors for various spin boxes and table headers in the style sheet. Requires reload. -Fix a bug in the bwraps variation which would produce different results than Chaotica and Apophysis. -Synth was allowed to be selected for random flame generation when using an Nvidia card but it shouldn't have been because Nvidia has a hard time compiling synth. -A casting bug in the OpenCL kernels for log scaling and density filtering was preventing successful compilations on Intel iGPUs. Fixed even though we don't support anything other than AMD and Nvidia. -Palette rotation (click and drag) position was not being reset when loading a new flame. -When the xform circles were hidden, opening and closing the options dialog would improperly reshow them. -Double click toggle was broken on integer spin boxes. -Fixed tab order of some controls. -Creating a palette from a jpg in the palette editor only produced a single color. --Needed to package imageformats/qjpeg.dll with the Windows installer. -The basic memory benchmark test flame was not really testing memory. Make it more spread out. -Remove the temporal samples field from the flame tab, it was never used because it's only an animation parameter which is specified in the final render dialog or on the command line with EmberAnimate. --Code changes: -Add IsEmpty() to Palette to determine if a palette is all black. -Attempt to avoid selecting a blank palette in PaletteList::GetRandomPalette(). -Add function ScanForChaosNodes() and some associated helper functions in XmlToEmber. -Make variation param name correction be case insensitive in XmlToEmber. -Report error when assigning a variation param value in XmlToEmber. -Add SubBatchPercentPerThread() method to RendererCL. -Override enterEvent() and leaveEvent() in DoubleSpinBox and SpinBox to prevent the context menu from showing up on right mouse up after already leaving the spinner. -Filtering the mouse wheel event in TableWidget no longer appears to be needed. It was probably an old Qt bug that has been fixed. -Gui/ember syncing code in the final render dialog needed to be reworked to accommodate absolute sizes.
2019-04-13 22:00:46 -04:00
<< "\t\t {\n"
<< "\t\t s = sin(ang) * " << outweight << ";\n"
<< "\t\t c = cos(ang) * " << outweight << ";\n"
<< "\t\t }\n"
<< "\t\t}\n"
<< "\t\telse\n"
<< "\t\t{\n"
<< "\t\t s = vIn.x * " << weight << ";\n"
<< "\t\t c = vIn.y * " << weight << ";\n"
<< "\t\t}\n"
<< "\n"
<< "\t\tvOut.x = s;\n"
<< "\t\tvOut.y = c;\n"
<< "\t\tvOut.z = " << DefaultZCl()
<< "\t}\n";
return ss.str();
}
virtual void Precalc() override
{
m_In = std::min(m_Inner, m_Outer);
m_Out = std::max(m_Inner, m_Outer);
m_OutWeight = m_Out * m_Weight;
}
protected:
void Init()
{
string prefix = Prefix();
m_Params.clear();
m_Params.push_back(ParamWithName<T>(&m_Inner, prefix + "circlecrop2_inner", T(0.5)));
m_Params.push_back(ParamWithName<T>(&m_Outer, prefix + "circlecrop2_outer", 1));
m_Params.push_back(ParamWithName<T>(&m_Zero, prefix + "circlecrop2_zero", 1, eParamType::INTEGER, 0, 1));
m_Params.push_back(ParamWithName<T>(true, &m_In, prefix + "circlecrop2_in"));
m_Params.push_back(ParamWithName<T>(true, &m_Out, prefix + "circlecrop2_out"));
m_Params.push_back(ParamWithName<T>(true, &m_OutWeight, prefix + "circlecrop2_out_weight"));
}
private:
T m_Inner;
T m_Outer;
T m_Zero;
T m_In;//Precalc.
T m_Out;
T m_OutWeight;
};
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
/// <summary>
/// julian3Dx.
/// </summary>
template <typename T>
--Code changes -Split VariationList into .h/cpp files. --Make VariationList.cpp the only place where the variations files are included from. This speeds up the build, but variations can no longer be directly instantiated. --Variations are no longer exported.
2016-02-18 21:58:24 -05:00
class Julian3DxVariation : public ParametricVariation<T>
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
public:
--Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums.
2016-01-04 19:50:15 -05:00
Julian3DxVariation(T weight = 1.0) : ParametricVariation<T>("julian3Dx", eVariationId::VAR_JULIAN3DX, weight, true, true)
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
Init();
}
PARVARCOPY(Julian3DxVariation)
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
const T z = helper.In.z / m_AbsN;
--User changes -Add new variations: bubbleT3D, crob, hexaplay3D, hexcrop, hexes, hexnix3D, loonie2, loonie3, nBlur, octapol and synth. -Allow for pre/post versions of dc_bubble, dc_cylinder and dc_linear whereas before they were omitted. -When saving a file with multiple embers in it, detect if time values are all the same and if so, start them at zero and increment by 1 for each ember. -Allow for numerous quality increases to be coalesced into one. It will pick up at the end of the current render. -Show selection highlight on variations tree in response to mouse hover. This makes it easier to see for which variation or param the current mouse wheel action will apply. -Make default temporal samples be 100, whereas before it was 1000 which was overkill. -Require the shift key to be held with delete for deleting an ember to prevent it from triggering when the user enters delete in the edit box. -This wasn't otherwise fixable without writing a lot more code. --Bug fixes -EmberGenome was crashing when generating a sequence from a source file with more than 2 embers in it. -EmberGenome was improperly handling the first frame of a merge after the last frame of the loop. -These bugs were due to a previous commit. Revert parts of that commit. -Prevent a zoom value of less than 0 when reading from xml. -Slight optimization of the crescents, and mask variations, if the compiler wasn't doing it already. -Unique file naming was broken because it was looking for _# and the default names ended with -#. -Disallow renaming of an ember in the library tree to an empty string. -Severe bug that prevented some variations from being read correctly from params generated outside this program. -Severe OpenCL randomization bug. The first x coordinates of the first points in the first kernel call of the first ember of a render since the OpenCL renderer object was created were not random and were mostly -1. -Severe bug when populating xform selection distributions that could sometimes cause a crash due to roundoff error. Fix by using double. -Limit the max number of variations in a random ember to MAX_CL_VARS, which is 8. This ensures they'll look the same on CPU and GPU. -Prevent user from saving stylesheet to default.qss, it's a special reserved filename. --Code changes -Generalize using the running sum output point inside of a variation for all cases: pre, reg and post. -Allow for array variables in variations where the address of each element is stored in m_Params. -Qualify all math functions with std:: -No longer use our own Clamp() in OpenCL, instead use the standard clamp(). -Redesign how functions are used in the variations OpenCL code. -Add tests to EmberTester to verify some of the new functionality. -Place more const and override qualifiers on functions where appropriate. -Add a global rand with a lock to be used very sparingly. -Use a map instead of a vector for bad param names in Xml parsing. -Prefix affine interpolation mode defines with "AFFINE_" to make their purpose more clear. -Allow for variations that change state during iteration by sending a separate copy of the ember to each rendering thread. -Implement this same functionality with a local struct in OpenCL. It's members are the total of all variables that need to change state within an ember. -Add Contains() function to Utils.h. -EmberRender: print names of kernels being printed with --dump_kernel option. -Clean up EmberTester to handle some of the recent changes. -Fix various casts. -Replace % 2 with & 1, even though the compiler was likely doing this already. -Add new file Variations06.h to accommodate new variations. -General cleanup.
2015-11-22 17:15:07 -05:00
const T radiusOut = m_Weight * std::pow(helper.m_PrecalcSumSquares + z * z, m_Cn);
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
const T x = m_A * helper.In.x + m_B * helper.In.y + m_E;
const T y = m_C * helper.In.x + m_D * helper.In.y + m_F;
More Linux work. Convert all casts to new style, away from legacy.
2014-12-07 02:51:44 -05:00
const T tempRand = T(int(rand.Frand01<T>() * m_AbsN));
--User changes -Add new variations: bubbleT3D, crob, hexaplay3D, hexcrop, hexes, hexnix3D, loonie2, loonie3, nBlur, octapol and synth. -Allow for pre/post versions of dc_bubble, dc_cylinder and dc_linear whereas before they were omitted. -When saving a file with multiple embers in it, detect if time values are all the same and if so, start them at zero and increment by 1 for each ember. -Allow for numerous quality increases to be coalesced into one. It will pick up at the end of the current render. -Show selection highlight on variations tree in response to mouse hover. This makes it easier to see for which variation or param the current mouse wheel action will apply. -Make default temporal samples be 100, whereas before it was 1000 which was overkill. -Require the shift key to be held with delete for deleting an ember to prevent it from triggering when the user enters delete in the edit box. -This wasn't otherwise fixable without writing a lot more code. --Bug fixes -EmberGenome was crashing when generating a sequence from a source file with more than 2 embers in it. -EmberGenome was improperly handling the first frame of a merge after the last frame of the loop. -These bugs were due to a previous commit. Revert parts of that commit. -Prevent a zoom value of less than 0 when reading from xml. -Slight optimization of the crescents, and mask variations, if the compiler wasn't doing it already. -Unique file naming was broken because it was looking for _# and the default names ended with -#. -Disallow renaming of an ember in the library tree to an empty string. -Severe bug that prevented some variations from being read correctly from params generated outside this program. -Severe OpenCL randomization bug. The first x coordinates of the first points in the first kernel call of the first ember of a render since the OpenCL renderer object was created were not random and were mostly -1. -Severe bug when populating xform selection distributions that could sometimes cause a crash due to roundoff error. Fix by using double. -Limit the max number of variations in a random ember to MAX_CL_VARS, which is 8. This ensures they'll look the same on CPU and GPU. -Prevent user from saving stylesheet to default.qss, it's a special reserved filename. --Code changes -Generalize using the running sum output point inside of a variation for all cases: pre, reg and post. -Allow for array variables in variations where the address of each element is stored in m_Params. -Qualify all math functions with std:: -No longer use our own Clamp() in OpenCL, instead use the standard clamp(). -Redesign how functions are used in the variations OpenCL code. -Add tests to EmberTester to verify some of the new functionality. -Place more const and override qualifiers on functions where appropriate. -Add a global rand with a lock to be used very sparingly. -Use a map instead of a vector for bad param names in Xml parsing. -Prefix affine interpolation mode defines with "AFFINE_" to make their purpose more clear. -Allow for variations that change state during iteration by sending a separate copy of the ember to each rendering thread. -Implement this same functionality with a local struct in OpenCL. It's members are the total of all variables that need to change state within an ember. -Add Contains() function to Utils.h. -EmberRender: print names of kernels being printed with --dump_kernel option. -Clean up EmberTester to handle some of the recent changes. -Fix various casts. -Replace % 2 with & 1, even though the compiler was likely doing this already. -Add new file Variations06.h to accommodate new variations. -General cleanup.
2015-11-22 17:15:07 -05:00
const T alpha = (std::atan2(y, x) + M_2PI * tempRand) / m_Power;
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
const T gamma = radiusOut * helper.m_PrecalcSqrtSumSquares;
--User changes -Add new variations: bubbleT3D, crob, hexaplay3D, hexcrop, hexes, hexnix3D, loonie2, loonie3, nBlur, octapol and synth. -Allow for pre/post versions of dc_bubble, dc_cylinder and dc_linear whereas before they were omitted. -When saving a file with multiple embers in it, detect if time values are all the same and if so, start them at zero and increment by 1 for each ember. -Allow for numerous quality increases to be coalesced into one. It will pick up at the end of the current render. -Show selection highlight on variations tree in response to mouse hover. This makes it easier to see for which variation or param the current mouse wheel action will apply. -Make default temporal samples be 100, whereas before it was 1000 which was overkill. -Require the shift key to be held with delete for deleting an ember to prevent it from triggering when the user enters delete in the edit box. -This wasn't otherwise fixable without writing a lot more code. --Bug fixes -EmberGenome was crashing when generating a sequence from a source file with more than 2 embers in it. -EmberGenome was improperly handling the first frame of a merge after the last frame of the loop. -These bugs were due to a previous commit. Revert parts of that commit. -Prevent a zoom value of less than 0 when reading from xml. -Slight optimization of the crescents, and mask variations, if the compiler wasn't doing it already. -Unique file naming was broken because it was looking for _# and the default names ended with -#. -Disallow renaming of an ember in the library tree to an empty string. -Severe bug that prevented some variations from being read correctly from params generated outside this program. -Severe OpenCL randomization bug. The first x coordinates of the first points in the first kernel call of the first ember of a render since the OpenCL renderer object was created were not random and were mostly -1. -Severe bug when populating xform selection distributions that could sometimes cause a crash due to roundoff error. Fix by using double. -Limit the max number of variations in a random ember to MAX_CL_VARS, which is 8. This ensures they'll look the same on CPU and GPU. -Prevent user from saving stylesheet to default.qss, it's a special reserved filename. --Code changes -Generalize using the running sum output point inside of a variation for all cases: pre, reg and post. -Allow for array variables in variations where the address of each element is stored in m_Params. -Qualify all math functions with std:: -No longer use our own Clamp() in OpenCL, instead use the standard clamp(). -Redesign how functions are used in the variations OpenCL code. -Add tests to EmberTester to verify some of the new functionality. -Place more const and override qualifiers on functions where appropriate. -Add a global rand with a lock to be used very sparingly. -Use a map instead of a vector for bad param names in Xml parsing. -Prefix affine interpolation mode defines with "AFFINE_" to make their purpose more clear. -Allow for variations that change state during iteration by sending a separate copy of the ember to each rendering thread. -Implement this same functionality with a local struct in OpenCL. It's members are the total of all variables that need to change state within an ember. -Add Contains() function to Utils.h. -EmberRender: print names of kernels being printed with --dump_kernel option. -Clean up EmberTester to handle some of the recent changes. -Fix various casts. -Replace % 2 with & 1, even though the compiler was likely doing this already. -Add new file Variations06.h to accommodate new variations. -General cleanup.
2015-11-22 17:15:07 -05:00
helper.Out.x = gamma * std::cos(alpha);
helper.Out.y = gamma * std::sin(alpha);
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
helper.Out.z = radiusOut * z;
}
--Code changes -Better use of const functions. -Ember.pro should not define BUILDING_EMBERCL, instead define BUILDING_EMBER.
2015-06-29 23:13:53 -04:00
virtual string OpenCLString() const override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
ostringstream ss, ss2;
0.4.1.3 Beta 10/14/2014 --User Changes Size is no longer fixed to the window size. Size scaling is done differently in the final render dialog. This fixes several bugs. Remove Xml saving size from settings and options dialog, it no longer applies. Final render can be broken into strips. Set default save path to the desktop if none is found in the settings file. Set default output size to 1920x1080 if none is found in the settings file. --Bug Fixes Better memory size reporting in final render dialog. --Code Changes Migrate to C++11, Qt 5.3.1, and Visual Studio 2013. Change most instances of unsigned int to size_t, and int to intmax_t. Add m_OrigPixPerUnit and m_ScaleType to Ember for scaling purposes. Replace some sprintf_s() calls in XmlToEmber with ostringstream. Move more non-templated members into RendererBase. Add CopyVec() overload that takes a per element function pointer. Add vector Memset(). Replace '&' with '+' instead of "&amp;" in XmlToEmber for much faster parsing. Break strips rendering out into EmberCommon and call from EmberRender and Fractorium. Make AddAndWriteBuffer() just call WriteBuffer(). Make AddAndWriteImage() delete the existing image first before replacing it. Add SetOutputTexture() to RendererCL to support making new textures in response to resize events. Remove multiple return statements in RendererCL, and replace with a bool that tracks results. Add ToDouble(), MakeEnd(), ToString() and Exists() wrappers in Fractorium. Add Size() wrapper in EmberFile. Make QString function arguments const QString&, and string with const string&. Make ShowCritical() wrapper for invoking a message box from another thread. Add combo box to TwoButtonWidget and rename.
2014-10-14 11:53:15 -04:00
intmax_t i = 0, varIndex = IndexInXform();
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
ss2 << "_" << XformIndexInEmber() << "]";
string index = ss2.str();
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
string weight = WeightDefineString();
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
string power = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
string dist = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string a = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string b = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string c = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string d = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string e = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string f = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string absn = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string cn = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
ss << "\t{\n"
<< "\t\tconst real_t z = vIn.z / " << absn << ";\n"
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
<< "\t\tconst real_t radiusOut = " << weight << " * pow(fma(z, z, precalcSumSquares), " << cn << ");\n"
<< "\t\tconst real_t x = fma(" << a << ", vIn.x, fma(" << b << ", vIn.y, " << e << "));\n"
<< "\t\tconst real_t y = fma(" << c << ", vIn.x, fma(" << d << ", vIn.y, " << f << "));\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t\tconst real_t rand = (int)(MwcNext01(mwc) * " << absn << ");\n"
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
<< "\t\tconst real_t alpha = fma(M_2PI, rand, atan2(y, x)) / " << power << ";\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t\tconst real_t gamma = radiusOut * precalcSqrtSumSquares;\n"
<< "\n"
<< "\t\tvOut.x = gamma * cos(alpha);\n"
<< "\t\tvOut.y = gamma * sin(alpha);\n"
<< "\t\tvOut.z = radiusOut * z;\n"
<< "\t}\n";
return ss.str();
}
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
virtual void Precalc() override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
--User changes -Add post_smartcrop. --Bug fixes -Fix bug in crackle. -Wrong point assignment in hexaplay3D, hexnix3D. -Improper Z assignment in rblur. -Fix inconsistency with original in circlecrop. -Put EMBER_ROOT bakc to ./../../../ in default.pri. This is TBD. --Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums. -Optimize crob, nBlur. -Fix weird assignment statement in falloff3. -Cleanup in VarFuncs::SimplexNoise3D(). -Replace fabs() with std::abs(). -General cleanup.
2016-01-12 23:42:12 -05:00
m_AbsN = std::abs(m_Power);
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
m_Cn = (m_Dist / m_Power - 1) / 2;
}
protected:
void Init()
{
string prefix = Prefix();
m_Params.clear();
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
m_Params.push_back(ParamWithName<T>(&m_Dist, prefix + "julian3Dx_dist", 1));
--Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums.
2016-01-04 19:50:15 -05:00
m_Params.push_back(ParamWithName<T>(&m_Power, prefix + "julian3Dx_power", 2, eParamType::INTEGER_NONZERO));
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
m_Params.push_back(ParamWithName<T>(&m_A, prefix + "julian3Dx_a", 1));
m_Params.push_back(ParamWithName<T>(&m_B, prefix + "julian3Dx_b"));
m_Params.push_back(ParamWithName<T>(&m_C, prefix + "julian3Dx_c"));
m_Params.push_back(ParamWithName<T>(&m_D, prefix + "julian3Dx_d", 1));
m_Params.push_back(ParamWithName<T>(&m_E, prefix + "julian3Dx_e"));
m_Params.push_back(ParamWithName<T>(&m_F, prefix + "julian3Dx_f"));
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
m_Params.push_back(ParamWithName<T>(true, &m_AbsN, prefix + "julian3Dx_absn"));//Precalc.
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
m_Params.push_back(ParamWithName<T>(true, &m_Cn, prefix + "julian3Dx_cn"));
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
}
private:
T m_Dist;//Params.
T m_Power;
T m_A;
T m_B;
T m_C;
T m_D;
T m_E;
T m_F;
T m_AbsN;//Precalc.
T m_Cn;
};
/// <summary>
/// fourth.
/// </summary>
template <typename T>
--Code changes -Split VariationList into .h/cpp files. --Make VariationList.cpp the only place where the variations files are included from. This speeds up the build, but variations can no longer be directly instantiated. --Variations are no longer exported.
2016-02-18 21:58:24 -05:00
class FourthVariation : public ParametricVariation<T>
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
public:
--Bug fixes -Fix variations: blob2 (broken on CPU), epispiral, hole. -Fix reading variations from Apophysis: supershape. -Bump render was broken. --Code changes -Flip precalc sina/cosa and apply everywhere by flipping the usage. Flam3 had these reversed and it made the code confusing to read.
2019-12-28 00:04:41 -05:00
FourthVariation(T weight = 1.0) : ParametricVariation<T>("fourth", eVariationId::VAR_FOURTH, weight, true, true, true, false, false)
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
Init();
}
PARVARCOPY(FourthVariation)
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
if (helper.In.x > 0 && helper.In.y > 0)//Quadrant IV: spherical.
{
T r = 1 / helper.m_PrecalcSqrtSumSquares;
--Bug fixes -Fix variations: blob2 (broken on CPU), epispiral, hole. -Fix reading variations from Apophysis: supershape. -Bump render was broken. --Code changes -Flip precalc sina/cosa and apply everywhere by flipping the usage. Flam3 had these reversed and it made the code confusing to read.
2019-12-28 00:04:41 -05:00
helper.Out.x = m_Weight * r * helper.m_PrecalcCosa;
helper.Out.y = m_Weight * r * helper.m_PrecalcSina;
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
}
else if (helper.In.x > 0 && helper.In.y < 0)//Quadrant I: loonie.
{
T r2 = helper.m_PrecalcSumSquares;
if (r2 < m_SqrWeight)
{
Explicitly namespace calls to sqrt and isnan osx/linux have alternative top-level implementations of these methods, so the calls are ambiguous given the - so these should be used explicitly
2015-04-13 07:32:58 -04:00
T r = m_Weight * std::sqrt(m_SqrWeight / r2 - 1);
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
helper.Out.x = r * helper.In.x;
helper.Out.y = r * helper.In.y;
}
else
{
helper.Out.x = m_Weight * helper.In.x;
helper.Out.y = m_Weight * helper.In.y;
}
}
else if (helper.In.x < 0 && helper.In.y > 0)//Quadrant III: susan.
{
T x = helper.In.x - m_X;
T y = helper.In.y + m_Y;
Explicitly namespace calls to sqrt and isnan osx/linux have alternative top-level implementations of these methods, so the calls are ambiguous given the - so these should be used explicitly
2015-04-13 07:32:58 -04:00
T r = std::sqrt(SQR(x) + SQR(y));
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
if (r < m_Weight)
{
--User changes -Add new variations: bubbleT3D, crob, hexaplay3D, hexcrop, hexes, hexnix3D, loonie2, loonie3, nBlur, octapol and synth. -Allow for pre/post versions of dc_bubble, dc_cylinder and dc_linear whereas before they were omitted. -When saving a file with multiple embers in it, detect if time values are all the same and if so, start them at zero and increment by 1 for each ember. -Allow for numerous quality increases to be coalesced into one. It will pick up at the end of the current render. -Show selection highlight on variations tree in response to mouse hover. This makes it easier to see for which variation or param the current mouse wheel action will apply. -Make default temporal samples be 100, whereas before it was 1000 which was overkill. -Require the shift key to be held with delete for deleting an ember to prevent it from triggering when the user enters delete in the edit box. -This wasn't otherwise fixable without writing a lot more code. --Bug fixes -EmberGenome was crashing when generating a sequence from a source file with more than 2 embers in it. -EmberGenome was improperly handling the first frame of a merge after the last frame of the loop. -These bugs were due to a previous commit. Revert parts of that commit. -Prevent a zoom value of less than 0 when reading from xml. -Slight optimization of the crescents, and mask variations, if the compiler wasn't doing it already. -Unique file naming was broken because it was looking for _# and the default names ended with -#. -Disallow renaming of an ember in the library tree to an empty string. -Severe bug that prevented some variations from being read correctly from params generated outside this program. -Severe OpenCL randomization bug. The first x coordinates of the first points in the first kernel call of the first ember of a render since the OpenCL renderer object was created were not random and were mostly -1. -Severe bug when populating xform selection distributions that could sometimes cause a crash due to roundoff error. Fix by using double. -Limit the max number of variations in a random ember to MAX_CL_VARS, which is 8. This ensures they'll look the same on CPU and GPU. -Prevent user from saving stylesheet to default.qss, it's a special reserved filename. --Code changes -Generalize using the running sum output point inside of a variation for all cases: pre, reg and post. -Allow for array variables in variations where the address of each element is stored in m_Params. -Qualify all math functions with std:: -No longer use our own Clamp() in OpenCL, instead use the standard clamp(). -Redesign how functions are used in the variations OpenCL code. -Add tests to EmberTester to verify some of the new functionality. -Place more const and override qualifiers on functions where appropriate. -Add a global rand with a lock to be used very sparingly. -Use a map instead of a vector for bad param names in Xml parsing. -Prefix affine interpolation mode defines with "AFFINE_" to make their purpose more clear. -Allow for variations that change state during iteration by sending a separate copy of the ember to each rendering thread. -Implement this same functionality with a local struct in OpenCL. It's members are the total of all variables that need to change state within an ember. -Add Contains() function to Utils.h. -EmberRender: print names of kernels being printed with --dump_kernel option. -Clean up EmberTester to handle some of the recent changes. -Fix various casts. -Replace % 2 with & 1, even though the compiler was likely doing this already. -Add new file Variations06.h to accommodate new variations. -General cleanup.
2015-11-22 17:15:07 -05:00
T a = std::atan2(y, x) + m_Spin + m_Twist * (m_Weight - r);
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
r *= m_Weight;
--User changes -Add new variations: bubbleT3D, crob, hexaplay3D, hexcrop, hexes, hexnix3D, loonie2, loonie3, nBlur, octapol and synth. -Allow for pre/post versions of dc_bubble, dc_cylinder and dc_linear whereas before they were omitted. -When saving a file with multiple embers in it, detect if time values are all the same and if so, start them at zero and increment by 1 for each ember. -Allow for numerous quality increases to be coalesced into one. It will pick up at the end of the current render. -Show selection highlight on variations tree in response to mouse hover. This makes it easier to see for which variation or param the current mouse wheel action will apply. -Make default temporal samples be 100, whereas before it was 1000 which was overkill. -Require the shift key to be held with delete for deleting an ember to prevent it from triggering when the user enters delete in the edit box. -This wasn't otherwise fixable without writing a lot more code. --Bug fixes -EmberGenome was crashing when generating a sequence from a source file with more than 2 embers in it. -EmberGenome was improperly handling the first frame of a merge after the last frame of the loop. -These bugs were due to a previous commit. Revert parts of that commit. -Prevent a zoom value of less than 0 when reading from xml. -Slight optimization of the crescents, and mask variations, if the compiler wasn't doing it already. -Unique file naming was broken because it was looking for _# and the default names ended with -#. -Disallow renaming of an ember in the library tree to an empty string. -Severe bug that prevented some variations from being read correctly from params generated outside this program. -Severe OpenCL randomization bug. The first x coordinates of the first points in the first kernel call of the first ember of a render since the OpenCL renderer object was created were not random and were mostly -1. -Severe bug when populating xform selection distributions that could sometimes cause a crash due to roundoff error. Fix by using double. -Limit the max number of variations in a random ember to MAX_CL_VARS, which is 8. This ensures they'll look the same on CPU and GPU. -Prevent user from saving stylesheet to default.qss, it's a special reserved filename. --Code changes -Generalize using the running sum output point inside of a variation for all cases: pre, reg and post. -Allow for array variables in variations where the address of each element is stored in m_Params. -Qualify all math functions with std:: -No longer use our own Clamp() in OpenCL, instead use the standard clamp(). -Redesign how functions are used in the variations OpenCL code. -Add tests to EmberTester to verify some of the new functionality. -Place more const and override qualifiers on functions where appropriate. -Add a global rand with a lock to be used very sparingly. -Use a map instead of a vector for bad param names in Xml parsing. -Prefix affine interpolation mode defines with "AFFINE_" to make their purpose more clear. -Allow for variations that change state during iteration by sending a separate copy of the ember to each rendering thread. -Implement this same functionality with a local struct in OpenCL. It's members are the total of all variables that need to change state within an ember. -Add Contains() function to Utils.h. -EmberRender: print names of kernels being printed with --dump_kernel option. -Clean up EmberTester to handle some of the recent changes. -Fix various casts. -Replace % 2 with & 1, even though the compiler was likely doing this already. -Add new file Variations06.h to accommodate new variations. -General cleanup.
2015-11-22 17:15:07 -05:00
helper.Out.x = r * std::cos(a) + m_X;
helper.Out.y = r * std::sin(a) - m_Y;
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
}
else
{
Numerous fixes 0.4.0.5 Beta 07/18/2014 --User Changes Allow for vibrancy values > 1. Add flatten and unflatten menu items. Automatically flatten like Apophysis does. Add plugin and new_linear tags to Xml to be compatible with Apophysis. --Bug Fixes Fix blur, blur3d, bubble, cropn, cross, curl, curl3d, epispiral, ho, julia3d, julia3dz, loonie, mirror_x, mirror_y, mirror_z, rotate_x, sinusoidal, spherical, spherical3d, stripes. Unique filename on final render was completely broken. Two severe OpenCL bugs. Random seeds were biased and fusing was being reset too often leading to results that differ from the CPU. Subtle, but sometimes severe bug in the setup of the xaos weights. Use properly defined epsilon by getting the value from std::numeric_limits, rather than hard coding 1e-6 or 1e-10. Omit incorrect usage of epsilon everywhere. It should not be automatically added to denominators. Rather, it should only be used if the denominator is zero. Force final render progress bars to 100 on completion. Sometimes they didn't seem to make it there. Make variation name and params comparisons be case insensitive. --Code Changes Make ForEach and FindIf wrappers around std::for_each and std::find_if.
2014-07-19 02:33:18 -04:00
r = m_Weight * (1 + m_Space / Zeps(r));
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
helper.Out.x = r * x + m_X;
helper.Out.y = r * y - m_Y;
}
}
else//Quadrant II: Linear.
{
helper.Out.x = m_Weight * helper.In.x;
helper.Out.y = m_Weight * helper.In.y;
}
Numerous fixes 0.4.0.5 Beta 07/18/2014 --User Changes Allow for vibrancy values > 1. Add flatten and unflatten menu items. Automatically flatten like Apophysis does. Add plugin and new_linear tags to Xml to be compatible with Apophysis. --Bug Fixes Fix blur, blur3d, bubble, cropn, cross, curl, curl3d, epispiral, ho, julia3d, julia3dz, loonie, mirror_x, mirror_y, mirror_z, rotate_x, sinusoidal, spherical, spherical3d, stripes. Unique filename on final render was completely broken. Two severe OpenCL bugs. Random seeds were biased and fusing was being reset too often leading to results that differ from the CPU. Subtle, but sometimes severe bug in the setup of the xaos weights. Use properly defined epsilon by getting the value from std::numeric_limits, rather than hard coding 1e-6 or 1e-10. Omit incorrect usage of epsilon everywhere. It should not be automatically added to denominators. Rather, it should only be used if the denominator is zero. Force final render progress bars to 100 on completion. Sometimes they didn't seem to make it there. Make variation name and params comparisons be case insensitive. --Code Changes Make ForEach and FindIf wrappers around std::for_each and std::find_if.
2014-07-19 02:33:18 -04:00
--User changes -Allow for pausing the renderer in the main window. This makes is more efficient when entering many parameters, such as when following a tutorial. -Add support for new variations: erf, gamma, jac_cn, jac_dn, jac_sn, logDB, pressure_wave, pRose3D, splits3D, w, waves2b, x, xerf, y, z. -Inform user of the start and stop of file parsing in EmberAnimate because the files could potentially be very large. -Move the follwing fields to a new table called Animation: Interpolation, Affine Interpolation, Temporal Samples, Temporal Filter Width, Temporal Filter Type. -These currently have no effect on the interactive renderer and instead are used when running flames through EmberGenome to generate sequences, and then animating them in Fractorium or EmberAnimate. -Add new parameter overrides for EmberRender and EmberAnimate which directly assign values to all flames being rendered, rather than scale: --quality --demin --demax --Bug fixes -Left pad instead of right pad names of sequence outputs from EmberGenome. -Unique file naming was broken for files which already had an underscore in them. -Properly report that png is the default format of EmberRender and EmberAnimate output instead of erroneously claiming it was jpg. -Make command line programs search these folders in this order for the palette file: ./ ~/.fractorium ~/.config/fractorium /usr/share/fractorium /usr/local/share/fractorium -Fix possible bad values in hexes. -Better assignment of Z variables. -Fix boarders due to use of poorly implemented rint() function from flam3. Use std::rint() now. -wedge_sph was completely wrong due to having accidentally swapped the mapping of two parameters. -Make juliascope juliascope_power parameter be of type REAL_NONZERO since it's used as a denominator. -Make Z assignment compatible with the originals in: -arch, bcircle, bCollide, bent, bent2, bisplit, blob, blur_linear, blur_square, bMod, boarders, boarders2, bSwirl, bTransform, butterfly, cardioid, cell, circleblur, circlize, circlize2, circus, collideoscope, cos, cosine, cosh, coth, cpow, cpow2, crescents, cropn, csc, csch, curl, curve, dc_gridout, deltaa, diamond, disc2, eclipse, eCollide, edisc, eJulia, elliptic, eMod, eMotion, ennepers, epispiral, ePush, eRotate, eScale, eSwirl, ex, exp, expo, exponential, fan, fdisc, fibonacci, fibonacci2, fisheye, flipcircle, flipy, flower, flux, funnel, glynnia, GlynnSim1, GlynnSim2, GlynnSim3, gridout, handkerchief, heart, hole, idisc, julia, julian2, juliaNab, kaleidoscope, lazyTravis, Lissajous, mask, MobiusN, mobius_strip, modulus, murl, murl2, npolar, ortho, oscilloscope, parabola, perspective, petal, phoenix_julia, pie (was also inconsistent between cpu and gpu), poincare, popcorn, popcorn2, power, pow_block, rational3, rays, rblur, rings, rippled, roundspher, sec, secant2, sigmoid, sin, sineblur, sinh, sinusgrid, sphericaln, spiralwing, spirograph, split, squarize, squirrel, squish, sschecks, starblur, stripes, stwin, super_shape, tan, tancos, tangent, tanh, TwinTrian, twoface, unpolar, waves, wavesn, wedge_julia, whorl, xheart, zblur, zscale. --Code changes -Generalize Variation::PrecalcHelper() and rename to PrePostPrecalcHelper(). --Do the same for the OpenCL version and rename it PrePostPrecalcOpenCLString(). -Rename Variation::m_AssignType to m_PrePostAssignType since it's only relevant to pre/post variations.
2016-01-29 20:02:15 -05:00
helper.Out.z = DefaultZ(helper);
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
}
--Code changes -Better use of const functions. -Ember.pro should not define BUILDING_EMBERCL, instead define BUILDING_EMBER.
2015-06-29 23:13:53 -04:00
virtual string OpenCLString() const override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
ostringstream ss, ss2;
0.4.1.3 Beta 10/14/2014 --User Changes Size is no longer fixed to the window size. Size scaling is done differently in the final render dialog. This fixes several bugs. Remove Xml saving size from settings and options dialog, it no longer applies. Final render can be broken into strips. Set default save path to the desktop if none is found in the settings file. Set default output size to 1920x1080 if none is found in the settings file. --Bug Fixes Better memory size reporting in final render dialog. --Code Changes Migrate to C++11, Qt 5.3.1, and Visual Studio 2013. Change most instances of unsigned int to size_t, and int to intmax_t. Add m_OrigPixPerUnit and m_ScaleType to Ember for scaling purposes. Replace some sprintf_s() calls in XmlToEmber with ostringstream. Move more non-templated members into RendererBase. Add CopyVec() overload that takes a per element function pointer. Add vector Memset(). Replace '&' with '+' instead of "&amp;" in XmlToEmber for much faster parsing. Break strips rendering out into EmberCommon and call from EmberRender and Fractorium. Make AddAndWriteBuffer() just call WriteBuffer(). Make AddAndWriteImage() delete the existing image first before replacing it. Add SetOutputTexture() to RendererCL to support making new textures in response to resize events. Remove multiple return statements in RendererCL, and replace with a bool that tracks results. Add ToDouble(), MakeEnd(), ToString() and Exists() wrappers in Fractorium. Add Size() wrapper in EmberFile. Make QString function arguments const QString&, and string with const string&. Make ShowCritical() wrapper for invoking a message box from another thread. Add combo box to TwoButtonWidget and rename.
2014-10-14 11:53:15 -04:00
intmax_t i = 0, varIndex = IndexInXform();
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
ss2 << "_" << XformIndexInEmber() << "]";
string index = ss2.str();
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
string weight = WeightDefineString();
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
string spin = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string space = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string twist = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string x = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string y = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
string sqrWeight = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
ss << "\t{\n"
--User changes -Add post_smartcrop. --Bug fixes -Fix bug in crackle. -Wrong point assignment in hexaplay3D, hexnix3D. -Improper Z assignment in rblur. -Fix inconsistency with original in circlecrop. -Put EMBER_ROOT bakc to ./../../../ in default.pri. This is TBD. --Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums. -Optimize crob, nBlur. -Fix weird assignment statement in falloff3. -Cleanup in VarFuncs::SimplexNoise3D(). -Replace fabs() with std::abs(). -General cleanup.
2016-01-12 23:42:12 -05:00
<< "\t\tif (vIn.x > 0 && vIn.y > 0)\n"
<< "\t\t{\n"
<< "\t\t real_t r = 1 / precalcSqrtSumSquares;\n"
<< "\n"
--Bug fixes -Fix variations: blob2 (broken on CPU), epispiral, hole. -Fix reading variations from Apophysis: supershape. -Bump render was broken. --Code changes -Flip precalc sina/cosa and apply everywhere by flipping the usage. Flam3 had these reversed and it made the code confusing to read.
2019-12-28 00:04:41 -05:00
<< "\t\t vOut.x = " << weight << " * r * precalcCosa;\n"
<< "\t\t vOut.y = " << weight << " * r * precalcSina;\n"
--User changes -Add post_smartcrop. --Bug fixes -Fix bug in crackle. -Wrong point assignment in hexaplay3D, hexnix3D. -Improper Z assignment in rblur. -Fix inconsistency with original in circlecrop. -Put EMBER_ROOT bakc to ./../../../ in default.pri. This is TBD. --Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums. -Optimize crob, nBlur. -Fix weird assignment statement in falloff3. -Cleanup in VarFuncs::SimplexNoise3D(). -Replace fabs() with std::abs(). -General cleanup.
2016-01-12 23:42:12 -05:00
<< "\t\t}\n"
<< "\t\telse if (vIn.x > 0 && vIn.y < 0)\n"
<< "\t\t{\n"
<< "\t\t real_t r2 = precalcSumSquares;\n"
<< "\n"
<< "\t\t if (r2 < " << sqrWeight << ")\n"
<< "\t\t {\n"
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
<< "\t\t real_t r = " << weight << " * sqrt(" << sqrWeight << " / r2 - 1);\n"
--User changes -Add post_smartcrop. --Bug fixes -Fix bug in crackle. -Wrong point assignment in hexaplay3D, hexnix3D. -Improper Z assignment in rblur. -Fix inconsistency with original in circlecrop. -Put EMBER_ROOT bakc to ./../../../ in default.pri. This is TBD. --Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums. -Optimize crob, nBlur. -Fix weird assignment statement in falloff3. -Cleanup in VarFuncs::SimplexNoise3D(). -Replace fabs() with std::abs(). -General cleanup.
2016-01-12 23:42:12 -05:00
<< "\n"
<< "\t\t vOut.x = r * vIn.x;\n"
<< "\t\t vOut.y = r * vIn.y;\n"
<< "\t\t }\n"
<< "\t\t else\n"
<< "\t\t {\n"
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
<< "\t\t vOut.x = " << weight << " * vIn.x;\n"
<< "\t\t vOut.y = " << weight << " * vIn.y;\n"
--User changes -Add post_smartcrop. --Bug fixes -Fix bug in crackle. -Wrong point assignment in hexaplay3D, hexnix3D. -Improper Z assignment in rblur. -Fix inconsistency with original in circlecrop. -Put EMBER_ROOT bakc to ./../../../ in default.pri. This is TBD. --Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums. -Optimize crob, nBlur. -Fix weird assignment statement in falloff3. -Cleanup in VarFuncs::SimplexNoise3D(). -Replace fabs() with std::abs(). -General cleanup.
2016-01-12 23:42:12 -05:00
<< "\t\t }\n"
<< "\t\t}\n"
<< "\t\telse if (vIn.x < 0 && vIn.y > 0)\n"
<< "\t\t{\n"
<< "\t\t real_t x = vIn.x - " << x << ";\n"
<< "\t\t real_t y = vIn.y + " << y << ";\n"
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
<< "\t\t real_t r = sqrt(fma(x, x, SQR(y)));\n"
--User changes -Add post_smartcrop. --Bug fixes -Fix bug in crackle. -Wrong point assignment in hexaplay3D, hexnix3D. -Improper Z assignment in rblur. -Fix inconsistency with original in circlecrop. -Put EMBER_ROOT bakc to ./../../../ in default.pri. This is TBD. --Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums. -Optimize crob, nBlur. -Fix weird assignment statement in falloff3. -Cleanup in VarFuncs::SimplexNoise3D(). -Replace fabs() with std::abs(). -General cleanup.
2016-01-12 23:42:12 -05:00
<< "\n"
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
<< "\t\t if (r < " << weight << ")\n"
--User changes -Add post_smartcrop. --Bug fixes -Fix bug in crackle. -Wrong point assignment in hexaplay3D, hexnix3D. -Improper Z assignment in rblur. -Fix inconsistency with original in circlecrop. -Put EMBER_ROOT bakc to ./../../../ in default.pri. This is TBD. --Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums. -Optimize crob, nBlur. -Fix weird assignment statement in falloff3. -Cleanup in VarFuncs::SimplexNoise3D(). -Replace fabs() with std::abs(). -General cleanup.
2016-01-12 23:42:12 -05:00
<< "\t\t {\n"
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
<< "\t\t real_t a = fma(" << twist << ", " << weight << " - r, atan2(y, x) + " << spin << ");\n"
--User changes -Add post_smartcrop. --Bug fixes -Fix bug in crackle. -Wrong point assignment in hexaplay3D, hexnix3D. -Improper Z assignment in rblur. -Fix inconsistency with original in circlecrop. -Put EMBER_ROOT bakc to ./../../../ in default.pri. This is TBD. --Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums. -Optimize crob, nBlur. -Fix weird assignment statement in falloff3. -Cleanup in VarFuncs::SimplexNoise3D(). -Replace fabs() with std::abs(). -General cleanup.
2016-01-12 23:42:12 -05:00
<< "\n"
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
<< "\t\t r *= " << weight << ";\n"
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
<< "\t\t vOut.x = fma(r, cos(a), " << x << ");\n"
<< "\t\t vOut.y = fma(r, sin(a), -" << y << ");\n"
--User changes -Add post_smartcrop. --Bug fixes -Fix bug in crackle. -Wrong point assignment in hexaplay3D, hexnix3D. -Improper Z assignment in rblur. -Fix inconsistency with original in circlecrop. -Put EMBER_ROOT bakc to ./../../../ in default.pri. This is TBD. --Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums. -Optimize crob, nBlur. -Fix weird assignment statement in falloff3. -Cleanup in VarFuncs::SimplexNoise3D(). -Replace fabs() with std::abs(). -General cleanup.
2016-01-12 23:42:12 -05:00
<< "\t\t }\n"
<< "\t\t else\n"
<< "\t\t {\n"
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
<< "\t\t r = " << weight << " * (1 + " << space << " / Zeps(r));\n"
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
<< "\t\t vOut.x = fma(r, x, " << x << ");\n"
<< "\t\t vOut.y = fma(r, y, -" << y << ");\n"
--User changes -Add post_smartcrop. --Bug fixes -Fix bug in crackle. -Wrong point assignment in hexaplay3D, hexnix3D. -Improper Z assignment in rblur. -Fix inconsistency with original in circlecrop. -Put EMBER_ROOT bakc to ./../../../ in default.pri. This is TBD. --Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums. -Optimize crob, nBlur. -Fix weird assignment statement in falloff3. -Cleanup in VarFuncs::SimplexNoise3D(). -Replace fabs() with std::abs(). -General cleanup.
2016-01-12 23:42:12 -05:00
<< "\t\t }\n"
<< "\t\t}\n"
<< "\t\telse\n"
<< "\t\t{\n"
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
<< "\t\t vOut.x = " << weight << " * vIn.x;\n"
<< "\t\t vOut.y = " << weight << " * vIn.y;\n"
--User changes -Add post_smartcrop. --Bug fixes -Fix bug in crackle. -Wrong point assignment in hexaplay3D, hexnix3D. -Improper Z assignment in rblur. -Fix inconsistency with original in circlecrop. -Put EMBER_ROOT bakc to ./../../../ in default.pri. This is TBD. --Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums. -Optimize crob, nBlur. -Fix weird assignment statement in falloff3. -Cleanup in VarFuncs::SimplexNoise3D(). -Replace fabs() with std::abs(). -General cleanup.
2016-01-12 23:42:12 -05:00
<< "\t\t}\n"
<< "\n"
--User changes -Allow for pausing the renderer in the main window. This makes is more efficient when entering many parameters, such as when following a tutorial. -Add support for new variations: erf, gamma, jac_cn, jac_dn, jac_sn, logDB, pressure_wave, pRose3D, splits3D, w, waves2b, x, xerf, y, z. -Inform user of the start and stop of file parsing in EmberAnimate because the files could potentially be very large. -Move the follwing fields to a new table called Animation: Interpolation, Affine Interpolation, Temporal Samples, Temporal Filter Width, Temporal Filter Type. -These currently have no effect on the interactive renderer and instead are used when running flames through EmberGenome to generate sequences, and then animating them in Fractorium or EmberAnimate. -Add new parameter overrides for EmberRender and EmberAnimate which directly assign values to all flames being rendered, rather than scale: --quality --demin --demax --Bug fixes -Left pad instead of right pad names of sequence outputs from EmberGenome. -Unique file naming was broken for files which already had an underscore in them. -Properly report that png is the default format of EmberRender and EmberAnimate output instead of erroneously claiming it was jpg. -Make command line programs search these folders in this order for the palette file: ./ ~/.fractorium ~/.config/fractorium /usr/share/fractorium /usr/local/share/fractorium -Fix possible bad values in hexes. -Better assignment of Z variables. -Fix boarders due to use of poorly implemented rint() function from flam3. Use std::rint() now. -wedge_sph was completely wrong due to having accidentally swapped the mapping of two parameters. -Make juliascope juliascope_power parameter be of type REAL_NONZERO since it's used as a denominator. -Make Z assignment compatible with the originals in: -arch, bcircle, bCollide, bent, bent2, bisplit, blob, blur_linear, blur_square, bMod, boarders, boarders2, bSwirl, bTransform, butterfly, cardioid, cell, circleblur, circlize, circlize2, circus, collideoscope, cos, cosine, cosh, coth, cpow, cpow2, crescents, cropn, csc, csch, curl, curve, dc_gridout, deltaa, diamond, disc2, eclipse, eCollide, edisc, eJulia, elliptic, eMod, eMotion, ennepers, epispiral, ePush, eRotate, eScale, eSwirl, ex, exp, expo, exponential, fan, fdisc, fibonacci, fibonacci2, fisheye, flipcircle, flipy, flower, flux, funnel, glynnia, GlynnSim1, GlynnSim2, GlynnSim3, gridout, handkerchief, heart, hole, idisc, julia, julian2, juliaNab, kaleidoscope, lazyTravis, Lissajous, mask, MobiusN, mobius_strip, modulus, murl, murl2, npolar, ortho, oscilloscope, parabola, perspective, petal, phoenix_julia, pie (was also inconsistent between cpu and gpu), poincare, popcorn, popcorn2, power, pow_block, rational3, rays, rblur, rings, rippled, roundspher, sec, secant2, sigmoid, sin, sineblur, sinh, sinusgrid, sphericaln, spiralwing, spirograph, split, squarize, squirrel, squish, sschecks, starblur, stripes, stwin, super_shape, tan, tancos, tangent, tanh, TwinTrian, twoface, unpolar, waves, wavesn, wedge_julia, whorl, xheart, zblur, zscale. --Code changes -Generalize Variation::PrecalcHelper() and rename to PrePostPrecalcHelper(). --Do the same for the OpenCL version and rename it PrePostPrecalcOpenCLString(). -Rename Variation::m_AssignType to m_PrePostAssignType since it's only relevant to pre/post variations.
2016-01-29 20:02:15 -05:00
<< "\t\tvOut.z = " << DefaultZCl()
--User changes -Add post_smartcrop. --Bug fixes -Fix bug in crackle. -Wrong point assignment in hexaplay3D, hexnix3D. -Improper Z assignment in rblur. -Fix inconsistency with original in circlecrop. -Put EMBER_ROOT bakc to ./../../../ in default.pri. This is TBD. --Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums. -Optimize crob, nBlur. -Fix weird assignment statement in falloff3. -Cleanup in VarFuncs::SimplexNoise3D(). -Replace fabs() with std::abs(). -General cleanup.
2016-01-12 23:42:12 -05:00
<< "\t}\n";
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
return ss.str();
}
--User changes -Add new variations: bubbleT3D, crob, hexaplay3D, hexcrop, hexes, hexnix3D, loonie2, loonie3, nBlur, octapol and synth. -Allow for pre/post versions of dc_bubble, dc_cylinder and dc_linear whereas before they were omitted. -When saving a file with multiple embers in it, detect if time values are all the same and if so, start them at zero and increment by 1 for each ember. -Allow for numerous quality increases to be coalesced into one. It will pick up at the end of the current render. -Show selection highlight on variations tree in response to mouse hover. This makes it easier to see for which variation or param the current mouse wheel action will apply. -Make default temporal samples be 100, whereas before it was 1000 which was overkill. -Require the shift key to be held with delete for deleting an ember to prevent it from triggering when the user enters delete in the edit box. -This wasn't otherwise fixable without writing a lot more code. --Bug fixes -EmberGenome was crashing when generating a sequence from a source file with more than 2 embers in it. -EmberGenome was improperly handling the first frame of a merge after the last frame of the loop. -These bugs were due to a previous commit. Revert parts of that commit. -Prevent a zoom value of less than 0 when reading from xml. -Slight optimization of the crescents, and mask variations, if the compiler wasn't doing it already. -Unique file naming was broken because it was looking for _# and the default names ended with -#. -Disallow renaming of an ember in the library tree to an empty string. -Severe bug that prevented some variations from being read correctly from params generated outside this program. -Severe OpenCL randomization bug. The first x coordinates of the first points in the first kernel call of the first ember of a render since the OpenCL renderer object was created were not random and were mostly -1. -Severe bug when populating xform selection distributions that could sometimes cause a crash due to roundoff error. Fix by using double. -Limit the max number of variations in a random ember to MAX_CL_VARS, which is 8. This ensures they'll look the same on CPU and GPU. -Prevent user from saving stylesheet to default.qss, it's a special reserved filename. --Code changes -Generalize using the running sum output point inside of a variation for all cases: pre, reg and post. -Allow for array variables in variations where the address of each element is stored in m_Params. -Qualify all math functions with std:: -No longer use our own Clamp() in OpenCL, instead use the standard clamp(). -Redesign how functions are used in the variations OpenCL code. -Add tests to EmberTester to verify some of the new functionality. -Place more const and override qualifiers on functions where appropriate. -Add a global rand with a lock to be used very sparingly. -Use a map instead of a vector for bad param names in Xml parsing. -Prefix affine interpolation mode defines with "AFFINE_" to make their purpose more clear. -Allow for variations that change state during iteration by sending a separate copy of the ember to each rendering thread. -Implement this same functionality with a local struct in OpenCL. It's members are the total of all variables that need to change state within an ember. -Add Contains() function to Utils.h. -EmberRender: print names of kernels being printed with --dump_kernel option. -Clean up EmberTester to handle some of the recent changes. -Fix various casts. -Replace % 2 with & 1, even though the compiler was likely doing this already. -Add new file Variations06.h to accommodate new variations. -General cleanup.
2015-11-22 17:15:07 -05:00
virtual vector<string> OpenCLGlobalFuncNames() const override
{
return vector<string> { "Zeps" };
}
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
virtual void Precalc() override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
m_SqrWeight = SQR(m_Weight);
}
protected:
void Init()
{
string prefix = Prefix();
m_Params.clear();
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
m_Params.push_back(ParamWithName<T>(&m_Spin, prefix + "fourth_spin", T(M_PI), eParamType::REAL_CYCLIC, 0, M_2PI));
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
m_Params.push_back(ParamWithName<T>(&m_Space, prefix + "fourth_space"));
m_Params.push_back(ParamWithName<T>(&m_Twist, prefix + "fourth_twist"));
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
m_Params.push_back(ParamWithName<T>(&m_X, prefix + "fourth_x"));
m_Params.push_back(ParamWithName<T>(&m_Y, prefix + "fourth_y"));
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
m_Params.push_back(ParamWithName<T>(true, &m_SqrWeight, prefix + "fourth_sqr_weight"));//Precalc.
}
private:
T m_Spin;
T m_Space;
T m_Twist;
T m_X;
T m_Y;
T m_SqrWeight;//Precalc.
};
/// <summary>
/// mobiq.
/// </summary>
template <typename T>
--Code changes -Split VariationList into .h/cpp files. --Make VariationList.cpp the only place where the variations files are included from. This speeds up the build, but variations can no longer be directly instantiated. --Variations are no longer exported.
2016-02-18 21:58:24 -05:00
class MobiqVariation : public ParametricVariation<T>
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
public:
--Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums.
2016-01-04 19:50:15 -05:00
MobiqVariation(T weight = 1.0) : ParametricVariation<T>("mobiq", eVariationId::VAR_MOBIQ, weight)
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
Init();
}
PARVARCOPY(MobiqVariation)
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
const T t1 = m_At;
const T t2 = helper.In.x;
const T t3 = m_Bt;
const T t4 = m_Ct;
const T t5 = m_Dt;
const T x1 = m_Ax;
const T x2 = helper.In.y;
const T x3 = m_Bx;
const T x4 = m_Cx;
const T x5 = m_Dx;
const T y1 = m_Ay;
const T y2 = helper.In.z;
const T y3 = m_By;
const T y4 = m_Cy;
const T y5 = m_Dy;
const T z1 = m_Az;
const T z3 = m_Bz;
const T z4 = m_Cz;
const T z5 = m_Dz;
T nt = t1 * t2 - x1 * x2 - y1 * y2 + t3;
T nx = t1 * x2 + x1 * t2 - z1 * y2 + x3;
T ny = t1 * y2 + y1 * t2 + z1 * x2 + y3;
T nz = z1 * t2 + x1 * y2 - y1 * x2 + z3;
T dt = t4 * t2 - x4 * x2 - y4 * y2 + t5;
T dx = t4 * x2 + x4 * t2 - z4 * y2 + x5;
T dy = t4 * y2 + y4 * t2 + z4 * x2 + y5;
T dz = z4 * t2 + x4 * y2 - y4 * x2 + z5;
T ni = m_Weight / (SQR(dt) + SQR(dx) + SQR(dy) + SQR(dz));
helper.Out.x = (nt * dt + nx * dx + ny * dy + nz * dz) * ni;
helper.Out.y = (nx * dt - nt * dx - ny * dz + nz * dy) * ni;
helper.Out.z = (ny * dt - nt * dy - nz * dx + nx * dz) * ni;
}
--Code changes -Better use of const functions. -Ember.pro should not define BUILDING_EMBERCL, instead define BUILDING_EMBER.
2015-06-29 23:13:53 -04:00
virtual string OpenCLString() const override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
ostringstream ss, ss2;
0.4.1.3 Beta 10/14/2014 --User Changes Size is no longer fixed to the window size. Size scaling is done differently in the final render dialog. This fixes several bugs. Remove Xml saving size from settings and options dialog, it no longer applies. Final render can be broken into strips. Set default save path to the desktop if none is found in the settings file. Set default output size to 1920x1080 if none is found in the settings file. --Bug Fixes Better memory size reporting in final render dialog. --Code Changes Migrate to C++11, Qt 5.3.1, and Visual Studio 2013. Change most instances of unsigned int to size_t, and int to intmax_t. Add m_OrigPixPerUnit and m_ScaleType to Ember for scaling purposes. Replace some sprintf_s() calls in XmlToEmber with ostringstream. Move more non-templated members into RendererBase. Add CopyVec() overload that takes a per element function pointer. Add vector Memset(). Replace '&' with '+' instead of "&amp;" in XmlToEmber for much faster parsing. Break strips rendering out into EmberCommon and call from EmberRender and Fractorium. Make AddAndWriteBuffer() just call WriteBuffer(). Make AddAndWriteImage() delete the existing image first before replacing it. Add SetOutputTexture() to RendererCL to support making new textures in response to resize events. Remove multiple return statements in RendererCL, and replace with a bool that tracks results. Add ToDouble(), MakeEnd(), ToString() and Exists() wrappers in Fractorium. Add Size() wrapper in EmberFile. Make QString function arguments const QString&, and string with const string&. Make ShowCritical() wrapper for invoking a message box from another thread. Add combo box to TwoButtonWidget and rename.
2014-10-14 11:53:15 -04:00
intmax_t i = 0, varIndex = IndexInXform();
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
ss2 << "_" << XformIndexInEmber() << "]";
string index = ss2.str();
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
string weight = WeightDefineString();
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
string at = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string ax = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string ay = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string az = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string bt = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string bx = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string by = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string bz = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string ct = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string cx = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string cy = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string cz = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string dt = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string dx = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string dy = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string dz = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
ss << "\t{\n"
<< "\t\tconst real_t t1 = " << at << ";\n"
<< "\t\tconst real_t t2 = vIn.x;\n"
<< "\t\tconst real_t t3 = " << bt << ";\n"
<< "\t\tconst real_t t4 = " << ct << ";\n"
<< "\t\tconst real_t t5 = " << dt << ";\n"
<< "\t\tconst real_t x1 = " << ax << ";\n"
<< "\t\tconst real_t x2 = vIn.y;\n"
<< "\t\tconst real_t x3 = " << bx << ";\n"
<< "\t\tconst real_t x4 = " << cx << ";\n"
<< "\t\tconst real_t x5 = " << dx << ";\n"
<< "\t\tconst real_t y1 = " << ay << ";\n"
<< "\t\tconst real_t y2 = vIn.z;\n"
<< "\t\tconst real_t y3 = " << by << ";\n"
<< "\t\tconst real_t y4 = " << cy << ";\n"
<< "\t\tconst real_t y5 = " << dy << ";\n"
<< "\t\tconst real_t z1 = " << az << ";\n"
<< "\t\tconst real_t z3 = " << bz << ";\n"
<< "\t\tconst real_t z4 = " << cz << ";\n"
<< "\t\tconst real_t z5 = " << dz << ";\n"
<< "\n"
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
<< "\t\treal_t nt = fma(t1, t2, -(x1 * x2)) - y1 * y2 + t3;\n"
<< "\t\treal_t nx = fma(t1, x2, x1 * t2) - z1 * y2 + x3;\n"
<< "\t\treal_t ny = fma(t1, y2, fma(y1, t2, fma(z1, x2, y3)));\n"
<< "\t\treal_t nz = fma(z1, t2, x1 * y2) - y1 * x2 + z3;\n"
<< "\t\treal_t dt = fma(t4, t2, -(x4 * x2)) - y4 * y2 + t5;\n"
<< "\t\treal_t dx = fma(t4, x2, x4 * t2) - z4 * y2 + x5;\n"
<< "\t\treal_t dy = fma(t4, y2, fma(y4, t2, fma(z4, x2, y5)));\n"
<< "\t\treal_t dz = fma(z4, t2, x4 * y2) - y4 * x2 + z5;\n"
<< "\t\treal_t ni = " << weight << " / fma(dt, dt, fma(dx, dx, fma(dy, dy, SQR(dz))));\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\n"
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
<< "\t\tvOut.x = fma(nt, dt, fma(nx, dx, fma(ny, dy, nz * dz))) * ni;\n"
<< "\t\tvOut.y = (fma(nx, dt, -(nt * dx)) - ny * dz + nz * dy) * ni;\n"
<< "\t\tvOut.z = (fma(ny, dt, -(nt * dy)) - nz * dx + nx * dz) * ni;\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t}\n";
return ss.str();
}
protected:
void Init()
{
string prefix = Prefix();
m_Params.clear();
m_Params.push_back(ParamWithName<T>(&m_At, prefix + "mobiq_at", 1));
m_Params.push_back(ParamWithName<T>(&m_Ax, prefix + "mobiq_ax"));
m_Params.push_back(ParamWithName<T>(&m_Ay, prefix + "mobiq_ay"));
m_Params.push_back(ParamWithName<T>(&m_Az, prefix + "mobiq_az"));
m_Params.push_back(ParamWithName<T>(&m_Bt, prefix + "mobiq_bt"));
m_Params.push_back(ParamWithName<T>(&m_Bx, prefix + "mobiq_bx"));
m_Params.push_back(ParamWithName<T>(&m_By, prefix + "mobiq_by"));
m_Params.push_back(ParamWithName<T>(&m_Bz, prefix + "mobiq_bz"));
m_Params.push_back(ParamWithName<T>(&m_Ct, prefix + "mobiq_ct"));
m_Params.push_back(ParamWithName<T>(&m_Cx, prefix + "mobiq_cx"));
m_Params.push_back(ParamWithName<T>(&m_Cy, prefix + "mobiq_cy"));
m_Params.push_back(ParamWithName<T>(&m_Cz, prefix + "mobiq_cz"));
m_Params.push_back(ParamWithName<T>(&m_Dt, prefix + "mobiq_dt", 1));
m_Params.push_back(ParamWithName<T>(&m_Dx, prefix + "mobiq_dx"));
m_Params.push_back(ParamWithName<T>(&m_Dy, prefix + "mobiq_dy"));
m_Params.push_back(ParamWithName<T>(&m_Dz, prefix + "mobiq_dz"));
}
private:
T m_At;
T m_Ax;
T m_Ay;
T m_Az;
T m_Bt;
T m_Bx;
T m_By;
T m_Bz;
T m_Ct;
T m_Cx;
T m_Cy;
T m_Cz;
T m_Dt;
T m_Dx;
T m_Dy;
T m_Dz;
};
/// <summary>
/// spherivoid.
/// </summary>
template <typename T>
--Code changes -Split VariationList into .h/cpp files. --Make VariationList.cpp the only place where the variations files are included from. This speeds up the build, but variations can no longer be directly instantiated. --Variations are no longer exported.
2016-02-18 21:58:24 -05:00
class SpherivoidVariation : public ParametricVariation<T>
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
public:
--Bug fixes -Fix variations: blob2 (broken on CPU), epispiral, hole. -Fix reading variations from Apophysis: supershape. -Bump render was broken. --Code changes -Flip precalc sina/cosa and apply everywhere by flipping the usage. Flam3 had these reversed and it made the code confusing to read.
2019-12-28 00:04:41 -05:00
SpherivoidVariation(T weight = 1.0) : ParametricVariation<T>("spherivoid", eVariationId::VAR_SPHERIVOID, weight, true, true, true, false, false)
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
Init();
}
PARVARCOPY(SpherivoidVariation)
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
--Code changes -Make usage of VarFuncs more consistent by moving some global functions there. -Make CircleTrans1 a little safer by calling Zeps() on m_Sc during precalc. -Begin prepping for 0.9.9.5 release.
2016-03-04 21:54:06 -05:00
const T zr = VarFuncs<T>::Hypot(helper.In.z, helper.m_PrecalcSqrtSumSquares);
--User changes -Add new variations: bubbleT3D, crob, hexaplay3D, hexcrop, hexes, hexnix3D, loonie2, loonie3, nBlur, octapol and synth. -Allow for pre/post versions of dc_bubble, dc_cylinder and dc_linear whereas before they were omitted. -When saving a file with multiple embers in it, detect if time values are all the same and if so, start them at zero and increment by 1 for each ember. -Allow for numerous quality increases to be coalesced into one. It will pick up at the end of the current render. -Show selection highlight on variations tree in response to mouse hover. This makes it easier to see for which variation or param the current mouse wheel action will apply. -Make default temporal samples be 100, whereas before it was 1000 which was overkill. -Require the shift key to be held with delete for deleting an ember to prevent it from triggering when the user enters delete in the edit box. -This wasn't otherwise fixable without writing a lot more code. --Bug fixes -EmberGenome was crashing when generating a sequence from a source file with more than 2 embers in it. -EmberGenome was improperly handling the first frame of a merge after the last frame of the loop. -These bugs were due to a previous commit. Revert parts of that commit. -Prevent a zoom value of less than 0 when reading from xml. -Slight optimization of the crescents, and mask variations, if the compiler wasn't doing it already. -Unique file naming was broken because it was looking for _# and the default names ended with -#. -Disallow renaming of an ember in the library tree to an empty string. -Severe bug that prevented some variations from being read correctly from params generated outside this program. -Severe OpenCL randomization bug. The first x coordinates of the first points in the first kernel call of the first ember of a render since the OpenCL renderer object was created were not random and were mostly -1. -Severe bug when populating xform selection distributions that could sometimes cause a crash due to roundoff error. Fix by using double. -Limit the max number of variations in a random ember to MAX_CL_VARS, which is 8. This ensures they'll look the same on CPU and GPU. -Prevent user from saving stylesheet to default.qss, it's a special reserved filename. --Code changes -Generalize using the running sum output point inside of a variation for all cases: pre, reg and post. -Allow for array variables in variations where the address of each element is stored in m_Params. -Qualify all math functions with std:: -No longer use our own Clamp() in OpenCL, instead use the standard clamp(). -Redesign how functions are used in the variations OpenCL code. -Add tests to EmberTester to verify some of the new functionality. -Place more const and override qualifiers on functions where appropriate. -Add a global rand with a lock to be used very sparingly. -Use a map instead of a vector for bad param names in Xml parsing. -Prefix affine interpolation mode defines with "AFFINE_" to make their purpose more clear. -Allow for variations that change state during iteration by sending a separate copy of the ember to each rendering thread. -Implement this same functionality with a local struct in OpenCL. It's members are the total of all variables that need to change state within an ember. -Add Contains() function to Utils.h. -EmberRender: print names of kernels being printed with --dump_kernel option. -Clean up EmberTester to handle some of the recent changes. -Fix various casts. -Replace % 2 with & 1, even though the compiler was likely doing this already. -Add new file Variations06.h to accommodate new variations. -General cleanup.
2015-11-22 17:15:07 -05:00
const T phi = std::acos(Clamp<T>(helper.In.z / zr, -1, 1));
const T ps = std::sin(phi);
const T pc = std::cos(phi);
--Bug fixes -Fix variations: blob2 (broken on CPU), epispiral, hole. -Fix reading variations from Apophysis: supershape. -Bump render was broken. --Code changes -Flip precalc sina/cosa and apply everywhere by flipping the usage. Flam3 had these reversed and it made the code confusing to read.
2019-12-28 00:04:41 -05:00
helper.Out.x = m_Weight * helper.m_PrecalcCosa * ps * (zr + m_Radius);
helper.Out.y = m_Weight * helper.m_PrecalcSina * ps * (zr + m_Radius);
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
helper.Out.z = m_Weight * pc * (zr + m_Radius);
}
--Code changes -Better use of const functions. -Ember.pro should not define BUILDING_EMBERCL, instead define BUILDING_EMBER.
2015-06-29 23:13:53 -04:00
virtual string OpenCLString() const override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
ostringstream ss, ss2;
0.4.1.3 Beta 10/14/2014 --User Changes Size is no longer fixed to the window size. Size scaling is done differently in the final render dialog. This fixes several bugs. Remove Xml saving size from settings and options dialog, it no longer applies. Final render can be broken into strips. Set default save path to the desktop if none is found in the settings file. Set default output size to 1920x1080 if none is found in the settings file. --Bug Fixes Better memory size reporting in final render dialog. --Code Changes Migrate to C++11, Qt 5.3.1, and Visual Studio 2013. Change most instances of unsigned int to size_t, and int to intmax_t. Add m_OrigPixPerUnit and m_ScaleType to Ember for scaling purposes. Replace some sprintf_s() calls in XmlToEmber with ostringstream. Move more non-templated members into RendererBase. Add CopyVec() overload that takes a per element function pointer. Add vector Memset(). Replace '&' with '+' instead of "&amp;" in XmlToEmber for much faster parsing. Break strips rendering out into EmberCommon and call from EmberRender and Fractorium. Make AddAndWriteBuffer() just call WriteBuffer(). Make AddAndWriteImage() delete the existing image first before replacing it. Add SetOutputTexture() to RendererCL to support making new textures in response to resize events. Remove multiple return statements in RendererCL, and replace with a bool that tracks results. Add ToDouble(), MakeEnd(), ToString() and Exists() wrappers in Fractorium. Add Size() wrapper in EmberFile. Make QString function arguments const QString&, and string with const string&. Make ShowCritical() wrapper for invoking a message box from another thread. Add combo box to TwoButtonWidget and rename.
2014-10-14 11:53:15 -04:00
intmax_t i = 0, varIndex = IndexInXform();
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
ss2 << "_" << XformIndexInEmber() << "]";
string index = ss2.str();
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
string weight = WeightDefineString();
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
string radius = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
ss << "\t{\n"
<< "\t\tconst real_t zr = Hypot(vIn.z, precalcSqrtSumSquares);\n"
--User changes -Add new variations: bubbleT3D, crob, hexaplay3D, hexcrop, hexes, hexnix3D, loonie2, loonie3, nBlur, octapol and synth. -Allow for pre/post versions of dc_bubble, dc_cylinder and dc_linear whereas before they were omitted. -When saving a file with multiple embers in it, detect if time values are all the same and if so, start them at zero and increment by 1 for each ember. -Allow for numerous quality increases to be coalesced into one. It will pick up at the end of the current render. -Show selection highlight on variations tree in response to mouse hover. This makes it easier to see for which variation or param the current mouse wheel action will apply. -Make default temporal samples be 100, whereas before it was 1000 which was overkill. -Require the shift key to be held with delete for deleting an ember to prevent it from triggering when the user enters delete in the edit box. -This wasn't otherwise fixable without writing a lot more code. --Bug fixes -EmberGenome was crashing when generating a sequence from a source file with more than 2 embers in it. -EmberGenome was improperly handling the first frame of a merge after the last frame of the loop. -These bugs were due to a previous commit. Revert parts of that commit. -Prevent a zoom value of less than 0 when reading from xml. -Slight optimization of the crescents, and mask variations, if the compiler wasn't doing it already. -Unique file naming was broken because it was looking for _# and the default names ended with -#. -Disallow renaming of an ember in the library tree to an empty string. -Severe bug that prevented some variations from being read correctly from params generated outside this program. -Severe OpenCL randomization bug. The first x coordinates of the first points in the first kernel call of the first ember of a render since the OpenCL renderer object was created were not random and were mostly -1. -Severe bug when populating xform selection distributions that could sometimes cause a crash due to roundoff error. Fix by using double. -Limit the max number of variations in a random ember to MAX_CL_VARS, which is 8. This ensures they'll look the same on CPU and GPU. -Prevent user from saving stylesheet to default.qss, it's a special reserved filename. --Code changes -Generalize using the running sum output point inside of a variation for all cases: pre, reg and post. -Allow for array variables in variations where the address of each element is stored in m_Params. -Qualify all math functions with std:: -No longer use our own Clamp() in OpenCL, instead use the standard clamp(). -Redesign how functions are used in the variations OpenCL code. -Add tests to EmberTester to verify some of the new functionality. -Place more const and override qualifiers on functions where appropriate. -Add a global rand with a lock to be used very sparingly. -Use a map instead of a vector for bad param names in Xml parsing. -Prefix affine interpolation mode defines with "AFFINE_" to make their purpose more clear. -Allow for variations that change state during iteration by sending a separate copy of the ember to each rendering thread. -Implement this same functionality with a local struct in OpenCL. It's members are the total of all variables that need to change state within an ember. -Add Contains() function to Utils.h. -EmberRender: print names of kernels being printed with --dump_kernel option. -Clean up EmberTester to handle some of the recent changes. -Fix various casts. -Replace % 2 with & 1, even though the compiler was likely doing this already. -Add new file Variations06.h to accommodate new variations. -General cleanup.
2015-11-22 17:15:07 -05:00
<< "\t\tconst real_t phi = acos(clamp(vIn.z / zr, -(real_t)(1.0), (real_t)(1.0)));\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t\tconst real_t ps = sin(phi);\n"
<< "\t\tconst real_t pc = cos(phi);\n"
<< "\n"
--Bug fixes -Fix variations: blob2 (broken on CPU), epispiral, hole. -Fix reading variations from Apophysis: supershape. -Bump render was broken. --Code changes -Flip precalc sina/cosa and apply everywhere by flipping the usage. Flam3 had these reversed and it made the code confusing to read.
2019-12-28 00:04:41 -05:00
<< "\t\tvOut.x = " << weight << " * precalcCosa * ps * (zr + " << radius << ");\n"
<< "\t\tvOut.y = " << weight << " * precalcSina * ps * (zr + " << radius << ");\n"
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
<< "\t\tvOut.z = " << weight << " * pc * (zr + " << radius << ");\n"
--User changes -Add post_smartcrop. --Bug fixes -Fix bug in crackle. -Wrong point assignment in hexaplay3D, hexnix3D. -Improper Z assignment in rblur. -Fix inconsistency with original in circlecrop. -Put EMBER_ROOT bakc to ./../../../ in default.pri. This is TBD. --Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums. -Optimize crob, nBlur. -Fix weird assignment statement in falloff3. -Cleanup in VarFuncs::SimplexNoise3D(). -Replace fabs() with std::abs(). -General cleanup.
2016-01-12 23:42:12 -05:00
<< "\t}\n";
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
return ss.str();
}
--User changes -Add new variations: bubbleT3D, crob, hexaplay3D, hexcrop, hexes, hexnix3D, loonie2, loonie3, nBlur, octapol and synth. -Allow for pre/post versions of dc_bubble, dc_cylinder and dc_linear whereas before they were omitted. -When saving a file with multiple embers in it, detect if time values are all the same and if so, start them at zero and increment by 1 for each ember. -Allow for numerous quality increases to be coalesced into one. It will pick up at the end of the current render. -Show selection highlight on variations tree in response to mouse hover. This makes it easier to see for which variation or param the current mouse wheel action will apply. -Make default temporal samples be 100, whereas before it was 1000 which was overkill. -Require the shift key to be held with delete for deleting an ember to prevent it from triggering when the user enters delete in the edit box. -This wasn't otherwise fixable without writing a lot more code. --Bug fixes -EmberGenome was crashing when generating a sequence from a source file with more than 2 embers in it. -EmberGenome was improperly handling the first frame of a merge after the last frame of the loop. -These bugs were due to a previous commit. Revert parts of that commit. -Prevent a zoom value of less than 0 when reading from xml. -Slight optimization of the crescents, and mask variations, if the compiler wasn't doing it already. -Unique file naming was broken because it was looking for _# and the default names ended with -#. -Disallow renaming of an ember in the library tree to an empty string. -Severe bug that prevented some variations from being read correctly from params generated outside this program. -Severe OpenCL randomization bug. The first x coordinates of the first points in the first kernel call of the first ember of a render since the OpenCL renderer object was created were not random and were mostly -1. -Severe bug when populating xform selection distributions that could sometimes cause a crash due to roundoff error. Fix by using double. -Limit the max number of variations in a random ember to MAX_CL_VARS, which is 8. This ensures they'll look the same on CPU and GPU. -Prevent user from saving stylesheet to default.qss, it's a special reserved filename. --Code changes -Generalize using the running sum output point inside of a variation for all cases: pre, reg and post. -Allow for array variables in variations where the address of each element is stored in m_Params. -Qualify all math functions with std:: -No longer use our own Clamp() in OpenCL, instead use the standard clamp(). -Redesign how functions are used in the variations OpenCL code. -Add tests to EmberTester to verify some of the new functionality. -Place more const and override qualifiers on functions where appropriate. -Add a global rand with a lock to be used very sparingly. -Use a map instead of a vector for bad param names in Xml parsing. -Prefix affine interpolation mode defines with "AFFINE_" to make their purpose more clear. -Allow for variations that change state during iteration by sending a separate copy of the ember to each rendering thread. -Implement this same functionality with a local struct in OpenCL. It's members are the total of all variables that need to change state within an ember. -Add Contains() function to Utils.h. -EmberRender: print names of kernels being printed with --dump_kernel option. -Clean up EmberTester to handle some of the recent changes. -Fix various casts. -Replace % 2 with & 1, even though the compiler was likely doing this already. -Add new file Variations06.h to accommodate new variations. -General cleanup.
2015-11-22 17:15:07 -05:00
virtual vector<string> OpenCLGlobalFuncNames() const override
{
return vector<string> { "Hypot" };
}
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
protected:
void Init()
{
string prefix = Prefix();
m_Params.clear();
m_Params.push_back(ParamWithName<T>(&m_Radius, prefix + "spherivoid_radius"));
}
private:
T m_Radius;
};
/// <summary>
/// farblur.
/// </summary>
template <typename T>
--Code changes -Split VariationList into .h/cpp files. --Make VariationList.cpp the only place where the variations files are included from. This speeds up the build, but variations can no longer be directly instantiated. --Variations are no longer exported.
2016-02-18 21:58:24 -05:00
class FarblurVariation : public ParametricVariation<T>
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
public:
--Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums.
2016-01-04 19:50:15 -05:00
FarblurVariation(T weight = 1.0) : ParametricVariation<T>("farblur", eVariationId::VAR_FARBLUR, weight)
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
Init();
}
PARVARCOPY(FarblurVariation)
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
T r = m_Weight * (Sqr(helper.In.x - m_XOrigin) +
--User changes -Add post_smartcrop. --Bug fixes -Fix bug in crackle. -Wrong point assignment in hexaplay3D, hexnix3D. -Improper Z assignment in rblur. -Fix inconsistency with original in circlecrop. -Put EMBER_ROOT bakc to ./../../../ in default.pri. This is TBD. --Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums. -Optimize crob, nBlur. -Fix weird assignment statement in falloff3. -Cleanup in VarFuncs::SimplexNoise3D(). -Replace fabs() with std::abs(). -General cleanup.
2016-01-12 23:42:12 -05:00
Sqr(helper.In.y - m_YOrigin) +
Sqr(helper.In.z - m_ZOrigin)) *
(rand.Frand01<T>() + rand.Frand01<T>() + rand.Frand01<T>() + rand.Frand01<T>() - 2);
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
T u = rand.Frand01<T>() * M_2PI;
--User changes -Add new variations: bubbleT3D, crob, hexaplay3D, hexcrop, hexes, hexnix3D, loonie2, loonie3, nBlur, octapol and synth. -Allow for pre/post versions of dc_bubble, dc_cylinder and dc_linear whereas before they were omitted. -When saving a file with multiple embers in it, detect if time values are all the same and if so, start them at zero and increment by 1 for each ember. -Allow for numerous quality increases to be coalesced into one. It will pick up at the end of the current render. -Show selection highlight on variations tree in response to mouse hover. This makes it easier to see for which variation or param the current mouse wheel action will apply. -Make default temporal samples be 100, whereas before it was 1000 which was overkill. -Require the shift key to be held with delete for deleting an ember to prevent it from triggering when the user enters delete in the edit box. -This wasn't otherwise fixable without writing a lot more code. --Bug fixes -EmberGenome was crashing when generating a sequence from a source file with more than 2 embers in it. -EmberGenome was improperly handling the first frame of a merge after the last frame of the loop. -These bugs were due to a previous commit. Revert parts of that commit. -Prevent a zoom value of less than 0 when reading from xml. -Slight optimization of the crescents, and mask variations, if the compiler wasn't doing it already. -Unique file naming was broken because it was looking for _# and the default names ended with -#. -Disallow renaming of an ember in the library tree to an empty string. -Severe bug that prevented some variations from being read correctly from params generated outside this program. -Severe OpenCL randomization bug. The first x coordinates of the first points in the first kernel call of the first ember of a render since the OpenCL renderer object was created were not random and were mostly -1. -Severe bug when populating xform selection distributions that could sometimes cause a crash due to roundoff error. Fix by using double. -Limit the max number of variations in a random ember to MAX_CL_VARS, which is 8. This ensures they'll look the same on CPU and GPU. -Prevent user from saving stylesheet to default.qss, it's a special reserved filename. --Code changes -Generalize using the running sum output point inside of a variation for all cases: pre, reg and post. -Allow for array variables in variations where the address of each element is stored in m_Params. -Qualify all math functions with std:: -No longer use our own Clamp() in OpenCL, instead use the standard clamp(). -Redesign how functions are used in the variations OpenCL code. -Add tests to EmberTester to verify some of the new functionality. -Place more const and override qualifiers on functions where appropriate. -Add a global rand with a lock to be used very sparingly. -Use a map instead of a vector for bad param names in Xml parsing. -Prefix affine interpolation mode defines with "AFFINE_" to make their purpose more clear. -Allow for variations that change state during iteration by sending a separate copy of the ember to each rendering thread. -Implement this same functionality with a local struct in OpenCL. It's members are the total of all variables that need to change state within an ember. -Add Contains() function to Utils.h. -EmberRender: print names of kernels being printed with --dump_kernel option. -Clean up EmberTester to handle some of the recent changes. -Fix various casts. -Replace % 2 with & 1, even though the compiler was likely doing this already. -Add new file Variations06.h to accommodate new variations. -General cleanup.
2015-11-22 17:15:07 -05:00
T su = std::sin(u);
T cu = std::cos(u);
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
T v = rand.Frand01<T>() * M_2PI;
--User changes -Add new variations: bubbleT3D, crob, hexaplay3D, hexcrop, hexes, hexnix3D, loonie2, loonie3, nBlur, octapol and synth. -Allow for pre/post versions of dc_bubble, dc_cylinder and dc_linear whereas before they were omitted. -When saving a file with multiple embers in it, detect if time values are all the same and if so, start them at zero and increment by 1 for each ember. -Allow for numerous quality increases to be coalesced into one. It will pick up at the end of the current render. -Show selection highlight on variations tree in response to mouse hover. This makes it easier to see for which variation or param the current mouse wheel action will apply. -Make default temporal samples be 100, whereas before it was 1000 which was overkill. -Require the shift key to be held with delete for deleting an ember to prevent it from triggering when the user enters delete in the edit box. -This wasn't otherwise fixable without writing a lot more code. --Bug fixes -EmberGenome was crashing when generating a sequence from a source file with more than 2 embers in it. -EmberGenome was improperly handling the first frame of a merge after the last frame of the loop. -These bugs were due to a previous commit. Revert parts of that commit. -Prevent a zoom value of less than 0 when reading from xml. -Slight optimization of the crescents, and mask variations, if the compiler wasn't doing it already. -Unique file naming was broken because it was looking for _# and the default names ended with -#. -Disallow renaming of an ember in the library tree to an empty string. -Severe bug that prevented some variations from being read correctly from params generated outside this program. -Severe OpenCL randomization bug. The first x coordinates of the first points in the first kernel call of the first ember of a render since the OpenCL renderer object was created were not random and were mostly -1. -Severe bug when populating xform selection distributions that could sometimes cause a crash due to roundoff error. Fix by using double. -Limit the max number of variations in a random ember to MAX_CL_VARS, which is 8. This ensures they'll look the same on CPU and GPU. -Prevent user from saving stylesheet to default.qss, it's a special reserved filename. --Code changes -Generalize using the running sum output point inside of a variation for all cases: pre, reg and post. -Allow for array variables in variations where the address of each element is stored in m_Params. -Qualify all math functions with std:: -No longer use our own Clamp() in OpenCL, instead use the standard clamp(). -Redesign how functions are used in the variations OpenCL code. -Add tests to EmberTester to verify some of the new functionality. -Place more const and override qualifiers on functions where appropriate. -Add a global rand with a lock to be used very sparingly. -Use a map instead of a vector for bad param names in Xml parsing. -Prefix affine interpolation mode defines with "AFFINE_" to make their purpose more clear. -Allow for variations that change state during iteration by sending a separate copy of the ember to each rendering thread. -Implement this same functionality with a local struct in OpenCL. It's members are the total of all variables that need to change state within an ember. -Add Contains() function to Utils.h. -EmberRender: print names of kernels being printed with --dump_kernel option. -Clean up EmberTester to handle some of the recent changes. -Fix various casts. -Replace % 2 with & 1, even though the compiler was likely doing this already. -Add new file Variations06.h to accommodate new variations. -General cleanup.
2015-11-22 17:15:07 -05:00
T sv = std::sin(v);
T cv = std::cos(v);
Numerous fixes 0.4.0.5 Beta 07/18/2014 --User Changes Allow for vibrancy values > 1. Add flatten and unflatten menu items. Automatically flatten like Apophysis does. Add plugin and new_linear tags to Xml to be compatible with Apophysis. --Bug Fixes Fix blur, blur3d, bubble, cropn, cross, curl, curl3d, epispiral, ho, julia3d, julia3dz, loonie, mirror_x, mirror_y, mirror_z, rotate_x, sinusoidal, spherical, spherical3d, stripes. Unique filename on final render was completely broken. Two severe OpenCL bugs. Random seeds were biased and fusing was being reset too often leading to results that differ from the CPU. Subtle, but sometimes severe bug in the setup of the xaos weights. Use properly defined epsilon by getting the value from std::numeric_limits, rather than hard coding 1e-6 or 1e-10. Omit incorrect usage of epsilon everywhere. It should not be automatically added to denominators. Rather, it should only be used if the denominator is zero. Force final render progress bars to 100 on completion. Sometimes they didn't seem to make it there. Make variation name and params comparisons be case insensitive. --Code Changes Make ForEach and FindIf wrappers around std::for_each and std::find_if.
2014-07-19 02:33:18 -04:00
helper.Out.x = m_X * r * sv * cu;
helper.Out.y = m_Y * r * sv * su;
helper.Out.z = m_Z * r * cv;
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
}
--Code changes -Better use of const functions. -Ember.pro should not define BUILDING_EMBERCL, instead define BUILDING_EMBER.
2015-06-29 23:13:53 -04:00
virtual string OpenCLString() const override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
ostringstream ss, ss2;
0.4.1.3 Beta 10/14/2014 --User Changes Size is no longer fixed to the window size. Size scaling is done differently in the final render dialog. This fixes several bugs. Remove Xml saving size from settings and options dialog, it no longer applies. Final render can be broken into strips. Set default save path to the desktop if none is found in the settings file. Set default output size to 1920x1080 if none is found in the settings file. --Bug Fixes Better memory size reporting in final render dialog. --Code Changes Migrate to C++11, Qt 5.3.1, and Visual Studio 2013. Change most instances of unsigned int to size_t, and int to intmax_t. Add m_OrigPixPerUnit and m_ScaleType to Ember for scaling purposes. Replace some sprintf_s() calls in XmlToEmber with ostringstream. Move more non-templated members into RendererBase. Add CopyVec() overload that takes a per element function pointer. Add vector Memset(). Replace '&' with '+' instead of "&amp;" in XmlToEmber for much faster parsing. Break strips rendering out into EmberCommon and call from EmberRender and Fractorium. Make AddAndWriteBuffer() just call WriteBuffer(). Make AddAndWriteImage() delete the existing image first before replacing it. Add SetOutputTexture() to RendererCL to support making new textures in response to resize events. Remove multiple return statements in RendererCL, and replace with a bool that tracks results. Add ToDouble(), MakeEnd(), ToString() and Exists() wrappers in Fractorium. Add Size() wrapper in EmberFile. Make QString function arguments const QString&, and string with const string&. Make ShowCritical() wrapper for invoking a message box from another thread. Add combo box to TwoButtonWidget and rename.
2014-10-14 11:53:15 -04:00
intmax_t i = 0, varIndex = IndexInXform();
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
ss2 << "_" << XformIndexInEmber() << "]";
string index = ss2.str();
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
string weight = WeightDefineString();
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
string x = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string y = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string z = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
string xOrigin = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string yOrigin = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string zOrigin = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
ss << "\t{\n"
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
<< "\t\treal_t xmx = vIn.x - " << xOrigin << ";\n"
<< "\t\treal_t ymy = vIn.y - " << yOrigin << ";\n"
<< "\t\treal_t zmz = vIn.z - " << zOrigin << ";\n"
<< "\t\treal_t r = " << weight << " * (fma(xmx, xmx, fma(ymy, ymy, SQR(zmz))))\n"
<< "\t\t * (MwcNext01(mwc) + MwcNext01(mwc) + MwcNext01(mwc) + MwcNext01(mwc) - 2);\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t\treal_t u = MwcNext01(mwc) * M_2PI;\n"
<< "\t\treal_t su = sin(u);\n"
<< "\t\treal_t cu = cos(u);\n"
<< "\t\treal_t v = MwcNext01(mwc) * M_2PI;\n"
<< "\t\treal_t sv = sin(v);\n"
<< "\t\treal_t cv = cos(v);\n"
<< "\n"
<< "\t\tvOut.x = " << x << " * r * sv * cu;\n"
<< "\t\tvOut.y = " << y << " * r * sv * su;\n"
<< "\t\tvOut.z = " << z << " * r * cv;\n"
<< "\t}\n";
return ss.str();
}
protected:
void Init()
{
string prefix = Prefix();
m_Params.clear();
m_Params.push_back(ParamWithName<T>(&m_X, prefix + "farblur_x", 1));
m_Params.push_back(ParamWithName<T>(&m_Y, prefix + "farblur_y", 1));
m_Params.push_back(ParamWithName<T>(&m_Z, prefix + "farblur_z", 1));
m_Params.push_back(ParamWithName<T>(&m_XOrigin, prefix + "farblur_x_origin"));
m_Params.push_back(ParamWithName<T>(&m_YOrigin, prefix + "farblur_y_origin"));
m_Params.push_back(ParamWithName<T>(&m_ZOrigin, prefix + "farblur_z_origin"));
}
private:
T m_X;
T m_Y;
T m_Z;
T m_XOrigin;
T m_YOrigin;
T m_ZOrigin;
};
/// <summary>
/// curl_sp.
/// </summary>
template <typename T>
--Code changes -Split VariationList into .h/cpp files. --Make VariationList.cpp the only place where the variations files are included from. This speeds up the build, but variations can no longer be directly instantiated. --Variations are no longer exported.
2016-02-18 21:58:24 -05:00
class CurlSPVariation : public ParametricVariation<T>
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
public:
--Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums.
2016-01-04 19:50:15 -05:00
CurlSPVariation(T weight = 1.0) : ParametricVariation<T>("curl_sp", eVariationId::VAR_CURL_SP, weight)
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
Init();
}
PARVARCOPY(CurlSPVariation)
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
--Code changes -Make usage of VarFuncs more consistent by moving some global functions there. -Make CircleTrans1 a little safer by calling Zeps() on m_Sc during precalc. -Begin prepping for 0.9.9.5 release.
2016-03-04 21:54:06 -05:00
const T x = VarFuncs<T>::Powq4c(helper.In.x, m_Power);
const T y = VarFuncs<T>::Powq4c(helper.In.y, m_Power);
const T z = VarFuncs<T>::Powq4c(helper.In.z, m_Power);
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
const T d = SQR(x) - SQR(y);
--Code changes -Make usage of VarFuncs more consistent by moving some global functions there. -Make CircleTrans1 a little safer by calling Zeps() on m_Sc during precalc. -Begin prepping for 0.9.9.5 release.
2016-03-04 21:54:06 -05:00
const T re = VarFuncs<T>::Spread(m_C1 * x + m_C2 * d, m_Sx) + 1;
const T im = VarFuncs<T>::Spread(m_C1 * y + m_C2x2 * x * y, m_Sy);
T c = Zeps(VarFuncs<T>::Powq4c(SQR(re) + SQR(im), m_PowerInv));
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
const T r = m_Weight / c;
helper.Out.x = (x * re + y * im) * r;
helper.Out.y = (y * re - x * im) * r;
helper.Out.z = (z * m_Weight) / c;
outPoint.m_ColorX = Clamp<T>(outPoint.m_ColorX + m_DcAdjust * c, 0, 1);
}
--Code changes -Better use of const functions. -Ember.pro should not define BUILDING_EMBERCL, instead define BUILDING_EMBER.
2015-06-29 23:13:53 -04:00
virtual string OpenCLString() const override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
ostringstream ss, ss2;
0.4.1.3 Beta 10/14/2014 --User Changes Size is no longer fixed to the window size. Size scaling is done differently in the final render dialog. This fixes several bugs. Remove Xml saving size from settings and options dialog, it no longer applies. Final render can be broken into strips. Set default save path to the desktop if none is found in the settings file. Set default output size to 1920x1080 if none is found in the settings file. --Bug Fixes Better memory size reporting in final render dialog. --Code Changes Migrate to C++11, Qt 5.3.1, and Visual Studio 2013. Change most instances of unsigned int to size_t, and int to intmax_t. Add m_OrigPixPerUnit and m_ScaleType to Ember for scaling purposes. Replace some sprintf_s() calls in XmlToEmber with ostringstream. Move more non-templated members into RendererBase. Add CopyVec() overload that takes a per element function pointer. Add vector Memset(). Replace '&' with '+' instead of "&amp;" in XmlToEmber for much faster parsing. Break strips rendering out into EmberCommon and call from EmberRender and Fractorium. Make AddAndWriteBuffer() just call WriteBuffer(). Make AddAndWriteImage() delete the existing image first before replacing it. Add SetOutputTexture() to RendererCL to support making new textures in response to resize events. Remove multiple return statements in RendererCL, and replace with a bool that tracks results. Add ToDouble(), MakeEnd(), ToString() and Exists() wrappers in Fractorium. Add Size() wrapper in EmberFile. Make QString function arguments const QString&, and string with const string&. Make ShowCritical() wrapper for invoking a message box from another thread. Add combo box to TwoButtonWidget and rename.
2014-10-14 11:53:15 -04:00
intmax_t i = 0, varIndex = IndexInXform();
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
ss2 << "_" << XformIndexInEmber() << "]";
string index = ss2.str();
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
string weight = WeightDefineString();
string power = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string c1 = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string c2 = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string sx = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string sy = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string dc = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string c2x2 = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
string dcAdjust = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string powerInv = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
ss << "\t{\n"
<< "\t\tconst real_t x = Powq4c(vIn.x, " << power << ");\n"
<< "\t\tconst real_t y = Powq4c(vIn.y, " << power << ");\n"
<< "\t\tconst real_t z = Powq4c(vIn.z, " << power << ");\n"
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
<< "\t\tconst real_t d = fma(x, x, -SQR(y));\n"
<< "\t\tconst real_t re = Spread(fma(" << c1 << ", x, " << c2 << " * d), " << sx << ") + (real_t)(1.0);\n"
<< "\t\tconst real_t im = Spread(fma(" << c1 << ", y, " << c2x2 << " * x * y), " << sy << ");\n"
<< "\t\treal_t c = Zeps(Powq4c(fma(re, re, SQR(im)), " << powerInv << "));\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\n"
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
<< "\t\tconst real_t r = " << weight << " / c;\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\n"
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
<< "\t\tvOut.x = fma(x, re, y * im) * r;\n"
<< "\t\tvOut.y = fma(y, re, -(x * im)) * r;\n"
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
<< "\t\tvOut.z = (z * " << weight << ") / c;\n"
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
<< "\t\toutPoint->m_ColorX = clamp(fma(" << dcAdjust << ", c, outPoint->m_ColorX), (real_t)(0.0), (real_t)(1.0));\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t}\n";
return ss.str();
}
--User changes -Add new variations: bubbleT3D, crob, hexaplay3D, hexcrop, hexes, hexnix3D, loonie2, loonie3, nBlur, octapol and synth. -Allow for pre/post versions of dc_bubble, dc_cylinder and dc_linear whereas before they were omitted. -When saving a file with multiple embers in it, detect if time values are all the same and if so, start them at zero and increment by 1 for each ember. -Allow for numerous quality increases to be coalesced into one. It will pick up at the end of the current render. -Show selection highlight on variations tree in response to mouse hover. This makes it easier to see for which variation or param the current mouse wheel action will apply. -Make default temporal samples be 100, whereas before it was 1000 which was overkill. -Require the shift key to be held with delete for deleting an ember to prevent it from triggering when the user enters delete in the edit box. -This wasn't otherwise fixable without writing a lot more code. --Bug fixes -EmberGenome was crashing when generating a sequence from a source file with more than 2 embers in it. -EmberGenome was improperly handling the first frame of a merge after the last frame of the loop. -These bugs were due to a previous commit. Revert parts of that commit. -Prevent a zoom value of less than 0 when reading from xml. -Slight optimization of the crescents, and mask variations, if the compiler wasn't doing it already. -Unique file naming was broken because it was looking for _# and the default names ended with -#. -Disallow renaming of an ember in the library tree to an empty string. -Severe bug that prevented some variations from being read correctly from params generated outside this program. -Severe OpenCL randomization bug. The first x coordinates of the first points in the first kernel call of the first ember of a render since the OpenCL renderer object was created were not random and were mostly -1. -Severe bug when populating xform selection distributions that could sometimes cause a crash due to roundoff error. Fix by using double. -Limit the max number of variations in a random ember to MAX_CL_VARS, which is 8. This ensures they'll look the same on CPU and GPU. -Prevent user from saving stylesheet to default.qss, it's a special reserved filename. --Code changes -Generalize using the running sum output point inside of a variation for all cases: pre, reg and post. -Allow for array variables in variations where the address of each element is stored in m_Params. -Qualify all math functions with std:: -No longer use our own Clamp() in OpenCL, instead use the standard clamp(). -Redesign how functions are used in the variations OpenCL code. -Add tests to EmberTester to verify some of the new functionality. -Place more const and override qualifiers on functions where appropriate. -Add a global rand with a lock to be used very sparingly. -Use a map instead of a vector for bad param names in Xml parsing. -Prefix affine interpolation mode defines with "AFFINE_" to make their purpose more clear. -Allow for variations that change state during iteration by sending a separate copy of the ember to each rendering thread. -Implement this same functionality with a local struct in OpenCL. It's members are the total of all variables that need to change state within an ember. -Add Contains() function to Utils.h. -EmberRender: print names of kernels being printed with --dump_kernel option. -Clean up EmberTester to handle some of the recent changes. -Fix various casts. -Replace % 2 with & 1, even though the compiler was likely doing this already. -Add new file Variations06.h to accommodate new variations. -General cleanup.
2015-11-22 17:15:07 -05:00
virtual vector<string> OpenCLGlobalFuncNames() const override
{
return vector<string> { "Hypot", "Spread", "SignNz", "Powq4", "Powq4c", "Zeps" };
}
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
virtual void Precalc() override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
m_C2x2 = 2 * m_C2;
m_DcAdjust = T(0.1) * m_Dc;
m_Power = Zeps(m_Power);
m_PowerInv = 1 / m_Power;
}
protected:
void Init()
{
string prefix = Prefix();
m_Params.clear();
--Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums.
2016-01-04 19:50:15 -05:00
m_Params.push_back(ParamWithName<T>(&m_Power, prefix + "curl_sp_pow", 1, eParamType::REAL_NONZERO));
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
m_Params.push_back(ParamWithName<T>(&m_C1, prefix + "curl_sp_c1"));
m_Params.push_back(ParamWithName<T>(&m_C2, prefix + "curl_sp_c2"));
m_Params.push_back(ParamWithName<T>(&m_Sx, prefix + "curl_sp_sx"));
m_Params.push_back(ParamWithName<T>(&m_Sy, prefix + "curl_sp_sy"));
m_Params.push_back(ParamWithName<T>(&m_Dc, prefix + "curl_sp_dc"));
m_Params.push_back(ParamWithName<T>(true, &m_C2x2, prefix + "curl_sp_c2_x2"));
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
m_Params.push_back(ParamWithName<T>(true, &m_DcAdjust, prefix + "curl_sp_dc_adjust"));
m_Params.push_back(ParamWithName<T>(true, &m_PowerInv, prefix + "curl_sp_power_inv"));
}
private:
T m_Power;
T m_C1;
T m_C2;
T m_Sx;
T m_Sy;
T m_Dc;
T m_C2x2;//Precalc.
T m_DcAdjust;
T m_PowerInv;
};
/// <summary>
/// heat.
/// </summary>
template <typename T>
--Code changes -Split VariationList into .h/cpp files. --Make VariationList.cpp the only place where the variations files are included from. This speeds up the build, but variations can no longer be directly instantiated. --Variations are no longer exported.
2016-02-18 21:58:24 -05:00
class HeatVariation : public ParametricVariation<T>
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
public:
--Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums.
2016-01-04 19:50:15 -05:00
HeatVariation(T weight = 1.0) : ParametricVariation<T>("heat", eVariationId::VAR_HEAT, weight, true, false, false, false, true)
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
Init();
}
PARVARCOPY(HeatVariation)
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
--User changes -Add post_smartcrop. --Bug fixes -Fix bug in crackle. -Wrong point assignment in hexaplay3D, hexnix3D. -Improper Z assignment in rblur. -Fix inconsistency with original in circlecrop. -Put EMBER_ROOT bakc to ./../../../ in default.pri. This is TBD. --Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums. -Optimize crob, nBlur. -Fix weird assignment statement in falloff3. -Cleanup in VarFuncs::SimplexNoise3D(). -Replace fabs() with std::abs(). -General cleanup.
2016-01-12 23:42:12 -05:00
T r = std::sqrt(std::abs(helper.m_PrecalcSumSquares + helper.In.z));
--User changes -Add new variations: bubbleT3D, crob, hexaplay3D, hexcrop, hexes, hexnix3D, loonie2, loonie3, nBlur, octapol and synth. -Allow for pre/post versions of dc_bubble, dc_cylinder and dc_linear whereas before they were omitted. -When saving a file with multiple embers in it, detect if time values are all the same and if so, start them at zero and increment by 1 for each ember. -Allow for numerous quality increases to be coalesced into one. It will pick up at the end of the current render. -Show selection highlight on variations tree in response to mouse hover. This makes it easier to see for which variation or param the current mouse wheel action will apply. -Make default temporal samples be 100, whereas before it was 1000 which was overkill. -Require the shift key to be held with delete for deleting an ember to prevent it from triggering when the user enters delete in the edit box. -This wasn't otherwise fixable without writing a lot more code. --Bug fixes -EmberGenome was crashing when generating a sequence from a source file with more than 2 embers in it. -EmberGenome was improperly handling the first frame of a merge after the last frame of the loop. -These bugs were due to a previous commit. Revert parts of that commit. -Prevent a zoom value of less than 0 when reading from xml. -Slight optimization of the crescents, and mask variations, if the compiler wasn't doing it already. -Unique file naming was broken because it was looking for _# and the default names ended with -#. -Disallow renaming of an ember in the library tree to an empty string. -Severe bug that prevented some variations from being read correctly from params generated outside this program. -Severe OpenCL randomization bug. The first x coordinates of the first points in the first kernel call of the first ember of a render since the OpenCL renderer object was created were not random and were mostly -1. -Severe bug when populating xform selection distributions that could sometimes cause a crash due to roundoff error. Fix by using double. -Limit the max number of variations in a random ember to MAX_CL_VARS, which is 8. This ensures they'll look the same on CPU and GPU. -Prevent user from saving stylesheet to default.qss, it's a special reserved filename. --Code changes -Generalize using the running sum output point inside of a variation for all cases: pre, reg and post. -Allow for array variables in variations where the address of each element is stored in m_Params. -Qualify all math functions with std:: -No longer use our own Clamp() in OpenCL, instead use the standard clamp(). -Redesign how functions are used in the variations OpenCL code. -Add tests to EmberTester to verify some of the new functionality. -Place more const and override qualifiers on functions where appropriate. -Add a global rand with a lock to be used very sparingly. -Use a map instead of a vector for bad param names in Xml parsing. -Prefix affine interpolation mode defines with "AFFINE_" to make their purpose more clear. -Allow for variations that change state during iteration by sending a separate copy of the ember to each rendering thread. -Implement this same functionality with a local struct in OpenCL. It's members are the total of all variables that need to change state within an ember. -Add Contains() function to Utils.h. -EmberRender: print names of kernels being printed with --dump_kernel option. -Clean up EmberTester to handle some of the recent changes. -Fix various casts. -Replace % 2 with & 1, even though the compiler was likely doing this already. -Add new file Variations06.h to accommodate new variations. -General cleanup.
2015-11-22 17:15:07 -05:00
r += m_Ar * std::sin(fma(m_Br, r, m_Cr));
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
if (r == 0)
Numerous fixes 0.4.0.5 Beta 07/18/2014 --User Changes Allow for vibrancy values > 1. Add flatten and unflatten menu items. Automatically flatten like Apophysis does. Add plugin and new_linear tags to Xml to be compatible with Apophysis. --Bug Fixes Fix blur, blur3d, bubble, cropn, cross, curl, curl3d, epispiral, ho, julia3d, julia3dz, loonie, mirror_x, mirror_y, mirror_z, rotate_x, sinusoidal, spherical, spherical3d, stripes. Unique filename on final render was completely broken. Two severe OpenCL bugs. Random seeds were biased and fusing was being reset too often leading to results that differ from the CPU. Subtle, but sometimes severe bug in the setup of the xaos weights. Use properly defined epsilon by getting the value from std::numeric_limits, rather than hard coding 1e-6 or 1e-10. Omit incorrect usage of epsilon everywhere. It should not be automatically added to denominators. Rather, it should only be used if the denominator is zero. Force final render progress bars to 100 on completion. Sometimes they didn't seem to make it there. Make variation name and params comparisons be case insensitive. --Code Changes Make ForEach and FindIf wrappers around std::for_each and std::find_if.
2014-07-19 02:33:18 -04:00
r = EPS;
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
--User changes -Add new variations: bubbleT3D, crob, hexaplay3D, hexcrop, hexes, hexnix3D, loonie2, loonie3, nBlur, octapol and synth. -Allow for pre/post versions of dc_bubble, dc_cylinder and dc_linear whereas before they were omitted. -When saving a file with multiple embers in it, detect if time values are all the same and if so, start them at zero and increment by 1 for each ember. -Allow for numerous quality increases to be coalesced into one. It will pick up at the end of the current render. -Show selection highlight on variations tree in response to mouse hover. This makes it easier to see for which variation or param the current mouse wheel action will apply. -Make default temporal samples be 100, whereas before it was 1000 which was overkill. -Require the shift key to be held with delete for deleting an ember to prevent it from triggering when the user enters delete in the edit box. -This wasn't otherwise fixable without writing a lot more code. --Bug fixes -EmberGenome was crashing when generating a sequence from a source file with more than 2 embers in it. -EmberGenome was improperly handling the first frame of a merge after the last frame of the loop. -These bugs were due to a previous commit. Revert parts of that commit. -Prevent a zoom value of less than 0 when reading from xml. -Slight optimization of the crescents, and mask variations, if the compiler wasn't doing it already. -Unique file naming was broken because it was looking for _# and the default names ended with -#. -Disallow renaming of an ember in the library tree to an empty string. -Severe bug that prevented some variations from being read correctly from params generated outside this program. -Severe OpenCL randomization bug. The first x coordinates of the first points in the first kernel call of the first ember of a render since the OpenCL renderer object was created were not random and were mostly -1. -Severe bug when populating xform selection distributions that could sometimes cause a crash due to roundoff error. Fix by using double. -Limit the max number of variations in a random ember to MAX_CL_VARS, which is 8. This ensures they'll look the same on CPU and GPU. -Prevent user from saving stylesheet to default.qss, it's a special reserved filename. --Code changes -Generalize using the running sum output point inside of a variation for all cases: pre, reg and post. -Allow for array variables in variations where the address of each element is stored in m_Params. -Qualify all math functions with std:: -No longer use our own Clamp() in OpenCL, instead use the standard clamp(). -Redesign how functions are used in the variations OpenCL code. -Add tests to EmberTester to verify some of the new functionality. -Place more const and override qualifiers on functions where appropriate. -Add a global rand with a lock to be used very sparingly. -Use a map instead of a vector for bad param names in Xml parsing. -Prefix affine interpolation mode defines with "AFFINE_" to make their purpose more clear. -Allow for variations that change state during iteration by sending a separate copy of the ember to each rendering thread. -Implement this same functionality with a local struct in OpenCL. It's members are the total of all variables that need to change state within an ember. -Add Contains() function to Utils.h. -EmberRender: print names of kernels being printed with --dump_kernel option. -Clean up EmberTester to handle some of the recent changes. -Fix various casts. -Replace % 2 with & 1, even though the compiler was likely doing this already. -Add new file Variations06.h to accommodate new variations. -General cleanup.
2015-11-22 17:15:07 -05:00
T temp = fma(m_At, std::sin(fma(m_Bt, r, m_Ct)), helper.m_PrecalcAtanyx);
T st = std::sin(temp);
T ct = std::cos(temp);
temp = fma(m_Ap, std::sin(fma(m_Bp, r, m_Cp)), std::acos(Clamp<T>(helper.In.z / r, -1, 1)));
T sp = std::sin(temp);
T cp = std::cos(temp);
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
helper.Out.x = r * ct * sp;
helper.Out.y = r * st * sp;
helper.Out.z = r * cp;
}
--Code changes -Better use of const functions. -Ember.pro should not define BUILDING_EMBERCL, instead define BUILDING_EMBER.
2015-06-29 23:13:53 -04:00
virtual string OpenCLString() const override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
ostringstream ss, ss2;
More C++11
2014-09-10 01:41:26 -04:00
int i = 0;
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
ss2 << "_" << XformIndexInEmber() << "]";
string index = ss2.str();
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
string weight = WeightDefineString();
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
string thetaPeriod = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
string thetaPhase = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string thetaAmp = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string phiPeriod = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string phiPhase = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string phiAmp = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string rperiod = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string rphase = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string ramp = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string at = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string bt = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string ct = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string ap = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string bp = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string cp = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string ar = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string br = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string cr = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
ss << "\t{\n"
<< "\t\treal_t r = sqrt(fabs(precalcSumSquares + vIn.z));\n"
<< "\n"
<< "\t\tr += " << ar << " * sin(fma(" << br << ", r, " << cr << "));\n"
<< "\n"
<< "\t\tif (r == 0)\n"
Numerous fixes 0.4.0.5 Beta 07/18/2014 --User Changes Allow for vibrancy values > 1. Add flatten and unflatten menu items. Automatically flatten like Apophysis does. Add plugin and new_linear tags to Xml to be compatible with Apophysis. --Bug Fixes Fix blur, blur3d, bubble, cropn, cross, curl, curl3d, epispiral, ho, julia3d, julia3dz, loonie, mirror_x, mirror_y, mirror_z, rotate_x, sinusoidal, spherical, spherical3d, stripes. Unique filename on final render was completely broken. Two severe OpenCL bugs. Random seeds were biased and fusing was being reset too often leading to results that differ from the CPU. Subtle, but sometimes severe bug in the setup of the xaos weights. Use properly defined epsilon by getting the value from std::numeric_limits, rather than hard coding 1e-6 or 1e-10. Omit incorrect usage of epsilon everywhere. It should not be automatically added to denominators. Rather, it should only be used if the denominator is zero. Force final render progress bars to 100 on completion. Sometimes they didn't seem to make it there. Make variation name and params comparisons be case insensitive. --Code Changes Make ForEach and FindIf wrappers around std::for_each and std::find_if.
2014-07-19 02:33:18 -04:00
<< "\t\t r = EPS;\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\n"
<< "\t\treal_t temp = fma(" << at << ", sin(fma(" << bt << ", r, " << ct << ")), precalcAtanyx);\n"
<< "\t\treal_t st = sin(temp);\n"
<< "\t\treal_t ct = cos(temp);\n"
<< "\n"
--Bug fixes -Ensure all OpenCL functions that take multiple arguments cast literal values to real_t.
2019-05-23 22:44:27 -04:00
<< "\t\ttemp = fma(" << ap << ", sin(fma(" << bp << ", r, " << cp << ")), acos(clamp(vIn.z / r, (real_t)(-1.0), (real_t)(1.0))));\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\n"
<< "\t\treal_t sp = sin(temp);\n"
<< "\t\treal_t cp = cos(temp);\n"
<< "\n"
<< "\t\tvOut.x = r * ct * sp;\n"
<< "\t\tvOut.y = r * st * sp;\n"
<< "\t\tvOut.z = r * cp;\n"
<< "\t}\n";
return ss.str();
}
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
virtual void Precalc() override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
T tx = m_ThetaPeriod == 0 ? 0 : (1 / m_ThetaPeriod);
T px = m_PhiPeriod == 0 ? 0 : (1 / m_PhiPeriod);
T rx = m_Rperiod == 0 ? 0 : (1 / m_Rperiod);
m_At = m_Weight * m_ThetaAmp;
m_Bt = M_2PI * tx;
m_Ct = m_ThetaPhase * tx;
m_Ap = m_Weight * m_PhiAmp;
m_Bp = M_2PI * px;
m_Cp = m_PhiPhase * px;
m_Ar = m_Weight * m_Ramp;
m_Br = M_2PI * rx;
m_Cr = m_Rphase * rx;
}
protected:
void Init()
{
string prefix = Prefix();
m_Params.clear();
m_Params.push_back(ParamWithName<T>(&m_ThetaPeriod, prefix + "heat_theta_period", 1));
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
m_Params.push_back(ParamWithName<T>(&m_ThetaPhase, prefix + "heat_theta_phase"));
m_Params.push_back(ParamWithName<T>(&m_ThetaAmp, prefix + "heat_theta_amp", 1));
m_Params.push_back(ParamWithName<T>(&m_PhiPeriod, prefix + "heat_phi_period", 1));
m_Params.push_back(ParamWithName<T>(&m_PhiPhase, prefix + "heat_phi_phase"));
m_Params.push_back(ParamWithName<T>(&m_PhiAmp, prefix + "heat_phi_amp"));
m_Params.push_back(ParamWithName<T>(&m_Rperiod, prefix + "heat_r_period", 1));
m_Params.push_back(ParamWithName<T>(&m_Rphase, prefix + "heat_r_phase"));
m_Params.push_back(ParamWithName<T>(&m_Ramp, prefix + "heat_r_amp"));
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
m_Params.push_back(ParamWithName<T>(true, &m_At, prefix + "heat_at"));
m_Params.push_back(ParamWithName<T>(true, &m_Bt, prefix + "heat_bt"));
m_Params.push_back(ParamWithName<T>(true, &m_Ct, prefix + "heat_ct"));
m_Params.push_back(ParamWithName<T>(true, &m_Ap, prefix + "heat_ap"));
m_Params.push_back(ParamWithName<T>(true, &m_Bp, prefix + "heat_bp"));
m_Params.push_back(ParamWithName<T>(true, &m_Cp, prefix + "heat_cp"));
m_Params.push_back(ParamWithName<T>(true, &m_Ar, prefix + "heat_ar"));
m_Params.push_back(ParamWithName<T>(true, &m_Br, prefix + "heat_br"));
m_Params.push_back(ParamWithName<T>(true, &m_Cr, prefix + "heat_cr"));
}
private:
T m_ThetaPeriod;
T m_ThetaPhase;
T m_ThetaAmp;
T m_PhiPeriod;
T m_PhiPhase;
T m_PhiAmp;
T m_Rperiod;
T m_Rphase;
T m_Ramp;
T m_At;//Precalc.
T m_Bt;
T m_Ct;
T m_Ap;
T m_Bp;
T m_Cp;
T m_Ar;
T m_Br;
T m_Cr;
};
/// <summary>
/// interference2.
/// </summary>
template <typename T>
--Code changes -Split VariationList into .h/cpp files. --Make VariationList.cpp the only place where the variations files are included from. This speeds up the build, but variations can no longer be directly instantiated. --Variations are no longer exported.
2016-02-18 21:58:24 -05:00
class Interference2Variation : public ParametricVariation<T>
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
public:
--Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums.
2016-01-04 19:50:15 -05:00
Interference2Variation(T weight = 1.0) : ParametricVariation<T>("interference2", eVariationId::VAR_INTERFERENCE2, weight)
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
Init();
}
PARVARCOPY(Interference2Variation)
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
T fp1x;
T fp1y;
T fp2x;
T fp2y;
More Linux work. Convert all casts to new style, away from legacy.
2014-12-07 02:51:44 -05:00
switch (int(m_T1))
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
case 0:
fp1x = Sine(m_A1, m_B1, m_C1, m_P1, helper.In.x);
fp1y = Sine(m_A1, m_B1, m_C1, m_P1, helper.In.y);
break;
--User changes -Add post_smartcrop. --Bug fixes -Fix bug in crackle. -Wrong point assignment in hexaplay3D, hexnix3D. -Improper Z assignment in rblur. -Fix inconsistency with original in circlecrop. -Put EMBER_ROOT bakc to ./../../../ in default.pri. This is TBD. --Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums. -Optimize crob, nBlur. -Fix weird assignment statement in falloff3. -Cleanup in VarFuncs::SimplexNoise3D(). -Replace fabs() with std::abs(). -General cleanup.
2016-01-12 23:42:12 -05:00
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
case 1:
fp1x = Tri(m_A1, m_B1, m_C1, m_P1, helper.In.x);
fp1y = Tri(m_A1, m_B1, m_C1, m_P1, helper.In.y);
break;
--User changes -Add post_smartcrop. --Bug fixes -Fix bug in crackle. -Wrong point assignment in hexaplay3D, hexnix3D. -Improper Z assignment in rblur. -Fix inconsistency with original in circlecrop. -Put EMBER_ROOT bakc to ./../../../ in default.pri. This is TBD. --Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums. -Optimize crob, nBlur. -Fix weird assignment statement in falloff3. -Cleanup in VarFuncs::SimplexNoise3D(). -Replace fabs() with std::abs(). -General cleanup.
2016-01-12 23:42:12 -05:00
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
case 2:
fp1x = Squ(m_A1, m_B1, m_C1, m_P1, helper.In.x);
fp1y = Squ(m_A1, m_B1, m_C1, m_P1, helper.In.y);
break;
--User changes -Add post_smartcrop. --Bug fixes -Fix bug in crackle. -Wrong point assignment in hexaplay3D, hexnix3D. -Improper Z assignment in rblur. -Fix inconsistency with original in circlecrop. -Put EMBER_ROOT bakc to ./../../../ in default.pri. This is TBD. --Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums. -Optimize crob, nBlur. -Fix weird assignment statement in falloff3. -Cleanup in VarFuncs::SimplexNoise3D(). -Replace fabs() with std::abs(). -General cleanup.
2016-01-12 23:42:12 -05:00
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
default:
fp1x = Sine(m_A1, m_B1, m_C1, m_P1, helper.In.x);
fp1y = Sine(m_A1, m_B1, m_C1, m_P1, helper.In.y);
break;
}
More Linux work. Convert all casts to new style, away from legacy.
2014-12-07 02:51:44 -05:00
switch (int(m_T2))
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
case 0:
fp2x = Sine(m_A2, m_B2, m_C2, m_P2, helper.In.x);
fp2y = Sine(m_A2, m_B2, m_C2, m_P2, helper.In.y);
break;
--User changes -Add post_smartcrop. --Bug fixes -Fix bug in crackle. -Wrong point assignment in hexaplay3D, hexnix3D. -Improper Z assignment in rblur. -Fix inconsistency with original in circlecrop. -Put EMBER_ROOT bakc to ./../../../ in default.pri. This is TBD. --Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums. -Optimize crob, nBlur. -Fix weird assignment statement in falloff3. -Cleanup in VarFuncs::SimplexNoise3D(). -Replace fabs() with std::abs(). -General cleanup.
2016-01-12 23:42:12 -05:00
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
case 1:
fp2x = Tri(m_A2, m_B2, m_C2, m_P2, helper.In.x);
fp2y = Tri(m_A2, m_B2, m_C2, m_P2, helper.In.y);
break;
--User changes -Add post_smartcrop. --Bug fixes -Fix bug in crackle. -Wrong point assignment in hexaplay3D, hexnix3D. -Improper Z assignment in rblur. -Fix inconsistency with original in circlecrop. -Put EMBER_ROOT bakc to ./../../../ in default.pri. This is TBD. --Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums. -Optimize crob, nBlur. -Fix weird assignment statement in falloff3. -Cleanup in VarFuncs::SimplexNoise3D(). -Replace fabs() with std::abs(). -General cleanup.
2016-01-12 23:42:12 -05:00
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
case 2:
fp2x = Squ(m_A2, m_B2, m_C2, m_P2, helper.In.x);
fp2y = Squ(m_A2, m_B2, m_C2, m_P2, helper.In.y);
break;
--User changes -Add post_smartcrop. --Bug fixes -Fix bug in crackle. -Wrong point assignment in hexaplay3D, hexnix3D. -Improper Z assignment in rblur. -Fix inconsistency with original in circlecrop. -Put EMBER_ROOT bakc to ./../../../ in default.pri. This is TBD. --Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums. -Optimize crob, nBlur. -Fix weird assignment statement in falloff3. -Cleanup in VarFuncs::SimplexNoise3D(). -Replace fabs() with std::abs(). -General cleanup.
2016-01-12 23:42:12 -05:00
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
default:
fp2x = Sine(m_A2, m_B2, m_C2, m_P2, helper.In.x);
fp2y = Sine(m_A2, m_B2, m_C2, m_P2, helper.In.y);
break;
}
helper.Out.x = m_Weight * (fp1x + fp2x);
helper.Out.y = m_Weight * (fp1y + fp2y);
--User changes -Allow for pausing the renderer in the main window. This makes is more efficient when entering many parameters, such as when following a tutorial. -Add support for new variations: erf, gamma, jac_cn, jac_dn, jac_sn, logDB, pressure_wave, pRose3D, splits3D, w, waves2b, x, xerf, y, z. -Inform user of the start and stop of file parsing in EmberAnimate because the files could potentially be very large. -Move the follwing fields to a new table called Animation: Interpolation, Affine Interpolation, Temporal Samples, Temporal Filter Width, Temporal Filter Type. -These currently have no effect on the interactive renderer and instead are used when running flames through EmberGenome to generate sequences, and then animating them in Fractorium or EmberAnimate. -Add new parameter overrides for EmberRender and EmberAnimate which directly assign values to all flames being rendered, rather than scale: --quality --demin --demax --Bug fixes -Left pad instead of right pad names of sequence outputs from EmberGenome. -Unique file naming was broken for files which already had an underscore in them. -Properly report that png is the default format of EmberRender and EmberAnimate output instead of erroneously claiming it was jpg. -Make command line programs search these folders in this order for the palette file: ./ ~/.fractorium ~/.config/fractorium /usr/share/fractorium /usr/local/share/fractorium -Fix possible bad values in hexes. -Better assignment of Z variables. -Fix boarders due to use of poorly implemented rint() function from flam3. Use std::rint() now. -wedge_sph was completely wrong due to having accidentally swapped the mapping of two parameters. -Make juliascope juliascope_power parameter be of type REAL_NONZERO since it's used as a denominator. -Make Z assignment compatible with the originals in: -arch, bcircle, bCollide, bent, bent2, bisplit, blob, blur_linear, blur_square, bMod, boarders, boarders2, bSwirl, bTransform, butterfly, cardioid, cell, circleblur, circlize, circlize2, circus, collideoscope, cos, cosine, cosh, coth, cpow, cpow2, crescents, cropn, csc, csch, curl, curve, dc_gridout, deltaa, diamond, disc2, eclipse, eCollide, edisc, eJulia, elliptic, eMod, eMotion, ennepers, epispiral, ePush, eRotate, eScale, eSwirl, ex, exp, expo, exponential, fan, fdisc, fibonacci, fibonacci2, fisheye, flipcircle, flipy, flower, flux, funnel, glynnia, GlynnSim1, GlynnSim2, GlynnSim3, gridout, handkerchief, heart, hole, idisc, julia, julian2, juliaNab, kaleidoscope, lazyTravis, Lissajous, mask, MobiusN, mobius_strip, modulus, murl, murl2, npolar, ortho, oscilloscope, parabola, perspective, petal, phoenix_julia, pie (was also inconsistent between cpu and gpu), poincare, popcorn, popcorn2, power, pow_block, rational3, rays, rblur, rings, rippled, roundspher, sec, secant2, sigmoid, sin, sineblur, sinh, sinusgrid, sphericaln, spiralwing, spirograph, split, squarize, squirrel, squish, sschecks, starblur, stripes, stwin, super_shape, tan, tancos, tangent, tanh, TwinTrian, twoface, unpolar, waves, wavesn, wedge_julia, whorl, xheart, zblur, zscale. --Code changes -Generalize Variation::PrecalcHelper() and rename to PrePostPrecalcHelper(). --Do the same for the OpenCL version and rename it PrePostPrecalcOpenCLString(). -Rename Variation::m_AssignType to m_PrePostAssignType since it's only relevant to pre/post variations.
2016-01-29 20:02:15 -05:00
helper.Out.z = DefaultZ(helper);
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
}
--Code changes -Better use of const functions. -Ember.pro should not define BUILDING_EMBERCL, instead define BUILDING_EMBER.
2015-06-29 23:13:53 -04:00
virtual string OpenCLString() const override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
ostringstream ss, ss2;
0.4.1.3 Beta 10/14/2014 --User Changes Size is no longer fixed to the window size. Size scaling is done differently in the final render dialog. This fixes several bugs. Remove Xml saving size from settings and options dialog, it no longer applies. Final render can be broken into strips. Set default save path to the desktop if none is found in the settings file. Set default output size to 1920x1080 if none is found in the settings file. --Bug Fixes Better memory size reporting in final render dialog. --Code Changes Migrate to C++11, Qt 5.3.1, and Visual Studio 2013. Change most instances of unsigned int to size_t, and int to intmax_t. Add m_OrigPixPerUnit and m_ScaleType to Ember for scaling purposes. Replace some sprintf_s() calls in XmlToEmber with ostringstream. Move more non-templated members into RendererBase. Add CopyVec() overload that takes a per element function pointer. Add vector Memset(). Replace '&' with '+' instead of "&amp;" in XmlToEmber for much faster parsing. Break strips rendering out into EmberCommon and call from EmberRender and Fractorium. Make AddAndWriteBuffer() just call WriteBuffer(). Make AddAndWriteImage() delete the existing image first before replacing it. Add SetOutputTexture() to RendererCL to support making new textures in response to resize events. Remove multiple return statements in RendererCL, and replace with a bool that tracks results. Add ToDouble(), MakeEnd(), ToString() and Exists() wrappers in Fractorium. Add Size() wrapper in EmberFile. Make QString function arguments const QString&, and string with const string&. Make ShowCritical() wrapper for invoking a message box from another thread. Add combo box to TwoButtonWidget and rename.
2014-10-14 11:53:15 -04:00
intmax_t i = 0, varIndex = IndexInXform();
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
ss2 << "_" << XformIndexInEmber() << "]";
string index = ss2.str();
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
string weight = WeightDefineString();
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
string a1 = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string b1 = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string c1 = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string p1 = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string t1 = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string a2 = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string b2 = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string c2 = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string p2 = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string t2 = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
ss << "\t{\n"
<< "\t\treal_t fp1x;\n"
<< "\t\treal_t fp1y;\n"
<< "\t\treal_t fp2x;\n"
<< "\t\treal_t fp2y;\n"
<< "\n"
<< "\t\tswitch ((int)" << t1 << ")\n"
<< "\t\t{\n"
<< "\t\t case 0:\n"
<< "\t\t fp1x = Interference2Sine(" << a1 << ", " << b1 << ", " << c1 << ", " << p1 << ", vIn.x);\n"
<< "\t\t fp1y = Interference2Sine(" << a1 << ", " << b1 << ", " << c1 << ", " << p1 << ", vIn.y);\n"
<< "\t\t break;\n"
<< "\t\t case 1:\n"
<< "\t\t fp1x = Interference2Tri(" << a1 << ", " << b1 << ", " << c1 << ", " << p1 << ", vIn.x);\n"
<< "\t\t fp1y = Interference2Tri(" << a1 << ", " << b1 << ", " << c1 << ", " << p1 << ", vIn.y);\n"
<< "\t\t break;\n"
<< "\t\t case 2:\n"
<< "\t\t fp1x = Interference2Squ(" << a1 << ", " << b1 << ", " << c1 << ", " << p1 << ", vIn.x);\n"
<< "\t\t fp1y = Interference2Squ(" << a1 << ", " << b1 << ", " << c1 << ", " << p1 << ", vIn.y);\n"
<< "\t\t break;\n"
<< "\t\t default:\n"
<< "\t\t fp1x = Interference2Sine(" << a1 << ", " << b1 << ", " << c1 << ", " << p1 << ", vIn.x);\n"
<< "\t\t fp1y = Interference2Sine(" << a1 << ", " << b1 << ", " << c1 << ", " << p1 << ", vIn.y);\n"
<< "\t\t break;\n"
<< "\t\t}\n"
<< "\n"
<< "\t\tswitch ((int)" << t2 << ")\n"
<< "\t\t{\n"
<< "\t\t case 0:\n"
<< "\t\t fp2x = Interference2Sine(" << a2 << ", " << b2 << ", " << c2 << ", " << p2 << ", vIn.x);\n"
<< "\t\t fp2y = Interference2Sine(" << a2 << ", " << b2 << ", " << c2 << ", " << p2 << ", vIn.y);\n"
<< "\t\t break;\n"
<< "\t\t case 1:\n"
<< "\t\t fp2x = Interference2Tri(" << a2 << ", " << b2 << ", " << c2 << ", " << p2 << ", vIn.x);\n"
<< "\t\t fp2y = Interference2Tri(" << a2 << ", " << b2 << ", " << c2 << ", " << p2 << ", vIn.y);\n"
<< "\t\t break;\n"
<< "\t\t case 2:\n"
<< "\t\t fp2x = Interference2Squ(" << a2 << ", " << b2 << ", " << c2 << ", " << p2 << ", vIn.x);\n"
<< "\t\t fp2y = Interference2Squ(" << a2 << ", " << b2 << ", " << c2 << ", " << p2 << ", vIn.y);\n"
<< "\t\t break;\n"
<< "\t\t default:\n"
<< "\t\t fp2x = Interference2Sine(" << a2 << ", " << b2 << ", " << c2 << ", " << p2 << ", vIn.x);\n"
<< "\t\t fp2y = Interference2Sine(" << a2 << ", " << b2 << ", " << c2 << ", " << p2 << ", vIn.y);\n"
<< "\t\t break;\n"
<< "\t\t}\n"
<< "\n"
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
<< "\t\tvOut.x = " << weight << " * (fp1x + fp2x);\n"
<< "\t\tvOut.y = " << weight << " * (fp1y + fp2y);\n"
--User changes -Allow for pausing the renderer in the main window. This makes is more efficient when entering many parameters, such as when following a tutorial. -Add support for new variations: erf, gamma, jac_cn, jac_dn, jac_sn, logDB, pressure_wave, pRose3D, splits3D, w, waves2b, x, xerf, y, z. -Inform user of the start and stop of file parsing in EmberAnimate because the files could potentially be very large. -Move the follwing fields to a new table called Animation: Interpolation, Affine Interpolation, Temporal Samples, Temporal Filter Width, Temporal Filter Type. -These currently have no effect on the interactive renderer and instead are used when running flames through EmberGenome to generate sequences, and then animating them in Fractorium or EmberAnimate. -Add new parameter overrides for EmberRender and EmberAnimate which directly assign values to all flames being rendered, rather than scale: --quality --demin --demax --Bug fixes -Left pad instead of right pad names of sequence outputs from EmberGenome. -Unique file naming was broken for files which already had an underscore in them. -Properly report that png is the default format of EmberRender and EmberAnimate output instead of erroneously claiming it was jpg. -Make command line programs search these folders in this order for the palette file: ./ ~/.fractorium ~/.config/fractorium /usr/share/fractorium /usr/local/share/fractorium -Fix possible bad values in hexes. -Better assignment of Z variables. -Fix boarders due to use of poorly implemented rint() function from flam3. Use std::rint() now. -wedge_sph was completely wrong due to having accidentally swapped the mapping of two parameters. -Make juliascope juliascope_power parameter be of type REAL_NONZERO since it's used as a denominator. -Make Z assignment compatible with the originals in: -arch, bcircle, bCollide, bent, bent2, bisplit, blob, blur_linear, blur_square, bMod, boarders, boarders2, bSwirl, bTransform, butterfly, cardioid, cell, circleblur, circlize, circlize2, circus, collideoscope, cos, cosine, cosh, coth, cpow, cpow2, crescents, cropn, csc, csch, curl, curve, dc_gridout, deltaa, diamond, disc2, eclipse, eCollide, edisc, eJulia, elliptic, eMod, eMotion, ennepers, epispiral, ePush, eRotate, eScale, eSwirl, ex, exp, expo, exponential, fan, fdisc, fibonacci, fibonacci2, fisheye, flipcircle, flipy, flower, flux, funnel, glynnia, GlynnSim1, GlynnSim2, GlynnSim3, gridout, handkerchief, heart, hole, idisc, julia, julian2, juliaNab, kaleidoscope, lazyTravis, Lissajous, mask, MobiusN, mobius_strip, modulus, murl, murl2, npolar, ortho, oscilloscope, parabola, perspective, petal, phoenix_julia, pie (was also inconsistent between cpu and gpu), poincare, popcorn, popcorn2, power, pow_block, rational3, rays, rblur, rings, rippled, roundspher, sec, secant2, sigmoid, sin, sineblur, sinh, sinusgrid, sphericaln, spiralwing, spirograph, split, squarize, squirrel, squish, sschecks, starblur, stripes, stwin, super_shape, tan, tancos, tangent, tanh, TwinTrian, twoface, unpolar, waves, wavesn, wedge_julia, whorl, xheart, zblur, zscale. --Code changes -Generalize Variation::PrecalcHelper() and rename to PrePostPrecalcHelper(). --Do the same for the OpenCL version and rename it PrePostPrecalcOpenCLString(). -Rename Variation::m_AssignType to m_PrePostAssignType since it's only relevant to pre/post variations.
2016-01-29 20:02:15 -05:00
<< "\t\tvOut.z = " << DefaultZCl()
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t}\n";
return ss.str();
}
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
--User changes -Add new variations: bubbleT3D, crob, hexaplay3D, hexcrop, hexes, hexnix3D, loonie2, loonie3, nBlur, octapol and synth. -Allow for pre/post versions of dc_bubble, dc_cylinder and dc_linear whereas before they were omitted. -When saving a file with multiple embers in it, detect if time values are all the same and if so, start them at zero and increment by 1 for each ember. -Allow for numerous quality increases to be coalesced into one. It will pick up at the end of the current render. -Show selection highlight on variations tree in response to mouse hover. This makes it easier to see for which variation or param the current mouse wheel action will apply. -Make default temporal samples be 100, whereas before it was 1000 which was overkill. -Require the shift key to be held with delete for deleting an ember to prevent it from triggering when the user enters delete in the edit box. -This wasn't otherwise fixable without writing a lot more code. --Bug fixes -EmberGenome was crashing when generating a sequence from a source file with more than 2 embers in it. -EmberGenome was improperly handling the first frame of a merge after the last frame of the loop. -These bugs were due to a previous commit. Revert parts of that commit. -Prevent a zoom value of less than 0 when reading from xml. -Slight optimization of the crescents, and mask variations, if the compiler wasn't doing it already. -Unique file naming was broken because it was looking for _# and the default names ended with -#. -Disallow renaming of an ember in the library tree to an empty string. -Severe bug that prevented some variations from being read correctly from params generated outside this program. -Severe OpenCL randomization bug. The first x coordinates of the first points in the first kernel call of the first ember of a render since the OpenCL renderer object was created were not random and were mostly -1. -Severe bug when populating xform selection distributions that could sometimes cause a crash due to roundoff error. Fix by using double. -Limit the max number of variations in a random ember to MAX_CL_VARS, which is 8. This ensures they'll look the same on CPU and GPU. -Prevent user from saving stylesheet to default.qss, it's a special reserved filename. --Code changes -Generalize using the running sum output point inside of a variation for all cases: pre, reg and post. -Allow for array variables in variations where the address of each element is stored in m_Params. -Qualify all math functions with std:: -No longer use our own Clamp() in OpenCL, instead use the standard clamp(). -Redesign how functions are used in the variations OpenCL code. -Add tests to EmberTester to verify some of the new functionality. -Place more const and override qualifiers on functions where appropriate. -Add a global rand with a lock to be used very sparingly. -Use a map instead of a vector for bad param names in Xml parsing. -Prefix affine interpolation mode defines with "AFFINE_" to make their purpose more clear. -Allow for variations that change state during iteration by sending a separate copy of the ember to each rendering thread. -Implement this same functionality with a local struct in OpenCL. It's members are the total of all variables that need to change state within an ember. -Add Contains() function to Utils.h. -EmberRender: print names of kernels being printed with --dump_kernel option. -Clean up EmberTester to handle some of the recent changes. -Fix various casts. -Replace % 2 with & 1, even though the compiler was likely doing this already. -Add new file Variations06.h to accommodate new variations. -General cleanup.
2015-11-22 17:15:07 -05:00
virtual string OpenCLFuncsString() const override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
return
"real_t Interference2Sine(real_t a, real_t b, real_t c, real_t p, real_t x)\n"
"{\n"
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
" return a * pow(fabs(sin(fma(b, x, c))), p);\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
"}\n"
"\n"
"real_t Interference2Tri(real_t a, real_t b, real_t c, real_t p, real_t x)\n"
"{\n"
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
" return a * 2 * pow(fabs(asin(cos(fma(b, x, c - MPI2)))) * M1PI, p);\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
"}\n"
"\n"
"real_t Interference2Squ(real_t a, real_t b, real_t c, real_t p, real_t x)\n"
"{\n"
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
" return a * pow(sin(fma(b, x, c)) < 0 ? EPS : 1, p);\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
"}\n"
"\n";
}
protected:
void Init()
{
string prefix = Prefix();
m_Params.clear();
m_Params.push_back(ParamWithName<T>(&m_A1, prefix + "interference2_a1", 1));//Original used a prefix of intrfr2_, which is incompatible with Ember's design.
m_Params.push_back(ParamWithName<T>(&m_B1, prefix + "interference2_b1", 1));
m_Params.push_back(ParamWithName<T>(&m_C1, prefix + "interference2_c1"));
m_Params.push_back(ParamWithName<T>(&m_P1, prefix + "interference2_p1", 1));
--Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums.
2016-01-04 19:50:15 -05:00
m_Params.push_back(ParamWithName<T>(&m_T1, prefix + "interference2_t1", 0, eParamType::INTEGER, 0, 2));
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
m_Params.push_back(ParamWithName<T>(&m_A2, prefix + "interference2_a2", 1));
m_Params.push_back(ParamWithName<T>(&m_B2, prefix + "interference2_b2", 1));
m_Params.push_back(ParamWithName<T>(&m_C2, prefix + "interference2_c2"));
m_Params.push_back(ParamWithName<T>(&m_P2, prefix + "interference2_p2", 1));
--Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums.
2016-01-04 19:50:15 -05:00
m_Params.push_back(ParamWithName<T>(&m_T2, prefix + "interference2_t2", 0, eParamType::INTEGER, 0, 2));
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
}
private:
inline static T Sine(T a, T b, T c, T p, T x)
{
--User changes -Add post_smartcrop. --Bug fixes -Fix bug in crackle. -Wrong point assignment in hexaplay3D, hexnix3D. -Improper Z assignment in rblur. -Fix inconsistency with original in circlecrop. -Put EMBER_ROOT bakc to ./../../../ in default.pri. This is TBD. --Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums. -Optimize crob, nBlur. -Fix weird assignment statement in falloff3. -Cleanup in VarFuncs::SimplexNoise3D(). -Replace fabs() with std::abs(). -General cleanup.
2016-01-12 23:42:12 -05:00
return a * std::pow(std::abs(std::sin(b * x + c)), p);//Original did not fabs().
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
}
inline static T Tri(T a, T b, T c, T p, T x)
{
--User changes -Add post_smartcrop. --Bug fixes -Fix bug in crackle. -Wrong point assignment in hexaplay3D, hexnix3D. -Improper Z assignment in rblur. -Fix inconsistency with original in circlecrop. -Put EMBER_ROOT bakc to ./../../../ in default.pri. This is TBD. --Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums. -Optimize crob, nBlur. -Fix weird assignment statement in falloff3. -Cleanup in VarFuncs::SimplexNoise3D(). -Replace fabs() with std::abs(). -General cleanup.
2016-01-12 23:42:12 -05:00
return a * 2 * std::pow(std::abs(std::asin(std::cos(b * x + c - T(M_PI_2)))) * T(M_1_PI), p);//Original did not fabs().
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
}
inline static T Squ(T a, T b, T c, T p, T x)
{
--User changes -Add new variations: bubbleT3D, crob, hexaplay3D, hexcrop, hexes, hexnix3D, loonie2, loonie3, nBlur, octapol and synth. -Allow for pre/post versions of dc_bubble, dc_cylinder and dc_linear whereas before they were omitted. -When saving a file with multiple embers in it, detect if time values are all the same and if so, start them at zero and increment by 1 for each ember. -Allow for numerous quality increases to be coalesced into one. It will pick up at the end of the current render. -Show selection highlight on variations tree in response to mouse hover. This makes it easier to see for which variation or param the current mouse wheel action will apply. -Make default temporal samples be 100, whereas before it was 1000 which was overkill. -Require the shift key to be held with delete for deleting an ember to prevent it from triggering when the user enters delete in the edit box. -This wasn't otherwise fixable without writing a lot more code. --Bug fixes -EmberGenome was crashing when generating a sequence from a source file with more than 2 embers in it. -EmberGenome was improperly handling the first frame of a merge after the last frame of the loop. -These bugs were due to a previous commit. Revert parts of that commit. -Prevent a zoom value of less than 0 when reading from xml. -Slight optimization of the crescents, and mask variations, if the compiler wasn't doing it already. -Unique file naming was broken because it was looking for _# and the default names ended with -#. -Disallow renaming of an ember in the library tree to an empty string. -Severe bug that prevented some variations from being read correctly from params generated outside this program. -Severe OpenCL randomization bug. The first x coordinates of the first points in the first kernel call of the first ember of a render since the OpenCL renderer object was created were not random and were mostly -1. -Severe bug when populating xform selection distributions that could sometimes cause a crash due to roundoff error. Fix by using double. -Limit the max number of variations in a random ember to MAX_CL_VARS, which is 8. This ensures they'll look the same on CPU and GPU. -Prevent user from saving stylesheet to default.qss, it's a special reserved filename. --Code changes -Generalize using the running sum output point inside of a variation for all cases: pre, reg and post. -Allow for array variables in variations where the address of each element is stored in m_Params. -Qualify all math functions with std:: -No longer use our own Clamp() in OpenCL, instead use the standard clamp(). -Redesign how functions are used in the variations OpenCL code. -Add tests to EmberTester to verify some of the new functionality. -Place more const and override qualifiers on functions where appropriate. -Add a global rand with a lock to be used very sparingly. -Use a map instead of a vector for bad param names in Xml parsing. -Prefix affine interpolation mode defines with "AFFINE_" to make their purpose more clear. -Allow for variations that change state during iteration by sending a separate copy of the ember to each rendering thread. -Implement this same functionality with a local struct in OpenCL. It's members are the total of all variables that need to change state within an ember. -Add Contains() function to Utils.h. -EmberRender: print names of kernels being printed with --dump_kernel option. -Clean up EmberTester to handle some of the recent changes. -Fix various casts. -Replace % 2 with & 1, even though the compiler was likely doing this already. -Add new file Variations06.h to accommodate new variations. -General cleanup.
2015-11-22 17:15:07 -05:00
return a * std::pow(std::sin(b * x + c) < 0 ? EPS : T(1), p);//Original passed -1 to pow if sin() was < 0. Doing so will return NaN, so EPS is passed instead.
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
}
T m_A1;
T m_B1;
T m_C1;
T m_P1;
T m_T1;
T m_A2;
T m_B2;
T m_C2;
T m_P2;
T m_T2;
};
/// <summary>
/// sinq.
/// </summary>
template <typename T>
--Code changes -Split VariationList into .h/cpp files. --Make VariationList.cpp the only place where the variations files are included from. This speeds up the build, but variations can no longer be directly instantiated. --Variations are no longer exported.
2016-02-18 21:58:24 -05:00
class SinqVariation : public Variation<T>
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
public:
--Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums.
2016-01-04 19:50:15 -05:00
SinqVariation(T weight = 1.0) : Variation<T>("sinq", eVariationId::VAR_SINQ, weight) { }
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
VARCOPY(SinqVariation)
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
--Code changes -Make usage of VarFuncs more consistent by moving some global functions there. -Make CircleTrans1 a little safer by calling Zeps() on m_Sc during precalc. -Begin prepping for 0.9.9.5 release.
2016-03-04 21:54:06 -05:00
T absV = VarFuncs<T>::Hypot(helper.In.y, helper.In.z);
--User changes -Add new variations: bubbleT3D, crob, hexaplay3D, hexcrop, hexes, hexnix3D, loonie2, loonie3, nBlur, octapol and synth. -Allow for pre/post versions of dc_bubble, dc_cylinder and dc_linear whereas before they were omitted. -When saving a file with multiple embers in it, detect if time values are all the same and if so, start them at zero and increment by 1 for each ember. -Allow for numerous quality increases to be coalesced into one. It will pick up at the end of the current render. -Show selection highlight on variations tree in response to mouse hover. This makes it easier to see for which variation or param the current mouse wheel action will apply. -Make default temporal samples be 100, whereas before it was 1000 which was overkill. -Require the shift key to be held with delete for deleting an ember to prevent it from triggering when the user enters delete in the edit box. -This wasn't otherwise fixable without writing a lot more code. --Bug fixes -EmberGenome was crashing when generating a sequence from a source file with more than 2 embers in it. -EmberGenome was improperly handling the first frame of a merge after the last frame of the loop. -These bugs were due to a previous commit. Revert parts of that commit. -Prevent a zoom value of less than 0 when reading from xml. -Slight optimization of the crescents, and mask variations, if the compiler wasn't doing it already. -Unique file naming was broken because it was looking for _# and the default names ended with -#. -Disallow renaming of an ember in the library tree to an empty string. -Severe bug that prevented some variations from being read correctly from params generated outside this program. -Severe OpenCL randomization bug. The first x coordinates of the first points in the first kernel call of the first ember of a render since the OpenCL renderer object was created were not random and were mostly -1. -Severe bug when populating xform selection distributions that could sometimes cause a crash due to roundoff error. Fix by using double. -Limit the max number of variations in a random ember to MAX_CL_VARS, which is 8. This ensures they'll look the same on CPU and GPU. -Prevent user from saving stylesheet to default.qss, it's a special reserved filename. --Code changes -Generalize using the running sum output point inside of a variation for all cases: pre, reg and post. -Allow for array variables in variations where the address of each element is stored in m_Params. -Qualify all math functions with std:: -No longer use our own Clamp() in OpenCL, instead use the standard clamp(). -Redesign how functions are used in the variations OpenCL code. -Add tests to EmberTester to verify some of the new functionality. -Place more const and override qualifiers on functions where appropriate. -Add a global rand with a lock to be used very sparingly. -Use a map instead of a vector for bad param names in Xml parsing. -Prefix affine interpolation mode defines with "AFFINE_" to make their purpose more clear. -Allow for variations that change state during iteration by sending a separate copy of the ember to each rendering thread. -Implement this same functionality with a local struct in OpenCL. It's members are the total of all variables that need to change state within an ember. -Add Contains() function to Utils.h. -EmberRender: print names of kernels being printed with --dump_kernel option. -Clean up EmberTester to handle some of the recent changes. -Fix various casts. -Replace % 2 with & 1, even though the compiler was likely doing this already. -Add new file Variations06.h to accommodate new variations. -General cleanup.
2015-11-22 17:15:07 -05:00
T s = std::sin(helper.In.x);
T c = std::cos(helper.In.x);
T sh = std::sinh(absV);
T ch = std::cosh(absV);
06/04/2017 --User changes -Make all fonts be MS Shell Dlg 2. This will require reloading dark.qss if users are already using it. -Limit size of the left side of the palette editor. -Disable create from image buttons in the palette editor when working on a fixed palette. --Bug fixes -The following variations were wrong: coshq, cothq. -During iteration, the color index could become nan if all xform color speeds were negative. This could lead to bad results on the GPU. Fix to check for nan. Minimal speed difference. --Code changes -Make the following variations safer by using Zeps(): sinq, sinhq, secq, sechq, tanq, tanhq, cosq, coshq, cotq, cothq, cscq, cschq, estiq. -Always pass -cl-no-signed-zeros -cl-denorms-are-zero to kernel compiles for both single and double. -Flush all denormals to zero for all executable programs. This will likely lead to a speedup for badly behaving programs.
2017-06-04 20:37:29 -04:00
T d = m_Weight * c * sh / Zeps(absV);
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
helper.Out.x = m_Weight * s * ch;
helper.Out.y = d * helper.In.y;
helper.Out.z = d * helper.In.z;
}
--Code changes -Better use of const functions. -Ember.pro should not define BUILDING_EMBERCL, instead define BUILDING_EMBER.
2015-06-29 23:13:53 -04:00
virtual string OpenCLString() const override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
ostringstream ss;
0.4.1.3 Beta 10/14/2014 --User Changes Size is no longer fixed to the window size. Size scaling is done differently in the final render dialog. This fixes several bugs. Remove Xml saving size from settings and options dialog, it no longer applies. Final render can be broken into strips. Set default save path to the desktop if none is found in the settings file. Set default output size to 1920x1080 if none is found in the settings file. --Bug Fixes Better memory size reporting in final render dialog. --Code Changes Migrate to C++11, Qt 5.3.1, and Visual Studio 2013. Change most instances of unsigned int to size_t, and int to intmax_t. Add m_OrigPixPerUnit and m_ScaleType to Ember for scaling purposes. Replace some sprintf_s() calls in XmlToEmber with ostringstream. Move more non-templated members into RendererBase. Add CopyVec() overload that takes a per element function pointer. Add vector Memset(). Replace '&' with '+' instead of "&amp;" in XmlToEmber for much faster parsing. Break strips rendering out into EmberCommon and call from EmberRender and Fractorium. Make AddAndWriteBuffer() just call WriteBuffer(). Make AddAndWriteImage() delete the existing image first before replacing it. Add SetOutputTexture() to RendererCL to support making new textures in response to resize events. Remove multiple return statements in RendererCL, and replace with a bool that tracks results. Add ToDouble(), MakeEnd(), ToString() and Exists() wrappers in Fractorium. Add Size() wrapper in EmberFile. Make QString function arguments const QString&, and string with const string&. Make ShowCritical() wrapper for invoking a message box from another thread. Add combo box to TwoButtonWidget and rename.
2014-10-14 11:53:15 -04:00
intmax_t varIndex = IndexInXform();
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
string weight = WeightDefineString();
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
ss << "\t{\n"
<< "\t\treal_t absV = Hypot(vIn.y, vIn.z);\n"
<< "\t\treal_t s = sin(vIn.x);\n"
<< "\t\treal_t c = cos(vIn.x);\n"
<< "\t\treal_t sh = sinh(absV);\n"
<< "\t\treal_t ch = cosh(absV);\n"
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
<< "\t\treal_t d = " << weight << " * c * sh / Zeps(absV);\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\n"
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
<< "\t\tvOut.x = " << weight << " * s * ch;\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t\tvOut.y = d * vIn.y;\n"
<< "\t\tvOut.z = d * vIn.z;\n"
<< "\t}\n";
return ss.str();
}
--User changes -Add new variations: bubbleT3D, crob, hexaplay3D, hexcrop, hexes, hexnix3D, loonie2, loonie3, nBlur, octapol and synth. -Allow for pre/post versions of dc_bubble, dc_cylinder and dc_linear whereas before they were omitted. -When saving a file with multiple embers in it, detect if time values are all the same and if so, start them at zero and increment by 1 for each ember. -Allow for numerous quality increases to be coalesced into one. It will pick up at the end of the current render. -Show selection highlight on variations tree in response to mouse hover. This makes it easier to see for which variation or param the current mouse wheel action will apply. -Make default temporal samples be 100, whereas before it was 1000 which was overkill. -Require the shift key to be held with delete for deleting an ember to prevent it from triggering when the user enters delete in the edit box. -This wasn't otherwise fixable without writing a lot more code. --Bug fixes -EmberGenome was crashing when generating a sequence from a source file with more than 2 embers in it. -EmberGenome was improperly handling the first frame of a merge after the last frame of the loop. -These bugs were due to a previous commit. Revert parts of that commit. -Prevent a zoom value of less than 0 when reading from xml. -Slight optimization of the crescents, and mask variations, if the compiler wasn't doing it already. -Unique file naming was broken because it was looking for _# and the default names ended with -#. -Disallow renaming of an ember in the library tree to an empty string. -Severe bug that prevented some variations from being read correctly from params generated outside this program. -Severe OpenCL randomization bug. The first x coordinates of the first points in the first kernel call of the first ember of a render since the OpenCL renderer object was created were not random and were mostly -1. -Severe bug when populating xform selection distributions that could sometimes cause a crash due to roundoff error. Fix by using double. -Limit the max number of variations in a random ember to MAX_CL_VARS, which is 8. This ensures they'll look the same on CPU and GPU. -Prevent user from saving stylesheet to default.qss, it's a special reserved filename. --Code changes -Generalize using the running sum output point inside of a variation for all cases: pre, reg and post. -Allow for array variables in variations where the address of each element is stored in m_Params. -Qualify all math functions with std:: -No longer use our own Clamp() in OpenCL, instead use the standard clamp(). -Redesign how functions are used in the variations OpenCL code. -Add tests to EmberTester to verify some of the new functionality. -Place more const and override qualifiers on functions where appropriate. -Add a global rand with a lock to be used very sparingly. -Use a map instead of a vector for bad param names in Xml parsing. -Prefix affine interpolation mode defines with "AFFINE_" to make their purpose more clear. -Allow for variations that change state during iteration by sending a separate copy of the ember to each rendering thread. -Implement this same functionality with a local struct in OpenCL. It's members are the total of all variables that need to change state within an ember. -Add Contains() function to Utils.h. -EmberRender: print names of kernels being printed with --dump_kernel option. -Clean up EmberTester to handle some of the recent changes. -Fix various casts. -Replace % 2 with & 1, even though the compiler was likely doing this already. -Add new file Variations06.h to accommodate new variations. -General cleanup.
2015-11-22 17:15:07 -05:00
virtual vector<string> OpenCLGlobalFuncNames() const override
{
06/04/2017 --User changes -Make all fonts be MS Shell Dlg 2. This will require reloading dark.qss if users are already using it. -Limit size of the left side of the palette editor. -Disable create from image buttons in the palette editor when working on a fixed palette. --Bug fixes -The following variations were wrong: coshq, cothq. -During iteration, the color index could become nan if all xform color speeds were negative. This could lead to bad results on the GPU. Fix to check for nan. Minimal speed difference. --Code changes -Make the following variations safer by using Zeps(): sinq, sinhq, secq, sechq, tanq, tanhq, cosq, coshq, cotq, cothq, cscq, cschq, estiq. -Always pass -cl-no-signed-zeros -cl-denorms-are-zero to kernel compiles for both single and double. -Flush all denormals to zero for all executable programs. This will likely lead to a speedup for badly behaving programs.
2017-06-04 20:37:29 -04:00
return vector<string> { "Hypot", "Zeps" };
--User changes -Add new variations: bubbleT3D, crob, hexaplay3D, hexcrop, hexes, hexnix3D, loonie2, loonie3, nBlur, octapol and synth. -Allow for pre/post versions of dc_bubble, dc_cylinder and dc_linear whereas before they were omitted. -When saving a file with multiple embers in it, detect if time values are all the same and if so, start them at zero and increment by 1 for each ember. -Allow for numerous quality increases to be coalesced into one. It will pick up at the end of the current render. -Show selection highlight on variations tree in response to mouse hover. This makes it easier to see for which variation or param the current mouse wheel action will apply. -Make default temporal samples be 100, whereas before it was 1000 which was overkill. -Require the shift key to be held with delete for deleting an ember to prevent it from triggering when the user enters delete in the edit box. -This wasn't otherwise fixable without writing a lot more code. --Bug fixes -EmberGenome was crashing when generating a sequence from a source file with more than 2 embers in it. -EmberGenome was improperly handling the first frame of a merge after the last frame of the loop. -These bugs were due to a previous commit. Revert parts of that commit. -Prevent a zoom value of less than 0 when reading from xml. -Slight optimization of the crescents, and mask variations, if the compiler wasn't doing it already. -Unique file naming was broken because it was looking for _# and the default names ended with -#. -Disallow renaming of an ember in the library tree to an empty string. -Severe bug that prevented some variations from being read correctly from params generated outside this program. -Severe OpenCL randomization bug. The first x coordinates of the first points in the first kernel call of the first ember of a render since the OpenCL renderer object was created were not random and were mostly -1. -Severe bug when populating xform selection distributions that could sometimes cause a crash due to roundoff error. Fix by using double. -Limit the max number of variations in a random ember to MAX_CL_VARS, which is 8. This ensures they'll look the same on CPU and GPU. -Prevent user from saving stylesheet to default.qss, it's a special reserved filename. --Code changes -Generalize using the running sum output point inside of a variation for all cases: pre, reg and post. -Allow for array variables in variations where the address of each element is stored in m_Params. -Qualify all math functions with std:: -No longer use our own Clamp() in OpenCL, instead use the standard clamp(). -Redesign how functions are used in the variations OpenCL code. -Add tests to EmberTester to verify some of the new functionality. -Place more const and override qualifiers on functions where appropriate. -Add a global rand with a lock to be used very sparingly. -Use a map instead of a vector for bad param names in Xml parsing. -Prefix affine interpolation mode defines with "AFFINE_" to make their purpose more clear. -Allow for variations that change state during iteration by sending a separate copy of the ember to each rendering thread. -Implement this same functionality with a local struct in OpenCL. It's members are the total of all variables that need to change state within an ember. -Add Contains() function to Utils.h. -EmberRender: print names of kernels being printed with --dump_kernel option. -Clean up EmberTester to handle some of the recent changes. -Fix various casts. -Replace % 2 with & 1, even though the compiler was likely doing this already. -Add new file Variations06.h to accommodate new variations. -General cleanup.
2015-11-22 17:15:07 -05:00
}
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
};
/// <summary>
/// sinhq.
/// </summary>
template <typename T>
--Code changes -Split VariationList into .h/cpp files. --Make VariationList.cpp the only place where the variations files are included from. This speeds up the build, but variations can no longer be directly instantiated. --Variations are no longer exported.
2016-02-18 21:58:24 -05:00
class SinhqVariation : public Variation<T>
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
public:
--Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums.
2016-01-04 19:50:15 -05:00
SinhqVariation(T weight = 1.0) : Variation<T>("sinhq", eVariationId::VAR_SINHQ, weight) { }
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
VARCOPY(SinhqVariation)
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
--Code changes -Make usage of VarFuncs more consistent by moving some global functions there. -Make CircleTrans1 a little safer by calling Zeps() on m_Sc during precalc. -Begin prepping for 0.9.9.5 release.
2016-03-04 21:54:06 -05:00
T absV = VarFuncs<T>::Hypot(helper.In.y, helper.In.z);
--User changes -Add new variations: bubbleT3D, crob, hexaplay3D, hexcrop, hexes, hexnix3D, loonie2, loonie3, nBlur, octapol and synth. -Allow for pre/post versions of dc_bubble, dc_cylinder and dc_linear whereas before they were omitted. -When saving a file with multiple embers in it, detect if time values are all the same and if so, start them at zero and increment by 1 for each ember. -Allow for numerous quality increases to be coalesced into one. It will pick up at the end of the current render. -Show selection highlight on variations tree in response to mouse hover. This makes it easier to see for which variation or param the current mouse wheel action will apply. -Make default temporal samples be 100, whereas before it was 1000 which was overkill. -Require the shift key to be held with delete for deleting an ember to prevent it from triggering when the user enters delete in the edit box. -This wasn't otherwise fixable without writing a lot more code. --Bug fixes -EmberGenome was crashing when generating a sequence from a source file with more than 2 embers in it. -EmberGenome was improperly handling the first frame of a merge after the last frame of the loop. -These bugs were due to a previous commit. Revert parts of that commit. -Prevent a zoom value of less than 0 when reading from xml. -Slight optimization of the crescents, and mask variations, if the compiler wasn't doing it already. -Unique file naming was broken because it was looking for _# and the default names ended with -#. -Disallow renaming of an ember in the library tree to an empty string. -Severe bug that prevented some variations from being read correctly from params generated outside this program. -Severe OpenCL randomization bug. The first x coordinates of the first points in the first kernel call of the first ember of a render since the OpenCL renderer object was created were not random and were mostly -1. -Severe bug when populating xform selection distributions that could sometimes cause a crash due to roundoff error. Fix by using double. -Limit the max number of variations in a random ember to MAX_CL_VARS, which is 8. This ensures they'll look the same on CPU and GPU. -Prevent user from saving stylesheet to default.qss, it's a special reserved filename. --Code changes -Generalize using the running sum output point inside of a variation for all cases: pre, reg and post. -Allow for array variables in variations where the address of each element is stored in m_Params. -Qualify all math functions with std:: -No longer use our own Clamp() in OpenCL, instead use the standard clamp(). -Redesign how functions are used in the variations OpenCL code. -Add tests to EmberTester to verify some of the new functionality. -Place more const and override qualifiers on functions where appropriate. -Add a global rand with a lock to be used very sparingly. -Use a map instead of a vector for bad param names in Xml parsing. -Prefix affine interpolation mode defines with "AFFINE_" to make their purpose more clear. -Allow for variations that change state during iteration by sending a separate copy of the ember to each rendering thread. -Implement this same functionality with a local struct in OpenCL. It's members are the total of all variables that need to change state within an ember. -Add Contains() function to Utils.h. -EmberRender: print names of kernels being printed with --dump_kernel option. -Clean up EmberTester to handle some of the recent changes. -Fix various casts. -Replace % 2 with & 1, even though the compiler was likely doing this already. -Add new file Variations06.h to accommodate new variations. -General cleanup.
2015-11-22 17:15:07 -05:00
T s = std::sin(absV);
T c = std::cos(absV);
T sh = std::sinh(helper.In.x);
T ch = std::cosh(helper.In.x);
06/04/2017 --User changes -Make all fonts be MS Shell Dlg 2. This will require reloading dark.qss if users are already using it. -Limit size of the left side of the palette editor. -Disable create from image buttons in the palette editor when working on a fixed palette. --Bug fixes -The following variations were wrong: coshq, cothq. -During iteration, the color index could become nan if all xform color speeds were negative. This could lead to bad results on the GPU. Fix to check for nan. Minimal speed difference. --Code changes -Make the following variations safer by using Zeps(): sinq, sinhq, secq, sechq, tanq, tanhq, cosq, coshq, cotq, cothq, cscq, cschq, estiq. -Always pass -cl-no-signed-zeros -cl-denorms-are-zero to kernel compiles for both single and double. -Flush all denormals to zero for all executable programs. This will likely lead to a speedup for badly behaving programs.
2017-06-04 20:37:29 -04:00
T d = m_Weight * ch * s / Zeps(absV);
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
helper.Out.x = m_Weight * sh * c;
helper.Out.y = d * helper.In.y;
helper.Out.z = d * helper.In.z;
}
--Code changes -Better use of const functions. -Ember.pro should not define BUILDING_EMBERCL, instead define BUILDING_EMBER.
2015-06-29 23:13:53 -04:00
virtual string OpenCLString() const override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
ostringstream ss;
0.4.1.3 Beta 10/14/2014 --User Changes Size is no longer fixed to the window size. Size scaling is done differently in the final render dialog. This fixes several bugs. Remove Xml saving size from settings and options dialog, it no longer applies. Final render can be broken into strips. Set default save path to the desktop if none is found in the settings file. Set default output size to 1920x1080 if none is found in the settings file. --Bug Fixes Better memory size reporting in final render dialog. --Code Changes Migrate to C++11, Qt 5.3.1, and Visual Studio 2013. Change most instances of unsigned int to size_t, and int to intmax_t. Add m_OrigPixPerUnit and m_ScaleType to Ember for scaling purposes. Replace some sprintf_s() calls in XmlToEmber with ostringstream. Move more non-templated members into RendererBase. Add CopyVec() overload that takes a per element function pointer. Add vector Memset(). Replace '&' with '+' instead of "&amp;" in XmlToEmber for much faster parsing. Break strips rendering out into EmberCommon and call from EmberRender and Fractorium. Make AddAndWriteBuffer() just call WriteBuffer(). Make AddAndWriteImage() delete the existing image first before replacing it. Add SetOutputTexture() to RendererCL to support making new textures in response to resize events. Remove multiple return statements in RendererCL, and replace with a bool that tracks results. Add ToDouble(), MakeEnd(), ToString() and Exists() wrappers in Fractorium. Add Size() wrapper in EmberFile. Make QString function arguments const QString&, and string with const string&. Make ShowCritical() wrapper for invoking a message box from another thread. Add combo box to TwoButtonWidget and rename.
2014-10-14 11:53:15 -04:00
intmax_t varIndex = IndexInXform();
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
string weight = WeightDefineString();
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
ss << "\t{\n"
<< "\t\treal_t absV = Hypot(vIn.y, vIn.z);\n"
<< "\t\treal_t s = sin(absV);\n"
<< "\t\treal_t c = cos(absV);\n"
<< "\t\treal_t sh = sinh(vIn.x);\n"
<< "\t\treal_t ch = cosh(vIn.x);\n"
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
<< "\t\treal_t d = " << weight << " * ch * s / Zeps(absV);\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\n"
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
<< "\t\tvOut.x = " << weight << " * sh * c;\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t\tvOut.y = d * vIn.y;\n"
<< "\t\tvOut.z = d * vIn.z;\n"
<< "\t}\n";
return ss.str();
}
--User changes -Add new variations: bubbleT3D, crob, hexaplay3D, hexcrop, hexes, hexnix3D, loonie2, loonie3, nBlur, octapol and synth. -Allow for pre/post versions of dc_bubble, dc_cylinder and dc_linear whereas before they were omitted. -When saving a file with multiple embers in it, detect if time values are all the same and if so, start them at zero and increment by 1 for each ember. -Allow for numerous quality increases to be coalesced into one. It will pick up at the end of the current render. -Show selection highlight on variations tree in response to mouse hover. This makes it easier to see for which variation or param the current mouse wheel action will apply. -Make default temporal samples be 100, whereas before it was 1000 which was overkill. -Require the shift key to be held with delete for deleting an ember to prevent it from triggering when the user enters delete in the edit box. -This wasn't otherwise fixable without writing a lot more code. --Bug fixes -EmberGenome was crashing when generating a sequence from a source file with more than 2 embers in it. -EmberGenome was improperly handling the first frame of a merge after the last frame of the loop. -These bugs were due to a previous commit. Revert parts of that commit. -Prevent a zoom value of less than 0 when reading from xml. -Slight optimization of the crescents, and mask variations, if the compiler wasn't doing it already. -Unique file naming was broken because it was looking for _# and the default names ended with -#. -Disallow renaming of an ember in the library tree to an empty string. -Severe bug that prevented some variations from being read correctly from params generated outside this program. -Severe OpenCL randomization bug. The first x coordinates of the first points in the first kernel call of the first ember of a render since the OpenCL renderer object was created were not random and were mostly -1. -Severe bug when populating xform selection distributions that could sometimes cause a crash due to roundoff error. Fix by using double. -Limit the max number of variations in a random ember to MAX_CL_VARS, which is 8. This ensures they'll look the same on CPU and GPU. -Prevent user from saving stylesheet to default.qss, it's a special reserved filename. --Code changes -Generalize using the running sum output point inside of a variation for all cases: pre, reg and post. -Allow for array variables in variations where the address of each element is stored in m_Params. -Qualify all math functions with std:: -No longer use our own Clamp() in OpenCL, instead use the standard clamp(). -Redesign how functions are used in the variations OpenCL code. -Add tests to EmberTester to verify some of the new functionality. -Place more const and override qualifiers on functions where appropriate. -Add a global rand with a lock to be used very sparingly. -Use a map instead of a vector for bad param names in Xml parsing. -Prefix affine interpolation mode defines with "AFFINE_" to make their purpose more clear. -Allow for variations that change state during iteration by sending a separate copy of the ember to each rendering thread. -Implement this same functionality with a local struct in OpenCL. It's members are the total of all variables that need to change state within an ember. -Add Contains() function to Utils.h. -EmberRender: print names of kernels being printed with --dump_kernel option. -Clean up EmberTester to handle some of the recent changes. -Fix various casts. -Replace % 2 with & 1, even though the compiler was likely doing this already. -Add new file Variations06.h to accommodate new variations. -General cleanup.
2015-11-22 17:15:07 -05:00
virtual vector<string> OpenCLGlobalFuncNames() const override
{
06/04/2017 --User changes -Make all fonts be MS Shell Dlg 2. This will require reloading dark.qss if users are already using it. -Limit size of the left side of the palette editor. -Disable create from image buttons in the palette editor when working on a fixed palette. --Bug fixes -The following variations were wrong: coshq, cothq. -During iteration, the color index could become nan if all xform color speeds were negative. This could lead to bad results on the GPU. Fix to check for nan. Minimal speed difference. --Code changes -Make the following variations safer by using Zeps(): sinq, sinhq, secq, sechq, tanq, tanhq, cosq, coshq, cotq, cothq, cscq, cschq, estiq. -Always pass -cl-no-signed-zeros -cl-denorms-are-zero to kernel compiles for both single and double. -Flush all denormals to zero for all executable programs. This will likely lead to a speedup for badly behaving programs.
2017-06-04 20:37:29 -04:00
return vector<string> { "Hypot", "Zeps" };
--User changes -Add new variations: bubbleT3D, crob, hexaplay3D, hexcrop, hexes, hexnix3D, loonie2, loonie3, nBlur, octapol and synth. -Allow for pre/post versions of dc_bubble, dc_cylinder and dc_linear whereas before they were omitted. -When saving a file with multiple embers in it, detect if time values are all the same and if so, start them at zero and increment by 1 for each ember. -Allow for numerous quality increases to be coalesced into one. It will pick up at the end of the current render. -Show selection highlight on variations tree in response to mouse hover. This makes it easier to see for which variation or param the current mouse wheel action will apply. -Make default temporal samples be 100, whereas before it was 1000 which was overkill. -Require the shift key to be held with delete for deleting an ember to prevent it from triggering when the user enters delete in the edit box. -This wasn't otherwise fixable without writing a lot more code. --Bug fixes -EmberGenome was crashing when generating a sequence from a source file with more than 2 embers in it. -EmberGenome was improperly handling the first frame of a merge after the last frame of the loop. -These bugs were due to a previous commit. Revert parts of that commit. -Prevent a zoom value of less than 0 when reading from xml. -Slight optimization of the crescents, and mask variations, if the compiler wasn't doing it already. -Unique file naming was broken because it was looking for _# and the default names ended with -#. -Disallow renaming of an ember in the library tree to an empty string. -Severe bug that prevented some variations from being read correctly from params generated outside this program. -Severe OpenCL randomization bug. The first x coordinates of the first points in the first kernel call of the first ember of a render since the OpenCL renderer object was created were not random and were mostly -1. -Severe bug when populating xform selection distributions that could sometimes cause a crash due to roundoff error. Fix by using double. -Limit the max number of variations in a random ember to MAX_CL_VARS, which is 8. This ensures they'll look the same on CPU and GPU. -Prevent user from saving stylesheet to default.qss, it's a special reserved filename. --Code changes -Generalize using the running sum output point inside of a variation for all cases: pre, reg and post. -Allow for array variables in variations where the address of each element is stored in m_Params. -Qualify all math functions with std:: -No longer use our own Clamp() in OpenCL, instead use the standard clamp(). -Redesign how functions are used in the variations OpenCL code. -Add tests to EmberTester to verify some of the new functionality. -Place more const and override qualifiers on functions where appropriate. -Add a global rand with a lock to be used very sparingly. -Use a map instead of a vector for bad param names in Xml parsing. -Prefix affine interpolation mode defines with "AFFINE_" to make their purpose more clear. -Allow for variations that change state during iteration by sending a separate copy of the ember to each rendering thread. -Implement this same functionality with a local struct in OpenCL. It's members are the total of all variables that need to change state within an ember. -Add Contains() function to Utils.h. -EmberRender: print names of kernels being printed with --dump_kernel option. -Clean up EmberTester to handle some of the recent changes. -Fix various casts. -Replace % 2 with & 1, even though the compiler was likely doing this already. -Add new file Variations06.h to accommodate new variations. -General cleanup.
2015-11-22 17:15:07 -05:00
}
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
};
/// <summary>
/// secq.
/// </summary>
template <typename T>
--Code changes -Split VariationList into .h/cpp files. --Make VariationList.cpp the only place where the variations files are included from. This speeds up the build, but variations can no longer be directly instantiated. --Variations are no longer exported.
2016-02-18 21:58:24 -05:00
class SecqVariation : public Variation<T>
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
public:
--Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums.
2016-01-04 19:50:15 -05:00
SecqVariation(T weight = 1.0) : Variation<T>("secq", eVariationId::VAR_SECQ, weight, true) { }
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
VARCOPY(SecqVariation)
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
--Code changes -Make usage of VarFuncs more consistent by moving some global functions there. -Make CircleTrans1 a little safer by calling Zeps() on m_Sc during precalc. -Begin prepping for 0.9.9.5 release.
2016-03-04 21:54:06 -05:00
T absV = VarFuncs<T>::Hypot(helper.In.y, helper.In.z);
06/04/2017 --User changes -Make all fonts be MS Shell Dlg 2. This will require reloading dark.qss if users are already using it. -Limit size of the left side of the palette editor. -Disable create from image buttons in the palette editor when working on a fixed palette. --Bug fixes -The following variations were wrong: coshq, cothq. -During iteration, the color index could become nan if all xform color speeds were negative. This could lead to bad results on the GPU. Fix to check for nan. Minimal speed difference. --Code changes -Make the following variations safer by using Zeps(): sinq, sinhq, secq, sechq, tanq, tanhq, cosq, coshq, cotq, cothq, cscq, cschq, estiq. -Always pass -cl-no-signed-zeros -cl-denorms-are-zero to kernel compiles for both single and double. -Flush all denormals to zero for all executable programs. This will likely lead to a speedup for badly behaving programs.
2017-06-04 20:37:29 -04:00
T ni = m_Weight / Zeps(helper.m_PrecalcSumSquares + SQR(helper.In.z));
--User changes -Add new variations: bubbleT3D, crob, hexaplay3D, hexcrop, hexes, hexnix3D, loonie2, loonie3, nBlur, octapol and synth. -Allow for pre/post versions of dc_bubble, dc_cylinder and dc_linear whereas before they were omitted. -When saving a file with multiple embers in it, detect if time values are all the same and if so, start them at zero and increment by 1 for each ember. -Allow for numerous quality increases to be coalesced into one. It will pick up at the end of the current render. -Show selection highlight on variations tree in response to mouse hover. This makes it easier to see for which variation or param the current mouse wheel action will apply. -Make default temporal samples be 100, whereas before it was 1000 which was overkill. -Require the shift key to be held with delete for deleting an ember to prevent it from triggering when the user enters delete in the edit box. -This wasn't otherwise fixable without writing a lot more code. --Bug fixes -EmberGenome was crashing when generating a sequence from a source file with more than 2 embers in it. -EmberGenome was improperly handling the first frame of a merge after the last frame of the loop. -These bugs were due to a previous commit. Revert parts of that commit. -Prevent a zoom value of less than 0 when reading from xml. -Slight optimization of the crescents, and mask variations, if the compiler wasn't doing it already. -Unique file naming was broken because it was looking for _# and the default names ended with -#. -Disallow renaming of an ember in the library tree to an empty string. -Severe bug that prevented some variations from being read correctly from params generated outside this program. -Severe OpenCL randomization bug. The first x coordinates of the first points in the first kernel call of the first ember of a render since the OpenCL renderer object was created were not random and were mostly -1. -Severe bug when populating xform selection distributions that could sometimes cause a crash due to roundoff error. Fix by using double. -Limit the max number of variations in a random ember to MAX_CL_VARS, which is 8. This ensures they'll look the same on CPU and GPU. -Prevent user from saving stylesheet to default.qss, it's a special reserved filename. --Code changes -Generalize using the running sum output point inside of a variation for all cases: pre, reg and post. -Allow for array variables in variations where the address of each element is stored in m_Params. -Qualify all math functions with std:: -No longer use our own Clamp() in OpenCL, instead use the standard clamp(). -Redesign how functions are used in the variations OpenCL code. -Add tests to EmberTester to verify some of the new functionality. -Place more const and override qualifiers on functions where appropriate. -Add a global rand with a lock to be used very sparingly. -Use a map instead of a vector for bad param names in Xml parsing. -Prefix affine interpolation mode defines with "AFFINE_" to make their purpose more clear. -Allow for variations that change state during iteration by sending a separate copy of the ember to each rendering thread. -Implement this same functionality with a local struct in OpenCL. It's members are the total of all variables that need to change state within an ember. -Add Contains() function to Utils.h. -EmberRender: print names of kernels being printed with --dump_kernel option. -Clean up EmberTester to handle some of the recent changes. -Fix various casts. -Replace % 2 with & 1, even though the compiler was likely doing this already. -Add new file Variations06.h to accommodate new variations. -General cleanup.
2015-11-22 17:15:07 -05:00
T s = std::sin(-helper.In.x);
T c = std::cos(-helper.In.x);
T sh = std::sinh(absV);
T ch = std::cosh(absV);
06/04/2017 --User changes -Make all fonts be MS Shell Dlg 2. This will require reloading dark.qss if users are already using it. -Limit size of the left side of the palette editor. -Disable create from image buttons in the palette editor when working on a fixed palette. --Bug fixes -The following variations were wrong: coshq, cothq. -During iteration, the color index could become nan if all xform color speeds were negative. This could lead to bad results on the GPU. Fix to check for nan. Minimal speed difference. --Code changes -Make the following variations safer by using Zeps(): sinq, sinhq, secq, sechq, tanq, tanhq, cosq, coshq, cotq, cothq, cscq, cschq, estiq. -Always pass -cl-no-signed-zeros -cl-denorms-are-zero to kernel compiles for both single and double. -Flush all denormals to zero for all executable programs. This will likely lead to a speedup for badly behaving programs.
2017-06-04 20:37:29 -04:00
T d = ni * s * sh / Zeps(absV);
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
helper.Out.x = c * ch * ni;
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
helper.Out.y = -(d * helper.In.y);
helper.Out.z = -(d * helper.In.z);
}
--Code changes -Better use of const functions. -Ember.pro should not define BUILDING_EMBERCL, instead define BUILDING_EMBER.
2015-06-29 23:13:53 -04:00
virtual string OpenCLString() const override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
ostringstream ss;
0.4.1.3 Beta 10/14/2014 --User Changes Size is no longer fixed to the window size. Size scaling is done differently in the final render dialog. This fixes several bugs. Remove Xml saving size from settings and options dialog, it no longer applies. Final render can be broken into strips. Set default save path to the desktop if none is found in the settings file. Set default output size to 1920x1080 if none is found in the settings file. --Bug Fixes Better memory size reporting in final render dialog. --Code Changes Migrate to C++11, Qt 5.3.1, and Visual Studio 2013. Change most instances of unsigned int to size_t, and int to intmax_t. Add m_OrigPixPerUnit and m_ScaleType to Ember for scaling purposes. Replace some sprintf_s() calls in XmlToEmber with ostringstream. Move more non-templated members into RendererBase. Add CopyVec() overload that takes a per element function pointer. Add vector Memset(). Replace '&' with '+' instead of "&amp;" in XmlToEmber for much faster parsing. Break strips rendering out into EmberCommon and call from EmberRender and Fractorium. Make AddAndWriteBuffer() just call WriteBuffer(). Make AddAndWriteImage() delete the existing image first before replacing it. Add SetOutputTexture() to RendererCL to support making new textures in response to resize events. Remove multiple return statements in RendererCL, and replace with a bool that tracks results. Add ToDouble(), MakeEnd(), ToString() and Exists() wrappers in Fractorium. Add Size() wrapper in EmberFile. Make QString function arguments const QString&, and string with const string&. Make ShowCritical() wrapper for invoking a message box from another thread. Add combo box to TwoButtonWidget and rename.
2014-10-14 11:53:15 -04:00
intmax_t varIndex = IndexInXform();
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
string weight = WeightDefineString();
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
ss << "\t{\n"
<< "\t\treal_t absV = Hypot(vIn.y, vIn.z);\n"
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
<< "\t\treal_t ni = " << weight << " / Zeps(fma(vIn.z, vIn.z, precalcSumSquares));\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t\treal_t s = sin(-vIn.x);\n"
<< "\t\treal_t c = cos(-vIn.x);\n"
<< "\t\treal_t sh = sinh(absV);\n"
<< "\t\treal_t ch = cosh(absV);\n"
06/04/2017 --User changes -Make all fonts be MS Shell Dlg 2. This will require reloading dark.qss if users are already using it. -Limit size of the left side of the palette editor. -Disable create from image buttons in the palette editor when working on a fixed palette. --Bug fixes -The following variations were wrong: coshq, cothq. -During iteration, the color index could become nan if all xform color speeds were negative. This could lead to bad results on the GPU. Fix to check for nan. Minimal speed difference. --Code changes -Make the following variations safer by using Zeps(): sinq, sinhq, secq, sechq, tanq, tanhq, cosq, coshq, cotq, cothq, cscq, cschq, estiq. -Always pass -cl-no-signed-zeros -cl-denorms-are-zero to kernel compiles for both single and double. -Flush all denormals to zero for all executable programs. This will likely lead to a speedup for badly behaving programs.
2017-06-04 20:37:29 -04:00
<< "\t\treal_t d = ni * s * sh / Zeps(absV);\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\n"
<< "\t\tvOut.x = c * ch * ni;\n"
<< "\t\tvOut.y = -(d * vIn.y);\n"
<< "\t\tvOut.z = -(d * vIn.z);\n"
<< "\t}\n";
return ss.str();
}
--User changes -Add new variations: bubbleT3D, crob, hexaplay3D, hexcrop, hexes, hexnix3D, loonie2, loonie3, nBlur, octapol and synth. -Allow for pre/post versions of dc_bubble, dc_cylinder and dc_linear whereas before they were omitted. -When saving a file with multiple embers in it, detect if time values are all the same and if so, start them at zero and increment by 1 for each ember. -Allow for numerous quality increases to be coalesced into one. It will pick up at the end of the current render. -Show selection highlight on variations tree in response to mouse hover. This makes it easier to see for which variation or param the current mouse wheel action will apply. -Make default temporal samples be 100, whereas before it was 1000 which was overkill. -Require the shift key to be held with delete for deleting an ember to prevent it from triggering when the user enters delete in the edit box. -This wasn't otherwise fixable without writing a lot more code. --Bug fixes -EmberGenome was crashing when generating a sequence from a source file with more than 2 embers in it. -EmberGenome was improperly handling the first frame of a merge after the last frame of the loop. -These bugs were due to a previous commit. Revert parts of that commit. -Prevent a zoom value of less than 0 when reading from xml. -Slight optimization of the crescents, and mask variations, if the compiler wasn't doing it already. -Unique file naming was broken because it was looking for _# and the default names ended with -#. -Disallow renaming of an ember in the library tree to an empty string. -Severe bug that prevented some variations from being read correctly from params generated outside this program. -Severe OpenCL randomization bug. The first x coordinates of the first points in the first kernel call of the first ember of a render since the OpenCL renderer object was created were not random and were mostly -1. -Severe bug when populating xform selection distributions that could sometimes cause a crash due to roundoff error. Fix by using double. -Limit the max number of variations in a random ember to MAX_CL_VARS, which is 8. This ensures they'll look the same on CPU and GPU. -Prevent user from saving stylesheet to default.qss, it's a special reserved filename. --Code changes -Generalize using the running sum output point inside of a variation for all cases: pre, reg and post. -Allow for array variables in variations where the address of each element is stored in m_Params. -Qualify all math functions with std:: -No longer use our own Clamp() in OpenCL, instead use the standard clamp(). -Redesign how functions are used in the variations OpenCL code. -Add tests to EmberTester to verify some of the new functionality. -Place more const and override qualifiers on functions where appropriate. -Add a global rand with a lock to be used very sparingly. -Use a map instead of a vector for bad param names in Xml parsing. -Prefix affine interpolation mode defines with "AFFINE_" to make their purpose more clear. -Allow for variations that change state during iteration by sending a separate copy of the ember to each rendering thread. -Implement this same functionality with a local struct in OpenCL. It's members are the total of all variables that need to change state within an ember. -Add Contains() function to Utils.h. -EmberRender: print names of kernels being printed with --dump_kernel option. -Clean up EmberTester to handle some of the recent changes. -Fix various casts. -Replace % 2 with & 1, even though the compiler was likely doing this already. -Add new file Variations06.h to accommodate new variations. -General cleanup.
2015-11-22 17:15:07 -05:00
virtual vector<string> OpenCLGlobalFuncNames() const override
{
06/04/2017 --User changes -Make all fonts be MS Shell Dlg 2. This will require reloading dark.qss if users are already using it. -Limit size of the left side of the palette editor. -Disable create from image buttons in the palette editor when working on a fixed palette. --Bug fixes -The following variations were wrong: coshq, cothq. -During iteration, the color index could become nan if all xform color speeds were negative. This could lead to bad results on the GPU. Fix to check for nan. Minimal speed difference. --Code changes -Make the following variations safer by using Zeps(): sinq, sinhq, secq, sechq, tanq, tanhq, cosq, coshq, cotq, cothq, cscq, cschq, estiq. -Always pass -cl-no-signed-zeros -cl-denorms-are-zero to kernel compiles for both single and double. -Flush all denormals to zero for all executable programs. This will likely lead to a speedup for badly behaving programs.
2017-06-04 20:37:29 -04:00
return vector<string> { "Hypot", "Zeps" };
--User changes -Add new variations: bubbleT3D, crob, hexaplay3D, hexcrop, hexes, hexnix3D, loonie2, loonie3, nBlur, octapol and synth. -Allow for pre/post versions of dc_bubble, dc_cylinder and dc_linear whereas before they were omitted. -When saving a file with multiple embers in it, detect if time values are all the same and if so, start them at zero and increment by 1 for each ember. -Allow for numerous quality increases to be coalesced into one. It will pick up at the end of the current render. -Show selection highlight on variations tree in response to mouse hover. This makes it easier to see for which variation or param the current mouse wheel action will apply. -Make default temporal samples be 100, whereas before it was 1000 which was overkill. -Require the shift key to be held with delete for deleting an ember to prevent it from triggering when the user enters delete in the edit box. -This wasn't otherwise fixable without writing a lot more code. --Bug fixes -EmberGenome was crashing when generating a sequence from a source file with more than 2 embers in it. -EmberGenome was improperly handling the first frame of a merge after the last frame of the loop. -These bugs were due to a previous commit. Revert parts of that commit. -Prevent a zoom value of less than 0 when reading from xml. -Slight optimization of the crescents, and mask variations, if the compiler wasn't doing it already. -Unique file naming was broken because it was looking for _# and the default names ended with -#. -Disallow renaming of an ember in the library tree to an empty string. -Severe bug that prevented some variations from being read correctly from params generated outside this program. -Severe OpenCL randomization bug. The first x coordinates of the first points in the first kernel call of the first ember of a render since the OpenCL renderer object was created were not random and were mostly -1. -Severe bug when populating xform selection distributions that could sometimes cause a crash due to roundoff error. Fix by using double. -Limit the max number of variations in a random ember to MAX_CL_VARS, which is 8. This ensures they'll look the same on CPU and GPU. -Prevent user from saving stylesheet to default.qss, it's a special reserved filename. --Code changes -Generalize using the running sum output point inside of a variation for all cases: pre, reg and post. -Allow for array variables in variations where the address of each element is stored in m_Params. -Qualify all math functions with std:: -No longer use our own Clamp() in OpenCL, instead use the standard clamp(). -Redesign how functions are used in the variations OpenCL code. -Add tests to EmberTester to verify some of the new functionality. -Place more const and override qualifiers on functions where appropriate. -Add a global rand with a lock to be used very sparingly. -Use a map instead of a vector for bad param names in Xml parsing. -Prefix affine interpolation mode defines with "AFFINE_" to make their purpose more clear. -Allow for variations that change state during iteration by sending a separate copy of the ember to each rendering thread. -Implement this same functionality with a local struct in OpenCL. It's members are the total of all variables that need to change state within an ember. -Add Contains() function to Utils.h. -EmberRender: print names of kernels being printed with --dump_kernel option. -Clean up EmberTester to handle some of the recent changes. -Fix various casts. -Replace % 2 with & 1, even though the compiler was likely doing this already. -Add new file Variations06.h to accommodate new variations. -General cleanup.
2015-11-22 17:15:07 -05:00
}
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
};
/// <summary>
/// sechq.
/// </summary>
template <typename T>
--Code changes -Split VariationList into .h/cpp files. --Make VariationList.cpp the only place where the variations files are included from. This speeds up the build, but variations can no longer be directly instantiated. --Variations are no longer exported.
2016-02-18 21:58:24 -05:00
class SechqVariation : public Variation<T>
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
public:
--Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums.
2016-01-04 19:50:15 -05:00
SechqVariation(T weight = 1.0) : Variation<T>("sechq", eVariationId::VAR_SECHQ, weight, true) { }
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
VARCOPY(SechqVariation)
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
--Code changes -Make usage of VarFuncs more consistent by moving some global functions there. -Make CircleTrans1 a little safer by calling Zeps() on m_Sc during precalc. -Begin prepping for 0.9.9.5 release.
2016-03-04 21:54:06 -05:00
T absV = VarFuncs<T>::Hypot(helper.In.y, helper.In.z);
06/04/2017 --User changes -Make all fonts be MS Shell Dlg 2. This will require reloading dark.qss if users are already using it. -Limit size of the left side of the palette editor. -Disable create from image buttons in the palette editor when working on a fixed palette. --Bug fixes -The following variations were wrong: coshq, cothq. -During iteration, the color index could become nan if all xform color speeds were negative. This could lead to bad results on the GPU. Fix to check for nan. Minimal speed difference. --Code changes -Make the following variations safer by using Zeps(): sinq, sinhq, secq, sechq, tanq, tanhq, cosq, coshq, cotq, cothq, cscq, cschq, estiq. -Always pass -cl-no-signed-zeros -cl-denorms-are-zero to kernel compiles for both single and double. -Flush all denormals to zero for all executable programs. This will likely lead to a speedup for badly behaving programs.
2017-06-04 20:37:29 -04:00
T ni = m_Weight / Zeps(helper.m_PrecalcSumSquares + SQR(helper.In.z));
--User changes -Add new variations: bubbleT3D, crob, hexaplay3D, hexcrop, hexes, hexnix3D, loonie2, loonie3, nBlur, octapol and synth. -Allow for pre/post versions of dc_bubble, dc_cylinder and dc_linear whereas before they were omitted. -When saving a file with multiple embers in it, detect if time values are all the same and if so, start them at zero and increment by 1 for each ember. -Allow for numerous quality increases to be coalesced into one. It will pick up at the end of the current render. -Show selection highlight on variations tree in response to mouse hover. This makes it easier to see for which variation or param the current mouse wheel action will apply. -Make default temporal samples be 100, whereas before it was 1000 which was overkill. -Require the shift key to be held with delete for deleting an ember to prevent it from triggering when the user enters delete in the edit box. -This wasn't otherwise fixable without writing a lot more code. --Bug fixes -EmberGenome was crashing when generating a sequence from a source file with more than 2 embers in it. -EmberGenome was improperly handling the first frame of a merge after the last frame of the loop. -These bugs were due to a previous commit. Revert parts of that commit. -Prevent a zoom value of less than 0 when reading from xml. -Slight optimization of the crescents, and mask variations, if the compiler wasn't doing it already. -Unique file naming was broken because it was looking for _# and the default names ended with -#. -Disallow renaming of an ember in the library tree to an empty string. -Severe bug that prevented some variations from being read correctly from params generated outside this program. -Severe OpenCL randomization bug. The first x coordinates of the first points in the first kernel call of the first ember of a render since the OpenCL renderer object was created were not random and were mostly -1. -Severe bug when populating xform selection distributions that could sometimes cause a crash due to roundoff error. Fix by using double. -Limit the max number of variations in a random ember to MAX_CL_VARS, which is 8. This ensures they'll look the same on CPU and GPU. -Prevent user from saving stylesheet to default.qss, it's a special reserved filename. --Code changes -Generalize using the running sum output point inside of a variation for all cases: pre, reg and post. -Allow for array variables in variations where the address of each element is stored in m_Params. -Qualify all math functions with std:: -No longer use our own Clamp() in OpenCL, instead use the standard clamp(). -Redesign how functions are used in the variations OpenCL code. -Add tests to EmberTester to verify some of the new functionality. -Place more const and override qualifiers on functions where appropriate. -Add a global rand with a lock to be used very sparingly. -Use a map instead of a vector for bad param names in Xml parsing. -Prefix affine interpolation mode defines with "AFFINE_" to make their purpose more clear. -Allow for variations that change state during iteration by sending a separate copy of the ember to each rendering thread. -Implement this same functionality with a local struct in OpenCL. It's members are the total of all variables that need to change state within an ember. -Add Contains() function to Utils.h. -EmberRender: print names of kernels being printed with --dump_kernel option. -Clean up EmberTester to handle some of the recent changes. -Fix various casts. -Replace % 2 with & 1, even though the compiler was likely doing this already. -Add new file Variations06.h to accommodate new variations. -General cleanup.
2015-11-22 17:15:07 -05:00
T s = std::sin(absV);
T c = std::cos(absV);
T sh = std::sinh(helper.In.x);
T ch = std::cosh(helper.In.x);
06/04/2017 --User changes -Make all fonts be MS Shell Dlg 2. This will require reloading dark.qss if users are already using it. -Limit size of the left side of the palette editor. -Disable create from image buttons in the palette editor when working on a fixed palette. --Bug fixes -The following variations were wrong: coshq, cothq. -During iteration, the color index could become nan if all xform color speeds were negative. This could lead to bad results on the GPU. Fix to check for nan. Minimal speed difference. --Code changes -Make the following variations safer by using Zeps(): sinq, sinhq, secq, sechq, tanq, tanhq, cosq, coshq, cotq, cothq, cscq, cschq, estiq. -Always pass -cl-no-signed-zeros -cl-denorms-are-zero to kernel compiles for both single and double. -Flush all denormals to zero for all executable programs. This will likely lead to a speedup for badly behaving programs.
2017-06-04 20:37:29 -04:00
T d = ni * sh * s / Zeps(absV);
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
helper.Out.x = ch * c * ni;
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
helper.Out.y = -(d * helper.In.y);
helper.Out.z = -(d * helper.In.z);
}
--Code changes -Better use of const functions. -Ember.pro should not define BUILDING_EMBERCL, instead define BUILDING_EMBER.
2015-06-29 23:13:53 -04:00
virtual string OpenCLString() const override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
ostringstream ss;
0.4.1.3 Beta 10/14/2014 --User Changes Size is no longer fixed to the window size. Size scaling is done differently in the final render dialog. This fixes several bugs. Remove Xml saving size from settings and options dialog, it no longer applies. Final render can be broken into strips. Set default save path to the desktop if none is found in the settings file. Set default output size to 1920x1080 if none is found in the settings file. --Bug Fixes Better memory size reporting in final render dialog. --Code Changes Migrate to C++11, Qt 5.3.1, and Visual Studio 2013. Change most instances of unsigned int to size_t, and int to intmax_t. Add m_OrigPixPerUnit and m_ScaleType to Ember for scaling purposes. Replace some sprintf_s() calls in XmlToEmber with ostringstream. Move more non-templated members into RendererBase. Add CopyVec() overload that takes a per element function pointer. Add vector Memset(). Replace '&' with '+' instead of "&amp;" in XmlToEmber for much faster parsing. Break strips rendering out into EmberCommon and call from EmberRender and Fractorium. Make AddAndWriteBuffer() just call WriteBuffer(). Make AddAndWriteImage() delete the existing image first before replacing it. Add SetOutputTexture() to RendererCL to support making new textures in response to resize events. Remove multiple return statements in RendererCL, and replace with a bool that tracks results. Add ToDouble(), MakeEnd(), ToString() and Exists() wrappers in Fractorium. Add Size() wrapper in EmberFile. Make QString function arguments const QString&, and string with const string&. Make ShowCritical() wrapper for invoking a message box from another thread. Add combo box to TwoButtonWidget and rename.
2014-10-14 11:53:15 -04:00
intmax_t varIndex = IndexInXform();
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
string weight = WeightDefineString();
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
ss << "\t{\n"
<< "\t\treal_t absV = Hypot(vIn.y, vIn.z);\n"
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
<< "\t\treal_t ni = " << weight << " / Zeps(fma(vIn.z, vIn.z, precalcSumSquares));\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t\treal_t s = sin(absV);\n"
<< "\t\treal_t c = cos(absV);\n"
<< "\t\treal_t sh = sinh(vIn.x);\n"
<< "\t\treal_t ch = cosh(vIn.x);\n"
<< "\t\treal_t d = ni * sh * s / absV;\n"
<< "\n"
<< "\t\tvOut.x = ch * c * ni;\n"
<< "\t\tvOut.y = -(d * vIn.y);\n"
<< "\t\tvOut.z = -(d * vIn.z);\n"
<< "\t}\n";
return ss.str();
}
--User changes -Add new variations: bubbleT3D, crob, hexaplay3D, hexcrop, hexes, hexnix3D, loonie2, loonie3, nBlur, octapol and synth. -Allow for pre/post versions of dc_bubble, dc_cylinder and dc_linear whereas before they were omitted. -When saving a file with multiple embers in it, detect if time values are all the same and if so, start them at zero and increment by 1 for each ember. -Allow for numerous quality increases to be coalesced into one. It will pick up at the end of the current render. -Show selection highlight on variations tree in response to mouse hover. This makes it easier to see for which variation or param the current mouse wheel action will apply. -Make default temporal samples be 100, whereas before it was 1000 which was overkill. -Require the shift key to be held with delete for deleting an ember to prevent it from triggering when the user enters delete in the edit box. -This wasn't otherwise fixable without writing a lot more code. --Bug fixes -EmberGenome was crashing when generating a sequence from a source file with more than 2 embers in it. -EmberGenome was improperly handling the first frame of a merge after the last frame of the loop. -These bugs were due to a previous commit. Revert parts of that commit. -Prevent a zoom value of less than 0 when reading from xml. -Slight optimization of the crescents, and mask variations, if the compiler wasn't doing it already. -Unique file naming was broken because it was looking for _# and the default names ended with -#. -Disallow renaming of an ember in the library tree to an empty string. -Severe bug that prevented some variations from being read correctly from params generated outside this program. -Severe OpenCL randomization bug. The first x coordinates of the first points in the first kernel call of the first ember of a render since the OpenCL renderer object was created were not random and were mostly -1. -Severe bug when populating xform selection distributions that could sometimes cause a crash due to roundoff error. Fix by using double. -Limit the max number of variations in a random ember to MAX_CL_VARS, which is 8. This ensures they'll look the same on CPU and GPU. -Prevent user from saving stylesheet to default.qss, it's a special reserved filename. --Code changes -Generalize using the running sum output point inside of a variation for all cases: pre, reg and post. -Allow for array variables in variations where the address of each element is stored in m_Params. -Qualify all math functions with std:: -No longer use our own Clamp() in OpenCL, instead use the standard clamp(). -Redesign how functions are used in the variations OpenCL code. -Add tests to EmberTester to verify some of the new functionality. -Place more const and override qualifiers on functions where appropriate. -Add a global rand with a lock to be used very sparingly. -Use a map instead of a vector for bad param names in Xml parsing. -Prefix affine interpolation mode defines with "AFFINE_" to make their purpose more clear. -Allow for variations that change state during iteration by sending a separate copy of the ember to each rendering thread. -Implement this same functionality with a local struct in OpenCL. It's members are the total of all variables that need to change state within an ember. -Add Contains() function to Utils.h. -EmberRender: print names of kernels being printed with --dump_kernel option. -Clean up EmberTester to handle some of the recent changes. -Fix various casts. -Replace % 2 with & 1, even though the compiler was likely doing this already. -Add new file Variations06.h to accommodate new variations. -General cleanup.
2015-11-22 17:15:07 -05:00
virtual vector<string> OpenCLGlobalFuncNames() const override
{
06/04/2017 --User changes -Make all fonts be MS Shell Dlg 2. This will require reloading dark.qss if users are already using it. -Limit size of the left side of the palette editor. -Disable create from image buttons in the palette editor when working on a fixed palette. --Bug fixes -The following variations were wrong: coshq, cothq. -During iteration, the color index could become nan if all xform color speeds were negative. This could lead to bad results on the GPU. Fix to check for nan. Minimal speed difference. --Code changes -Make the following variations safer by using Zeps(): sinq, sinhq, secq, sechq, tanq, tanhq, cosq, coshq, cotq, cothq, cscq, cschq, estiq. -Always pass -cl-no-signed-zeros -cl-denorms-are-zero to kernel compiles for both single and double. -Flush all denormals to zero for all executable programs. This will likely lead to a speedup for badly behaving programs.
2017-06-04 20:37:29 -04:00
return vector<string> { "Hypot", "Zeps" };
--User changes -Add new variations: bubbleT3D, crob, hexaplay3D, hexcrop, hexes, hexnix3D, loonie2, loonie3, nBlur, octapol and synth. -Allow for pre/post versions of dc_bubble, dc_cylinder and dc_linear whereas before they were omitted. -When saving a file with multiple embers in it, detect if time values are all the same and if so, start them at zero and increment by 1 for each ember. -Allow for numerous quality increases to be coalesced into one. It will pick up at the end of the current render. -Show selection highlight on variations tree in response to mouse hover. This makes it easier to see for which variation or param the current mouse wheel action will apply. -Make default temporal samples be 100, whereas before it was 1000 which was overkill. -Require the shift key to be held with delete for deleting an ember to prevent it from triggering when the user enters delete in the edit box. -This wasn't otherwise fixable without writing a lot more code. --Bug fixes -EmberGenome was crashing when generating a sequence from a source file with more than 2 embers in it. -EmberGenome was improperly handling the first frame of a merge after the last frame of the loop. -These bugs were due to a previous commit. Revert parts of that commit. -Prevent a zoom value of less than 0 when reading from xml. -Slight optimization of the crescents, and mask variations, if the compiler wasn't doing it already. -Unique file naming was broken because it was looking for _# and the default names ended with -#. -Disallow renaming of an ember in the library tree to an empty string. -Severe bug that prevented some variations from being read correctly from params generated outside this program. -Severe OpenCL randomization bug. The first x coordinates of the first points in the first kernel call of the first ember of a render since the OpenCL renderer object was created were not random and were mostly -1. -Severe bug when populating xform selection distributions that could sometimes cause a crash due to roundoff error. Fix by using double. -Limit the max number of variations in a random ember to MAX_CL_VARS, which is 8. This ensures they'll look the same on CPU and GPU. -Prevent user from saving stylesheet to default.qss, it's a special reserved filename. --Code changes -Generalize using the running sum output point inside of a variation for all cases: pre, reg and post. -Allow for array variables in variations where the address of each element is stored in m_Params. -Qualify all math functions with std:: -No longer use our own Clamp() in OpenCL, instead use the standard clamp(). -Redesign how functions are used in the variations OpenCL code. -Add tests to EmberTester to verify some of the new functionality. -Place more const and override qualifiers on functions where appropriate. -Add a global rand with a lock to be used very sparingly. -Use a map instead of a vector for bad param names in Xml parsing. -Prefix affine interpolation mode defines with "AFFINE_" to make their purpose more clear. -Allow for variations that change state during iteration by sending a separate copy of the ember to each rendering thread. -Implement this same functionality with a local struct in OpenCL. It's members are the total of all variables that need to change state within an ember. -Add Contains() function to Utils.h. -EmberRender: print names of kernels being printed with --dump_kernel option. -Clean up EmberTester to handle some of the recent changes. -Fix various casts. -Replace % 2 with & 1, even though the compiler was likely doing this already. -Add new file Variations06.h to accommodate new variations. -General cleanup.
2015-11-22 17:15:07 -05:00
}
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
};
/// <summary>
/// tanq.
/// </summary>
template <typename T>
--Code changes -Split VariationList into .h/cpp files. --Make VariationList.cpp the only place where the variations files are included from. This speeds up the build, but variations can no longer be directly instantiated. --Variations are no longer exported.
2016-02-18 21:58:24 -05:00
class TanqVariation : public Variation<T>
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
public:
--Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums.
2016-01-04 19:50:15 -05:00
TanqVariation(T weight = 1.0) : Variation<T>("tanq", eVariationId::VAR_TANQ, weight) { }
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
VARCOPY(TanqVariation)
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
T sysz = SQR(helper.In.y) + SQR(helper.In.z);
Explicitly namespace calls to sqrt and isnan osx/linux have alternative top-level implementations of these methods, so the calls are ambiguous given the - so these should be used explicitly
2015-04-13 07:32:58 -04:00
T absV = std::sqrt(sysz);
06/04/2017 --User changes -Make all fonts be MS Shell Dlg 2. This will require reloading dark.qss if users are already using it. -Limit size of the left side of the palette editor. -Disable create from image buttons in the palette editor when working on a fixed palette. --Bug fixes -The following variations were wrong: coshq, cothq. -During iteration, the color index could become nan if all xform color speeds were negative. This could lead to bad results on the GPU. Fix to check for nan. Minimal speed difference. --Code changes -Make the following variations safer by using Zeps(): sinq, sinhq, secq, sechq, tanq, tanhq, cosq, coshq, cotq, cothq, cscq, cschq, estiq. -Always pass -cl-no-signed-zeros -cl-denorms-are-zero to kernel compiles for both single and double. -Flush all denormals to zero for all executable programs. This will likely lead to a speedup for badly behaving programs.
2017-06-04 20:37:29 -04:00
T ni = m_Weight / Zeps(SQR(helper.In.x) + sysz);
--User changes -Add new variations: bubbleT3D, crob, hexaplay3D, hexcrop, hexes, hexnix3D, loonie2, loonie3, nBlur, octapol and synth. -Allow for pre/post versions of dc_bubble, dc_cylinder and dc_linear whereas before they were omitted. -When saving a file with multiple embers in it, detect if time values are all the same and if so, start them at zero and increment by 1 for each ember. -Allow for numerous quality increases to be coalesced into one. It will pick up at the end of the current render. -Show selection highlight on variations tree in response to mouse hover. This makes it easier to see for which variation or param the current mouse wheel action will apply. -Make default temporal samples be 100, whereas before it was 1000 which was overkill. -Require the shift key to be held with delete for deleting an ember to prevent it from triggering when the user enters delete in the edit box. -This wasn't otherwise fixable without writing a lot more code. --Bug fixes -EmberGenome was crashing when generating a sequence from a source file with more than 2 embers in it. -EmberGenome was improperly handling the first frame of a merge after the last frame of the loop. -These bugs were due to a previous commit. Revert parts of that commit. -Prevent a zoom value of less than 0 when reading from xml. -Slight optimization of the crescents, and mask variations, if the compiler wasn't doing it already. -Unique file naming was broken because it was looking for _# and the default names ended with -#. -Disallow renaming of an ember in the library tree to an empty string. -Severe bug that prevented some variations from being read correctly from params generated outside this program. -Severe OpenCL randomization bug. The first x coordinates of the first points in the first kernel call of the first ember of a render since the OpenCL renderer object was created were not random and were mostly -1. -Severe bug when populating xform selection distributions that could sometimes cause a crash due to roundoff error. Fix by using double. -Limit the max number of variations in a random ember to MAX_CL_VARS, which is 8. This ensures they'll look the same on CPU and GPU. -Prevent user from saving stylesheet to default.qss, it's a special reserved filename. --Code changes -Generalize using the running sum output point inside of a variation for all cases: pre, reg and post. -Allow for array variables in variations where the address of each element is stored in m_Params. -Qualify all math functions with std:: -No longer use our own Clamp() in OpenCL, instead use the standard clamp(). -Redesign how functions are used in the variations OpenCL code. -Add tests to EmberTester to verify some of the new functionality. -Place more const and override qualifiers on functions where appropriate. -Add a global rand with a lock to be used very sparingly. -Use a map instead of a vector for bad param names in Xml parsing. -Prefix affine interpolation mode defines with "AFFINE_" to make their purpose more clear. -Allow for variations that change state during iteration by sending a separate copy of the ember to each rendering thread. -Implement this same functionality with a local struct in OpenCL. It's members are the total of all variables that need to change state within an ember. -Add Contains() function to Utils.h. -EmberRender: print names of kernels being printed with --dump_kernel option. -Clean up EmberTester to handle some of the recent changes. -Fix various casts. -Replace % 2 with & 1, even though the compiler was likely doing this already. -Add new file Variations06.h to accommodate new variations. -General cleanup.
2015-11-22 17:15:07 -05:00
T s = std::sin(helper.In.x);
T c = std::cos(helper.In.x);
T sh = std::sinh(absV);
T ch = std::cosh(absV);
06/04/2017 --User changes -Make all fonts be MS Shell Dlg 2. This will require reloading dark.qss if users are already using it. -Limit size of the left side of the palette editor. -Disable create from image buttons in the palette editor when working on a fixed palette. --Bug fixes -The following variations were wrong: coshq, cothq. -During iteration, the color index could become nan if all xform color speeds were negative. This could lead to bad results on the GPU. Fix to check for nan. Minimal speed difference. --Code changes -Make the following variations safer by using Zeps(): sinq, sinhq, secq, sechq, tanq, tanhq, cosq, coshq, cotq, cothq, cscq, cschq, estiq. -Always pass -cl-no-signed-zeros -cl-denorms-are-zero to kernel compiles for both single and double. -Flush all denormals to zero for all executable programs. This will likely lead to a speedup for badly behaving programs.
2017-06-04 20:37:29 -04:00
T d = c * sh / Zeps(absV);
T b = -s * sh / Zeps(absV);
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
T stcv = s * ch;
T nstcv = -stcv;
T ctcv = c * ch;
helper.Out.x = (stcv * ctcv + d * b * sysz) * ni;
helper.Out.y = (nstcv * b * helper.In.y + d * helper.In.y * ctcv) * ni;
helper.Out.z = (nstcv * b * helper.In.z + d * helper.In.z * ctcv) * ni;
}
--Code changes -Better use of const functions. -Ember.pro should not define BUILDING_EMBERCL, instead define BUILDING_EMBER.
2015-06-29 23:13:53 -04:00
virtual string OpenCLString() const override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
ostringstream ss;
0.4.1.3 Beta 10/14/2014 --User Changes Size is no longer fixed to the window size. Size scaling is done differently in the final render dialog. This fixes several bugs. Remove Xml saving size from settings and options dialog, it no longer applies. Final render can be broken into strips. Set default save path to the desktop if none is found in the settings file. Set default output size to 1920x1080 if none is found in the settings file. --Bug Fixes Better memory size reporting in final render dialog. --Code Changes Migrate to C++11, Qt 5.3.1, and Visual Studio 2013. Change most instances of unsigned int to size_t, and int to intmax_t. Add m_OrigPixPerUnit and m_ScaleType to Ember for scaling purposes. Replace some sprintf_s() calls in XmlToEmber with ostringstream. Move more non-templated members into RendererBase. Add CopyVec() overload that takes a per element function pointer. Add vector Memset(). Replace '&' with '+' instead of "&amp;" in XmlToEmber for much faster parsing. Break strips rendering out into EmberCommon and call from EmberRender and Fractorium. Make AddAndWriteBuffer() just call WriteBuffer(). Make AddAndWriteImage() delete the existing image first before replacing it. Add SetOutputTexture() to RendererCL to support making new textures in response to resize events. Remove multiple return statements in RendererCL, and replace with a bool that tracks results. Add ToDouble(), MakeEnd(), ToString() and Exists() wrappers in Fractorium. Add Size() wrapper in EmberFile. Make QString function arguments const QString&, and string with const string&. Make ShowCritical() wrapper for invoking a message box from another thread. Add combo box to TwoButtonWidget and rename.
2014-10-14 11:53:15 -04:00
intmax_t varIndex = IndexInXform();
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
string weight = WeightDefineString();
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
ss << "\t{\n"
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
<< "\t\treal_t sysz = fma(vIn.y, vIn.y, SQR(vIn.z));\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t\treal_t absV = sqrt(sysz);\n"
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
<< "\t\treal_t ni = " << weight << " / Zeps(fma(vIn.x, vIn.x, sysz));\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t\treal_t s = sin(vIn.x);\n"
<< "\t\treal_t c = cos(vIn.x);\n"
<< "\t\treal_t sh = sinh(absV);\n"
<< "\t\treal_t ch = cosh(absV);\n"
06/04/2017 --User changes -Make all fonts be MS Shell Dlg 2. This will require reloading dark.qss if users are already using it. -Limit size of the left side of the palette editor. -Disable create from image buttons in the palette editor when working on a fixed palette. --Bug fixes -The following variations were wrong: coshq, cothq. -During iteration, the color index could become nan if all xform color speeds were negative. This could lead to bad results on the GPU. Fix to check for nan. Minimal speed difference. --Code changes -Make the following variations safer by using Zeps(): sinq, sinhq, secq, sechq, tanq, tanhq, cosq, coshq, cotq, cothq, cscq, cschq, estiq. -Always pass -cl-no-signed-zeros -cl-denorms-are-zero to kernel compiles for both single and double. -Flush all denormals to zero for all executable programs. This will likely lead to a speedup for badly behaving programs.
2017-06-04 20:37:29 -04:00
<< "\t\treal_t d = c * sh / Zeps(absV);\n"
<< "\t\treal_t b = -s * sh / Zeps(absV);\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t\treal_t stcv = s * ch;\n"
<< "\t\treal_t nstcv = -stcv;\n"
<< "\t\treal_t ctcv = c * ch;\n"
<< "\n"
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
<< "\t\tvOut.x = fma(stcv, ctcv, d * b * sysz) * ni;\n"
<< "\t\tvOut.y = fma(nstcv, b * vIn.y, d * vIn.y * ctcv) * ni;\n"
<< "\t\tvOut.z = fma(nstcv, b * vIn.z, d * vIn.z * ctcv) * ni;\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t}\n";
return ss.str();
}
06/04/2017 --User changes -Make all fonts be MS Shell Dlg 2. This will require reloading dark.qss if users are already using it. -Limit size of the left side of the palette editor. -Disable create from image buttons in the palette editor when working on a fixed palette. --Bug fixes -The following variations were wrong: coshq, cothq. -During iteration, the color index could become nan if all xform color speeds were negative. This could lead to bad results on the GPU. Fix to check for nan. Minimal speed difference. --Code changes -Make the following variations safer by using Zeps(): sinq, sinhq, secq, sechq, tanq, tanhq, cosq, coshq, cotq, cothq, cscq, cschq, estiq. -Always pass -cl-no-signed-zeros -cl-denorms-are-zero to kernel compiles for both single and double. -Flush all denormals to zero for all executable programs. This will likely lead to a speedup for badly behaving programs.
2017-06-04 20:37:29 -04:00
virtual vector<string> OpenCLGlobalFuncNames() const override
{
return vector<string> { "Zeps" };
}
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
};
/// <summary>
/// tanhq.
/// </summary>
template <typename T>
--Code changes -Split VariationList into .h/cpp files. --Make VariationList.cpp the only place where the variations files are included from. This speeds up the build, but variations can no longer be directly instantiated. --Variations are no longer exported.
2016-02-18 21:58:24 -05:00
class TanhqVariation : public Variation<T>
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
public:
--Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums.
2016-01-04 19:50:15 -05:00
TanhqVariation(T weight = 1.0) : Variation<T>("tanhq", eVariationId::VAR_TANHQ, weight) { }
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
VARCOPY(TanhqVariation)
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
T sysz = SQR(helper.In.y) + SQR(helper.In.z);
Explicitly namespace calls to sqrt and isnan osx/linux have alternative top-level implementations of these methods, so the calls are ambiguous given the - so these should be used explicitly
2015-04-13 07:32:58 -04:00
T absV = std::sqrt(sysz);
06/04/2017 --User changes -Make all fonts be MS Shell Dlg 2. This will require reloading dark.qss if users are already using it. -Limit size of the left side of the palette editor. -Disable create from image buttons in the palette editor when working on a fixed palette. --Bug fixes -The following variations were wrong: coshq, cothq. -During iteration, the color index could become nan if all xform color speeds were negative. This could lead to bad results on the GPU. Fix to check for nan. Minimal speed difference. --Code changes -Make the following variations safer by using Zeps(): sinq, sinhq, secq, sechq, tanq, tanhq, cosq, coshq, cotq, cothq, cscq, cschq, estiq. -Always pass -cl-no-signed-zeros -cl-denorms-are-zero to kernel compiles for both single and double. -Flush all denormals to zero for all executable programs. This will likely lead to a speedup for badly behaving programs.
2017-06-04 20:37:29 -04:00
T ni = m_Weight / Zeps(SQR(helper.In.x) + sysz);
--User changes -Add new variations: bubbleT3D, crob, hexaplay3D, hexcrop, hexes, hexnix3D, loonie2, loonie3, nBlur, octapol and synth. -Allow for pre/post versions of dc_bubble, dc_cylinder and dc_linear whereas before they were omitted. -When saving a file with multiple embers in it, detect if time values are all the same and if so, start them at zero and increment by 1 for each ember. -Allow for numerous quality increases to be coalesced into one. It will pick up at the end of the current render. -Show selection highlight on variations tree in response to mouse hover. This makes it easier to see for which variation or param the current mouse wheel action will apply. -Make default temporal samples be 100, whereas before it was 1000 which was overkill. -Require the shift key to be held with delete for deleting an ember to prevent it from triggering when the user enters delete in the edit box. -This wasn't otherwise fixable without writing a lot more code. --Bug fixes -EmberGenome was crashing when generating a sequence from a source file with more than 2 embers in it. -EmberGenome was improperly handling the first frame of a merge after the last frame of the loop. -These bugs were due to a previous commit. Revert parts of that commit. -Prevent a zoom value of less than 0 when reading from xml. -Slight optimization of the crescents, and mask variations, if the compiler wasn't doing it already. -Unique file naming was broken because it was looking for _# and the default names ended with -#. -Disallow renaming of an ember in the library tree to an empty string. -Severe bug that prevented some variations from being read correctly from params generated outside this program. -Severe OpenCL randomization bug. The first x coordinates of the first points in the first kernel call of the first ember of a render since the OpenCL renderer object was created were not random and were mostly -1. -Severe bug when populating xform selection distributions that could sometimes cause a crash due to roundoff error. Fix by using double. -Limit the max number of variations in a random ember to MAX_CL_VARS, which is 8. This ensures they'll look the same on CPU and GPU. -Prevent user from saving stylesheet to default.qss, it's a special reserved filename. --Code changes -Generalize using the running sum output point inside of a variation for all cases: pre, reg and post. -Allow for array variables in variations where the address of each element is stored in m_Params. -Qualify all math functions with std:: -No longer use our own Clamp() in OpenCL, instead use the standard clamp(). -Redesign how functions are used in the variations OpenCL code. -Add tests to EmberTester to verify some of the new functionality. -Place more const and override qualifiers on functions where appropriate. -Add a global rand with a lock to be used very sparingly. -Use a map instead of a vector for bad param names in Xml parsing. -Prefix affine interpolation mode defines with "AFFINE_" to make their purpose more clear. -Allow for variations that change state during iteration by sending a separate copy of the ember to each rendering thread. -Implement this same functionality with a local struct in OpenCL. It's members are the total of all variables that need to change state within an ember. -Add Contains() function to Utils.h. -EmberRender: print names of kernels being printed with --dump_kernel option. -Clean up EmberTester to handle some of the recent changes. -Fix various casts. -Replace % 2 with & 1, even though the compiler was likely doing this already. -Add new file Variations06.h to accommodate new variations. -General cleanup.
2015-11-22 17:15:07 -05:00
T s = std::sin(absV);
T c = std::cos(absV);
T sh = std::sinh(helper.In.x);
T ch = std::cosh(helper.In.x);
06/04/2017 --User changes -Make all fonts be MS Shell Dlg 2. This will require reloading dark.qss if users are already using it. -Limit size of the left side of the palette editor. -Disable create from image buttons in the palette editor when working on a fixed palette. --Bug fixes -The following variations were wrong: coshq, cothq. -During iteration, the color index could become nan if all xform color speeds were negative. This could lead to bad results on the GPU. Fix to check for nan. Minimal speed difference. --Code changes -Make the following variations safer by using Zeps(): sinq, sinhq, secq, sechq, tanq, tanhq, cosq, coshq, cotq, cothq, cscq, cschq, estiq. -Always pass -cl-no-signed-zeros -cl-denorms-are-zero to kernel compiles for both single and double. -Flush all denormals to zero for all executable programs. This will likely lead to a speedup for badly behaving programs.
2017-06-04 20:37:29 -04:00
T d = ch * s / Zeps(absV);
T b = sh * s / Zeps(absV);
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
T stcv = sh * c;
T nstcv = -stcv;
T ctcv = c * ch;
helper.Out.x = (stcv * ctcv + d * b * sysz) * ni;
helper.Out.y = (nstcv * b * helper.In.y + d * helper.In.y * ctcv) * ni;
helper.Out.z = (nstcv * b * helper.In.z + d * helper.In.z * ctcv) * ni;
}
--Code changes -Better use of const functions. -Ember.pro should not define BUILDING_EMBERCL, instead define BUILDING_EMBER.
2015-06-29 23:13:53 -04:00
virtual string OpenCLString() const override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
ostringstream ss;
0.4.1.3 Beta 10/14/2014 --User Changes Size is no longer fixed to the window size. Size scaling is done differently in the final render dialog. This fixes several bugs. Remove Xml saving size from settings and options dialog, it no longer applies. Final render can be broken into strips. Set default save path to the desktop if none is found in the settings file. Set default output size to 1920x1080 if none is found in the settings file. --Bug Fixes Better memory size reporting in final render dialog. --Code Changes Migrate to C++11, Qt 5.3.1, and Visual Studio 2013. Change most instances of unsigned int to size_t, and int to intmax_t. Add m_OrigPixPerUnit and m_ScaleType to Ember for scaling purposes. Replace some sprintf_s() calls in XmlToEmber with ostringstream. Move more non-templated members into RendererBase. Add CopyVec() overload that takes a per element function pointer. Add vector Memset(). Replace '&' with '+' instead of "&amp;" in XmlToEmber for much faster parsing. Break strips rendering out into EmberCommon and call from EmberRender and Fractorium. Make AddAndWriteBuffer() just call WriteBuffer(). Make AddAndWriteImage() delete the existing image first before replacing it. Add SetOutputTexture() to RendererCL to support making new textures in response to resize events. Remove multiple return statements in RendererCL, and replace with a bool that tracks results. Add ToDouble(), MakeEnd(), ToString() and Exists() wrappers in Fractorium. Add Size() wrapper in EmberFile. Make QString function arguments const QString&, and string with const string&. Make ShowCritical() wrapper for invoking a message box from another thread. Add combo box to TwoButtonWidget and rename.
2014-10-14 11:53:15 -04:00
intmax_t varIndex = IndexInXform();
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
string weight = WeightDefineString();
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
ss << "\t{\n"
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
<< "\t\treal_t sysz = fma(vIn.y, vIn.y, SQR(vIn.z));\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t\treal_t absV = sqrt(sysz);\n"
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
<< "\t\treal_t ni = " << weight << " / Zeps(fma(vIn.x, vIn.x, sysz));\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t\treal_t s = sin(absV);\n"
<< "\t\treal_t c = cos(absV);\n"
<< "\t\treal_t sh = sinh(vIn.x);\n"
<< "\t\treal_t ch = cosh(vIn.x);\n"
06/04/2017 --User changes -Make all fonts be MS Shell Dlg 2. This will require reloading dark.qss if users are already using it. -Limit size of the left side of the palette editor. -Disable create from image buttons in the palette editor when working on a fixed palette. --Bug fixes -The following variations were wrong: coshq, cothq. -During iteration, the color index could become nan if all xform color speeds were negative. This could lead to bad results on the GPU. Fix to check for nan. Minimal speed difference. --Code changes -Make the following variations safer by using Zeps(): sinq, sinhq, secq, sechq, tanq, tanhq, cosq, coshq, cotq, cothq, cscq, cschq, estiq. -Always pass -cl-no-signed-zeros -cl-denorms-are-zero to kernel compiles for both single and double. -Flush all denormals to zero for all executable programs. This will likely lead to a speedup for badly behaving programs.
2017-06-04 20:37:29 -04:00
<< "\t\treal_t d = ch * s / Zeps(absV);\n"
<< "\t\treal_t b = sh * s / Zeps(absV);\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t\treal_t stcv = sh * c;\n"
<< "\t\treal_t nstcv = -stcv;\n"
<< "\t\treal_t ctcv = c * ch;\n"
<< "\n"
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
<< "\t\tvOut.x = fma(stcv, ctcv, d * b * sysz) * ni;\n"
<< "\t\tvOut.y = fma(nstcv, b * vIn.y, d * vIn.y * ctcv) * ni;\n"
<< "\t\tvOut.z = fma(nstcv, b * vIn.z, d * vIn.z * ctcv) * ni;\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t}\n";
return ss.str();
}
06/04/2017 --User changes -Make all fonts be MS Shell Dlg 2. This will require reloading dark.qss if users are already using it. -Limit size of the left side of the palette editor. -Disable create from image buttons in the palette editor when working on a fixed palette. --Bug fixes -The following variations were wrong: coshq, cothq. -During iteration, the color index could become nan if all xform color speeds were negative. This could lead to bad results on the GPU. Fix to check for nan. Minimal speed difference. --Code changes -Make the following variations safer by using Zeps(): sinq, sinhq, secq, sechq, tanq, tanhq, cosq, coshq, cotq, cothq, cscq, cschq, estiq. -Always pass -cl-no-signed-zeros -cl-denorms-are-zero to kernel compiles for both single and double. -Flush all denormals to zero for all executable programs. This will likely lead to a speedup for badly behaving programs.
2017-06-04 20:37:29 -04:00
virtual vector<string> OpenCLGlobalFuncNames() const override
{
return vector<string> { "Zeps" };
}
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
};
/// <summary>
/// cosq.
/// </summary>
template <typename T>
--Code changes -Split VariationList into .h/cpp files. --Make VariationList.cpp the only place where the variations files are included from. This speeds up the build, but variations can no longer be directly instantiated. --Variations are no longer exported.
2016-02-18 21:58:24 -05:00
class CosqVariation : public Variation<T>
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
public:
--Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums.
2016-01-04 19:50:15 -05:00
CosqVariation(T weight = 1.0) : Variation<T>("cosq", eVariationId::VAR_COSQ, weight) { }
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
VARCOPY(CosqVariation)
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
--Code changes -Make usage of VarFuncs more consistent by moving some global functions there. -Make CircleTrans1 a little safer by calling Zeps() on m_Sc during precalc. -Begin prepping for 0.9.9.5 release.
2016-03-04 21:54:06 -05:00
T absV = VarFuncs<T>::Hypot(helper.In.y, helper.In.z);
--User changes -Add new variations: bubbleT3D, crob, hexaplay3D, hexcrop, hexes, hexnix3D, loonie2, loonie3, nBlur, octapol and synth. -Allow for pre/post versions of dc_bubble, dc_cylinder and dc_linear whereas before they were omitted. -When saving a file with multiple embers in it, detect if time values are all the same and if so, start them at zero and increment by 1 for each ember. -Allow for numerous quality increases to be coalesced into one. It will pick up at the end of the current render. -Show selection highlight on variations tree in response to mouse hover. This makes it easier to see for which variation or param the current mouse wheel action will apply. -Make default temporal samples be 100, whereas before it was 1000 which was overkill. -Require the shift key to be held with delete for deleting an ember to prevent it from triggering when the user enters delete in the edit box. -This wasn't otherwise fixable without writing a lot more code. --Bug fixes -EmberGenome was crashing when generating a sequence from a source file with more than 2 embers in it. -EmberGenome was improperly handling the first frame of a merge after the last frame of the loop. -These bugs were due to a previous commit. Revert parts of that commit. -Prevent a zoom value of less than 0 when reading from xml. -Slight optimization of the crescents, and mask variations, if the compiler wasn't doing it already. -Unique file naming was broken because it was looking for _# and the default names ended with -#. -Disallow renaming of an ember in the library tree to an empty string. -Severe bug that prevented some variations from being read correctly from params generated outside this program. -Severe OpenCL randomization bug. The first x coordinates of the first points in the first kernel call of the first ember of a render since the OpenCL renderer object was created were not random and were mostly -1. -Severe bug when populating xform selection distributions that could sometimes cause a crash due to roundoff error. Fix by using double. -Limit the max number of variations in a random ember to MAX_CL_VARS, which is 8. This ensures they'll look the same on CPU and GPU. -Prevent user from saving stylesheet to default.qss, it's a special reserved filename. --Code changes -Generalize using the running sum output point inside of a variation for all cases: pre, reg and post. -Allow for array variables in variations where the address of each element is stored in m_Params. -Qualify all math functions with std:: -No longer use our own Clamp() in OpenCL, instead use the standard clamp(). -Redesign how functions are used in the variations OpenCL code. -Add tests to EmberTester to verify some of the new functionality. -Place more const and override qualifiers on functions where appropriate. -Add a global rand with a lock to be used very sparingly. -Use a map instead of a vector for bad param names in Xml parsing. -Prefix affine interpolation mode defines with "AFFINE_" to make their purpose more clear. -Allow for variations that change state during iteration by sending a separate copy of the ember to each rendering thread. -Implement this same functionality with a local struct in OpenCL. It's members are the total of all variables that need to change state within an ember. -Add Contains() function to Utils.h. -EmberRender: print names of kernels being printed with --dump_kernel option. -Clean up EmberTester to handle some of the recent changes. -Fix various casts. -Replace % 2 with & 1, even though the compiler was likely doing this already. -Add new file Variations06.h to accommodate new variations. -General cleanup.
2015-11-22 17:15:07 -05:00
T s = std::sin(helper.In.x);
T c = std::cos(helper.In.x);
T sh = std::sinh(absV);
T ch = std::cosh(absV);
06/04/2017 --User changes -Make all fonts be MS Shell Dlg 2. This will require reloading dark.qss if users are already using it. -Limit size of the left side of the palette editor. -Disable create from image buttons in the palette editor when working on a fixed palette. --Bug fixes -The following variations were wrong: coshq, cothq. -During iteration, the color index could become nan if all xform color speeds were negative. This could lead to bad results on the GPU. Fix to check for nan. Minimal speed difference. --Code changes -Make the following variations safer by using Zeps(): sinq, sinhq, secq, sechq, tanq, tanhq, cosq, coshq, cotq, cothq, cscq, cschq, estiq. -Always pass -cl-no-signed-zeros -cl-denorms-are-zero to kernel compiles for both single and double. -Flush all denormals to zero for all executable programs. This will likely lead to a speedup for badly behaving programs.
2017-06-04 20:37:29 -04:00
T d = -m_Weight * s * sh / Zeps(absV);
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
helper.Out.x = m_Weight * c * ch;
helper.Out.y = d * helper.In.y;
helper.Out.z = d * helper.In.z;
}
--Code changes -Better use of const functions. -Ember.pro should not define BUILDING_EMBERCL, instead define BUILDING_EMBER.
2015-06-29 23:13:53 -04:00
virtual string OpenCLString() const override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
ostringstream ss;
0.4.1.3 Beta 10/14/2014 --User Changes Size is no longer fixed to the window size. Size scaling is done differently in the final render dialog. This fixes several bugs. Remove Xml saving size from settings and options dialog, it no longer applies. Final render can be broken into strips. Set default save path to the desktop if none is found in the settings file. Set default output size to 1920x1080 if none is found in the settings file. --Bug Fixes Better memory size reporting in final render dialog. --Code Changes Migrate to C++11, Qt 5.3.1, and Visual Studio 2013. Change most instances of unsigned int to size_t, and int to intmax_t. Add m_OrigPixPerUnit and m_ScaleType to Ember for scaling purposes. Replace some sprintf_s() calls in XmlToEmber with ostringstream. Move more non-templated members into RendererBase. Add CopyVec() overload that takes a per element function pointer. Add vector Memset(). Replace '&' with '+' instead of "&amp;" in XmlToEmber for much faster parsing. Break strips rendering out into EmberCommon and call from EmberRender and Fractorium. Make AddAndWriteBuffer() just call WriteBuffer(). Make AddAndWriteImage() delete the existing image first before replacing it. Add SetOutputTexture() to RendererCL to support making new textures in response to resize events. Remove multiple return statements in RendererCL, and replace with a bool that tracks results. Add ToDouble(), MakeEnd(), ToString() and Exists() wrappers in Fractorium. Add Size() wrapper in EmberFile. Make QString function arguments const QString&, and string with const string&. Make ShowCritical() wrapper for invoking a message box from another thread. Add combo box to TwoButtonWidget and rename.
2014-10-14 11:53:15 -04:00
intmax_t varIndex = IndexInXform();
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
string weight = WeightDefineString();
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
ss << "\t{\n"
<< "\t\treal_t absV = Hypot(vIn.y, vIn.z);\n"
<< "\t\treal_t s = sin(vIn.x);\n"
<< "\t\treal_t c = cos(vIn.x);\n"
<< "\t\treal_t sh = sinh(absV);\n"
<< "\t\treal_t ch = cosh(absV);\n"
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
<< "\t\treal_t d = -" << weight << " * s * sh / Zeps(absV);\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\n"
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
<< "\t\tvOut.x = " << weight << " * c * ch;\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t\tvOut.y = d * vIn.y;\n"
<< "\t\tvOut.z = d * vIn.z;\n"
<< "\t}\n";
return ss.str();
}
--User changes -Add new variations: bubbleT3D, crob, hexaplay3D, hexcrop, hexes, hexnix3D, loonie2, loonie3, nBlur, octapol and synth. -Allow for pre/post versions of dc_bubble, dc_cylinder and dc_linear whereas before they were omitted. -When saving a file with multiple embers in it, detect if time values are all the same and if so, start them at zero and increment by 1 for each ember. -Allow for numerous quality increases to be coalesced into one. It will pick up at the end of the current render. -Show selection highlight on variations tree in response to mouse hover. This makes it easier to see for which variation or param the current mouse wheel action will apply. -Make default temporal samples be 100, whereas before it was 1000 which was overkill. -Require the shift key to be held with delete for deleting an ember to prevent it from triggering when the user enters delete in the edit box. -This wasn't otherwise fixable without writing a lot more code. --Bug fixes -EmberGenome was crashing when generating a sequence from a source file with more than 2 embers in it. -EmberGenome was improperly handling the first frame of a merge after the last frame of the loop. -These bugs were due to a previous commit. Revert parts of that commit. -Prevent a zoom value of less than 0 when reading from xml. -Slight optimization of the crescents, and mask variations, if the compiler wasn't doing it already. -Unique file naming was broken because it was looking for _# and the default names ended with -#. -Disallow renaming of an ember in the library tree to an empty string. -Severe bug that prevented some variations from being read correctly from params generated outside this program. -Severe OpenCL randomization bug. The first x coordinates of the first points in the first kernel call of the first ember of a render since the OpenCL renderer object was created were not random and were mostly -1. -Severe bug when populating xform selection distributions that could sometimes cause a crash due to roundoff error. Fix by using double. -Limit the max number of variations in a random ember to MAX_CL_VARS, which is 8. This ensures they'll look the same on CPU and GPU. -Prevent user from saving stylesheet to default.qss, it's a special reserved filename. --Code changes -Generalize using the running sum output point inside of a variation for all cases: pre, reg and post. -Allow for array variables in variations where the address of each element is stored in m_Params. -Qualify all math functions with std:: -No longer use our own Clamp() in OpenCL, instead use the standard clamp(). -Redesign how functions are used in the variations OpenCL code. -Add tests to EmberTester to verify some of the new functionality. -Place more const and override qualifiers on functions where appropriate. -Add a global rand with a lock to be used very sparingly. -Use a map instead of a vector for bad param names in Xml parsing. -Prefix affine interpolation mode defines with "AFFINE_" to make their purpose more clear. -Allow for variations that change state during iteration by sending a separate copy of the ember to each rendering thread. -Implement this same functionality with a local struct in OpenCL. It's members are the total of all variables that need to change state within an ember. -Add Contains() function to Utils.h. -EmberRender: print names of kernels being printed with --dump_kernel option. -Clean up EmberTester to handle some of the recent changes. -Fix various casts. -Replace % 2 with & 1, even though the compiler was likely doing this already. -Add new file Variations06.h to accommodate new variations. -General cleanup.
2015-11-22 17:15:07 -05:00
virtual vector<string> OpenCLGlobalFuncNames() const override
{
06/04/2017 --User changes -Make all fonts be MS Shell Dlg 2. This will require reloading dark.qss if users are already using it. -Limit size of the left side of the palette editor. -Disable create from image buttons in the palette editor when working on a fixed palette. --Bug fixes -The following variations were wrong: coshq, cothq. -During iteration, the color index could become nan if all xform color speeds were negative. This could lead to bad results on the GPU. Fix to check for nan. Minimal speed difference. --Code changes -Make the following variations safer by using Zeps(): sinq, sinhq, secq, sechq, tanq, tanhq, cosq, coshq, cotq, cothq, cscq, cschq, estiq. -Always pass -cl-no-signed-zeros -cl-denorms-are-zero to kernel compiles for both single and double. -Flush all denormals to zero for all executable programs. This will likely lead to a speedup for badly behaving programs.
2017-06-04 20:37:29 -04:00
return vector<string> { "Hypot", "Zeps" };
--User changes -Add new variations: bubbleT3D, crob, hexaplay3D, hexcrop, hexes, hexnix3D, loonie2, loonie3, nBlur, octapol and synth. -Allow for pre/post versions of dc_bubble, dc_cylinder and dc_linear whereas before they were omitted. -When saving a file with multiple embers in it, detect if time values are all the same and if so, start them at zero and increment by 1 for each ember. -Allow for numerous quality increases to be coalesced into one. It will pick up at the end of the current render. -Show selection highlight on variations tree in response to mouse hover. This makes it easier to see for which variation or param the current mouse wheel action will apply. -Make default temporal samples be 100, whereas before it was 1000 which was overkill. -Require the shift key to be held with delete for deleting an ember to prevent it from triggering when the user enters delete in the edit box. -This wasn't otherwise fixable without writing a lot more code. --Bug fixes -EmberGenome was crashing when generating a sequence from a source file with more than 2 embers in it. -EmberGenome was improperly handling the first frame of a merge after the last frame of the loop. -These bugs were due to a previous commit. Revert parts of that commit. -Prevent a zoom value of less than 0 when reading from xml. -Slight optimization of the crescents, and mask variations, if the compiler wasn't doing it already. -Unique file naming was broken because it was looking for _# and the default names ended with -#. -Disallow renaming of an ember in the library tree to an empty string. -Severe bug that prevented some variations from being read correctly from params generated outside this program. -Severe OpenCL randomization bug. The first x coordinates of the first points in the first kernel call of the first ember of a render since the OpenCL renderer object was created were not random and were mostly -1. -Severe bug when populating xform selection distributions that could sometimes cause a crash due to roundoff error. Fix by using double. -Limit the max number of variations in a random ember to MAX_CL_VARS, which is 8. This ensures they'll look the same on CPU and GPU. -Prevent user from saving stylesheet to default.qss, it's a special reserved filename. --Code changes -Generalize using the running sum output point inside of a variation for all cases: pre, reg and post. -Allow for array variables in variations where the address of each element is stored in m_Params. -Qualify all math functions with std:: -No longer use our own Clamp() in OpenCL, instead use the standard clamp(). -Redesign how functions are used in the variations OpenCL code. -Add tests to EmberTester to verify some of the new functionality. -Place more const and override qualifiers on functions where appropriate. -Add a global rand with a lock to be used very sparingly. -Use a map instead of a vector for bad param names in Xml parsing. -Prefix affine interpolation mode defines with "AFFINE_" to make their purpose more clear. -Allow for variations that change state during iteration by sending a separate copy of the ember to each rendering thread. -Implement this same functionality with a local struct in OpenCL. It's members are the total of all variables that need to change state within an ember. -Add Contains() function to Utils.h. -EmberRender: print names of kernels being printed with --dump_kernel option. -Clean up EmberTester to handle some of the recent changes. -Fix various casts. -Replace % 2 with & 1, even though the compiler was likely doing this already. -Add new file Variations06.h to accommodate new variations. -General cleanup.
2015-11-22 17:15:07 -05:00
}
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
};
/// <summary>
/// coshq.
/// </summary>
template <typename T>
--Code changes -Split VariationList into .h/cpp files. --Make VariationList.cpp the only place where the variations files are included from. This speeds up the build, but variations can no longer be directly instantiated. --Variations are no longer exported.
2016-02-18 21:58:24 -05:00
class CoshqVariation : public Variation<T>
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
public:
--Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums.
2016-01-04 19:50:15 -05:00
CoshqVariation(T weight = 1.0) : Variation<T>("coshq", eVariationId::VAR_COSHQ, weight) { }
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
VARCOPY(CoshqVariation)
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
--Code changes -Make usage of VarFuncs more consistent by moving some global functions there. -Make CircleTrans1 a little safer by calling Zeps() on m_Sc during precalc. -Begin prepping for 0.9.9.5 release.
2016-03-04 21:54:06 -05:00
T absV = VarFuncs<T>::Hypot(helper.In.y, helper.In.z);
--User changes -Add new variations: bubbleT3D, crob, hexaplay3D, hexcrop, hexes, hexnix3D, loonie2, loonie3, nBlur, octapol and synth. -Allow for pre/post versions of dc_bubble, dc_cylinder and dc_linear whereas before they were omitted. -When saving a file with multiple embers in it, detect if time values are all the same and if so, start them at zero and increment by 1 for each ember. -Allow for numerous quality increases to be coalesced into one. It will pick up at the end of the current render. -Show selection highlight on variations tree in response to mouse hover. This makes it easier to see for which variation or param the current mouse wheel action will apply. -Make default temporal samples be 100, whereas before it was 1000 which was overkill. -Require the shift key to be held with delete for deleting an ember to prevent it from triggering when the user enters delete in the edit box. -This wasn't otherwise fixable without writing a lot more code. --Bug fixes -EmberGenome was crashing when generating a sequence from a source file with more than 2 embers in it. -EmberGenome was improperly handling the first frame of a merge after the last frame of the loop. -These bugs were due to a previous commit. Revert parts of that commit. -Prevent a zoom value of less than 0 when reading from xml. -Slight optimization of the crescents, and mask variations, if the compiler wasn't doing it already. -Unique file naming was broken because it was looking for _# and the default names ended with -#. -Disallow renaming of an ember in the library tree to an empty string. -Severe bug that prevented some variations from being read correctly from params generated outside this program. -Severe OpenCL randomization bug. The first x coordinates of the first points in the first kernel call of the first ember of a render since the OpenCL renderer object was created were not random and were mostly -1. -Severe bug when populating xform selection distributions that could sometimes cause a crash due to roundoff error. Fix by using double. -Limit the max number of variations in a random ember to MAX_CL_VARS, which is 8. This ensures they'll look the same on CPU and GPU. -Prevent user from saving stylesheet to default.qss, it's a special reserved filename. --Code changes -Generalize using the running sum output point inside of a variation for all cases: pre, reg and post. -Allow for array variables in variations where the address of each element is stored in m_Params. -Qualify all math functions with std:: -No longer use our own Clamp() in OpenCL, instead use the standard clamp(). -Redesign how functions are used in the variations OpenCL code. -Add tests to EmberTester to verify some of the new functionality. -Place more const and override qualifiers on functions where appropriate. -Add a global rand with a lock to be used very sparingly. -Use a map instead of a vector for bad param names in Xml parsing. -Prefix affine interpolation mode defines with "AFFINE_" to make their purpose more clear. -Allow for variations that change state during iteration by sending a separate copy of the ember to each rendering thread. -Implement this same functionality with a local struct in OpenCL. It's members are the total of all variables that need to change state within an ember. -Add Contains() function to Utils.h. -EmberRender: print names of kernels being printed with --dump_kernel option. -Clean up EmberTester to handle some of the recent changes. -Fix various casts. -Replace % 2 with & 1, even though the compiler was likely doing this already. -Add new file Variations06.h to accommodate new variations. -General cleanup.
2015-11-22 17:15:07 -05:00
T s = std::sin(absV);
T c = std::cos(absV);
T sh = std::sinh(helper.In.x);
T ch = std::cosh(helper.In.x);
06/04/2017 --User changes -Make all fonts be MS Shell Dlg 2. This will require reloading dark.qss if users are already using it. -Limit size of the left side of the palette editor. -Disable create from image buttons in the palette editor when working on a fixed palette. --Bug fixes -The following variations were wrong: coshq, cothq. -During iteration, the color index could become nan if all xform color speeds were negative. This could lead to bad results on the GPU. Fix to check for nan. Minimal speed difference. --Code changes -Make the following variations safer by using Zeps(): sinq, sinhq, secq, sechq, tanq, tanhq, cosq, coshq, cotq, cothq, cscq, cschq, estiq. -Always pass -cl-no-signed-zeros -cl-denorms-are-zero to kernel compiles for both single and double. -Flush all denormals to zero for all executable programs. This will likely lead to a speedup for badly behaving programs.
2017-06-04 20:37:29 -04:00
T d = m_Weight * sh * s / Zeps(absV);
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
helper.Out.x = m_Weight * c * ch;
helper.Out.y = d * helper.In.y;
helper.Out.z = d * helper.In.z;
}
--Code changes -Better use of const functions. -Ember.pro should not define BUILDING_EMBERCL, instead define BUILDING_EMBER.
2015-06-29 23:13:53 -04:00
virtual string OpenCLString() const override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
ostringstream ss;
0.4.1.3 Beta 10/14/2014 --User Changes Size is no longer fixed to the window size. Size scaling is done differently in the final render dialog. This fixes several bugs. Remove Xml saving size from settings and options dialog, it no longer applies. Final render can be broken into strips. Set default save path to the desktop if none is found in the settings file. Set default output size to 1920x1080 if none is found in the settings file. --Bug Fixes Better memory size reporting in final render dialog. --Code Changes Migrate to C++11, Qt 5.3.1, and Visual Studio 2013. Change most instances of unsigned int to size_t, and int to intmax_t. Add m_OrigPixPerUnit and m_ScaleType to Ember for scaling purposes. Replace some sprintf_s() calls in XmlToEmber with ostringstream. Move more non-templated members into RendererBase. Add CopyVec() overload that takes a per element function pointer. Add vector Memset(). Replace '&' with '+' instead of "&amp;" in XmlToEmber for much faster parsing. Break strips rendering out into EmberCommon and call from EmberRender and Fractorium. Make AddAndWriteBuffer() just call WriteBuffer(). Make AddAndWriteImage() delete the existing image first before replacing it. Add SetOutputTexture() to RendererCL to support making new textures in response to resize events. Remove multiple return statements in RendererCL, and replace with a bool that tracks results. Add ToDouble(), MakeEnd(), ToString() and Exists() wrappers in Fractorium. Add Size() wrapper in EmberFile. Make QString function arguments const QString&, and string with const string&. Make ShowCritical() wrapper for invoking a message box from another thread. Add combo box to TwoButtonWidget and rename.
2014-10-14 11:53:15 -04:00
intmax_t varIndex = IndexInXform();
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
string weight = WeightDefineString();
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
ss << "\t{\n"
<< "\t\treal_t absV = Hypot(vIn.y, vIn.z);\n"
<< "\t\treal_t s = sin(absV);\n"
<< "\t\treal_t c = cos(absV);\n"
<< "\t\treal_t sh = sinh(vIn.x);\n"
<< "\t\treal_t ch = cosh(vIn.x);\n"
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
<< "\t\treal_t d = " << weight << " * sh * s / Zeps(absV);\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\n"
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
<< "\t\tvOut.x = " << weight << " * c * ch;\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t\tvOut.y = d * vIn.y;\n"
<< "\t\tvOut.z = d * vIn.z;\n"
<< "\t}\n";
return ss.str();
}
--User changes -Add new variations: bubbleT3D, crob, hexaplay3D, hexcrop, hexes, hexnix3D, loonie2, loonie3, nBlur, octapol and synth. -Allow for pre/post versions of dc_bubble, dc_cylinder and dc_linear whereas before they were omitted. -When saving a file with multiple embers in it, detect if time values are all the same and if so, start them at zero and increment by 1 for each ember. -Allow for numerous quality increases to be coalesced into one. It will pick up at the end of the current render. -Show selection highlight on variations tree in response to mouse hover. This makes it easier to see for which variation or param the current mouse wheel action will apply. -Make default temporal samples be 100, whereas before it was 1000 which was overkill. -Require the shift key to be held with delete for deleting an ember to prevent it from triggering when the user enters delete in the edit box. -This wasn't otherwise fixable without writing a lot more code. --Bug fixes -EmberGenome was crashing when generating a sequence from a source file with more than 2 embers in it. -EmberGenome was improperly handling the first frame of a merge after the last frame of the loop. -These bugs were due to a previous commit. Revert parts of that commit. -Prevent a zoom value of less than 0 when reading from xml. -Slight optimization of the crescents, and mask variations, if the compiler wasn't doing it already. -Unique file naming was broken because it was looking for _# and the default names ended with -#. -Disallow renaming of an ember in the library tree to an empty string. -Severe bug that prevented some variations from being read correctly from params generated outside this program. -Severe OpenCL randomization bug. The first x coordinates of the first points in the first kernel call of the first ember of a render since the OpenCL renderer object was created were not random and were mostly -1. -Severe bug when populating xform selection distributions that could sometimes cause a crash due to roundoff error. Fix by using double. -Limit the max number of variations in a random ember to MAX_CL_VARS, which is 8. This ensures they'll look the same on CPU and GPU. -Prevent user from saving stylesheet to default.qss, it's a special reserved filename. --Code changes -Generalize using the running sum output point inside of a variation for all cases: pre, reg and post. -Allow for array variables in variations where the address of each element is stored in m_Params. -Qualify all math functions with std:: -No longer use our own Clamp() in OpenCL, instead use the standard clamp(). -Redesign how functions are used in the variations OpenCL code. -Add tests to EmberTester to verify some of the new functionality. -Place more const and override qualifiers on functions where appropriate. -Add a global rand with a lock to be used very sparingly. -Use a map instead of a vector for bad param names in Xml parsing. -Prefix affine interpolation mode defines with "AFFINE_" to make their purpose more clear. -Allow for variations that change state during iteration by sending a separate copy of the ember to each rendering thread. -Implement this same functionality with a local struct in OpenCL. It's members are the total of all variables that need to change state within an ember. -Add Contains() function to Utils.h. -EmberRender: print names of kernels being printed with --dump_kernel option. -Clean up EmberTester to handle some of the recent changes. -Fix various casts. -Replace % 2 with & 1, even though the compiler was likely doing this already. -Add new file Variations06.h to accommodate new variations. -General cleanup.
2015-11-22 17:15:07 -05:00
virtual vector<string> OpenCLGlobalFuncNames() const override
{
06/04/2017 --User changes -Make all fonts be MS Shell Dlg 2. This will require reloading dark.qss if users are already using it. -Limit size of the left side of the palette editor. -Disable create from image buttons in the palette editor when working on a fixed palette. --Bug fixes -The following variations were wrong: coshq, cothq. -During iteration, the color index could become nan if all xform color speeds were negative. This could lead to bad results on the GPU. Fix to check for nan. Minimal speed difference. --Code changes -Make the following variations safer by using Zeps(): sinq, sinhq, secq, sechq, tanq, tanhq, cosq, coshq, cotq, cothq, cscq, cschq, estiq. -Always pass -cl-no-signed-zeros -cl-denorms-are-zero to kernel compiles for both single and double. -Flush all denormals to zero for all executable programs. This will likely lead to a speedup for badly behaving programs.
2017-06-04 20:37:29 -04:00
return vector<string> { "Hypot", "Zeps" };
--User changes -Add new variations: bubbleT3D, crob, hexaplay3D, hexcrop, hexes, hexnix3D, loonie2, loonie3, nBlur, octapol and synth. -Allow for pre/post versions of dc_bubble, dc_cylinder and dc_linear whereas before they were omitted. -When saving a file with multiple embers in it, detect if time values are all the same and if so, start them at zero and increment by 1 for each ember. -Allow for numerous quality increases to be coalesced into one. It will pick up at the end of the current render. -Show selection highlight on variations tree in response to mouse hover. This makes it easier to see for which variation or param the current mouse wheel action will apply. -Make default temporal samples be 100, whereas before it was 1000 which was overkill. -Require the shift key to be held with delete for deleting an ember to prevent it from triggering when the user enters delete in the edit box. -This wasn't otherwise fixable without writing a lot more code. --Bug fixes -EmberGenome was crashing when generating a sequence from a source file with more than 2 embers in it. -EmberGenome was improperly handling the first frame of a merge after the last frame of the loop. -These bugs were due to a previous commit. Revert parts of that commit. -Prevent a zoom value of less than 0 when reading from xml. -Slight optimization of the crescents, and mask variations, if the compiler wasn't doing it already. -Unique file naming was broken because it was looking for _# and the default names ended with -#. -Disallow renaming of an ember in the library tree to an empty string. -Severe bug that prevented some variations from being read correctly from params generated outside this program. -Severe OpenCL randomization bug. The first x coordinates of the first points in the first kernel call of the first ember of a render since the OpenCL renderer object was created were not random and were mostly -1. -Severe bug when populating xform selection distributions that could sometimes cause a crash due to roundoff error. Fix by using double. -Limit the max number of variations in a random ember to MAX_CL_VARS, which is 8. This ensures they'll look the same on CPU and GPU. -Prevent user from saving stylesheet to default.qss, it's a special reserved filename. --Code changes -Generalize using the running sum output point inside of a variation for all cases: pre, reg and post. -Allow for array variables in variations where the address of each element is stored in m_Params. -Qualify all math functions with std:: -No longer use our own Clamp() in OpenCL, instead use the standard clamp(). -Redesign how functions are used in the variations OpenCL code. -Add tests to EmberTester to verify some of the new functionality. -Place more const and override qualifiers on functions where appropriate. -Add a global rand with a lock to be used very sparingly. -Use a map instead of a vector for bad param names in Xml parsing. -Prefix affine interpolation mode defines with "AFFINE_" to make their purpose more clear. -Allow for variations that change state during iteration by sending a separate copy of the ember to each rendering thread. -Implement this same functionality with a local struct in OpenCL. It's members are the total of all variables that need to change state within an ember. -Add Contains() function to Utils.h. -EmberRender: print names of kernels being printed with --dump_kernel option. -Clean up EmberTester to handle some of the recent changes. -Fix various casts. -Replace % 2 with & 1, even though the compiler was likely doing this already. -Add new file Variations06.h to accommodate new variations. -General cleanup.
2015-11-22 17:15:07 -05:00
}
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
};
/// <summary>
/// cotq.
/// </summary>
template <typename T>
--Code changes -Split VariationList into .h/cpp files. --Make VariationList.cpp the only place where the variations files are included from. This speeds up the build, but variations can no longer be directly instantiated. --Variations are no longer exported.
2016-02-18 21:58:24 -05:00
class CotqVariation : public Variation<T>
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
public:
--Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums.
2016-01-04 19:50:15 -05:00
CotqVariation(T weight = 1.0) : Variation<T>("cotq", eVariationId::VAR_COTQ, weight) { }
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
VARCOPY(CotqVariation)
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
T sysz = SQR(helper.In.y) + SQR(helper.In.z);
Explicitly namespace calls to sqrt and isnan osx/linux have alternative top-level implementations of these methods, so the calls are ambiguous given the - so these should be used explicitly
2015-04-13 07:32:58 -04:00
T absV = std::sqrt(sysz);
06/04/2017 --User changes -Make all fonts be MS Shell Dlg 2. This will require reloading dark.qss if users are already using it. -Limit size of the left side of the palette editor. -Disable create from image buttons in the palette editor when working on a fixed palette. --Bug fixes -The following variations were wrong: coshq, cothq. -During iteration, the color index could become nan if all xform color speeds were negative. This could lead to bad results on the GPU. Fix to check for nan. Minimal speed difference. --Code changes -Make the following variations safer by using Zeps(): sinq, sinhq, secq, sechq, tanq, tanhq, cosq, coshq, cotq, cothq, cscq, cschq, estiq. -Always pass -cl-no-signed-zeros -cl-denorms-are-zero to kernel compiles for both single and double. -Flush all denormals to zero for all executable programs. This will likely lead to a speedup for badly behaving programs.
2017-06-04 20:37:29 -04:00
T ni = m_Weight / Zeps(SQR(helper.In.x) + sysz);
--User changes -Add new variations: bubbleT3D, crob, hexaplay3D, hexcrop, hexes, hexnix3D, loonie2, loonie3, nBlur, octapol and synth. -Allow for pre/post versions of dc_bubble, dc_cylinder and dc_linear whereas before they were omitted. -When saving a file with multiple embers in it, detect if time values are all the same and if so, start them at zero and increment by 1 for each ember. -Allow for numerous quality increases to be coalesced into one. It will pick up at the end of the current render. -Show selection highlight on variations tree in response to mouse hover. This makes it easier to see for which variation or param the current mouse wheel action will apply. -Make default temporal samples be 100, whereas before it was 1000 which was overkill. -Require the shift key to be held with delete for deleting an ember to prevent it from triggering when the user enters delete in the edit box. -This wasn't otherwise fixable without writing a lot more code. --Bug fixes -EmberGenome was crashing when generating a sequence from a source file with more than 2 embers in it. -EmberGenome was improperly handling the first frame of a merge after the last frame of the loop. -These bugs were due to a previous commit. Revert parts of that commit. -Prevent a zoom value of less than 0 when reading from xml. -Slight optimization of the crescents, and mask variations, if the compiler wasn't doing it already. -Unique file naming was broken because it was looking for _# and the default names ended with -#. -Disallow renaming of an ember in the library tree to an empty string. -Severe bug that prevented some variations from being read correctly from params generated outside this program. -Severe OpenCL randomization bug. The first x coordinates of the first points in the first kernel call of the first ember of a render since the OpenCL renderer object was created were not random and were mostly -1. -Severe bug when populating xform selection distributions that could sometimes cause a crash due to roundoff error. Fix by using double. -Limit the max number of variations in a random ember to MAX_CL_VARS, which is 8. This ensures they'll look the same on CPU and GPU. -Prevent user from saving stylesheet to default.qss, it's a special reserved filename. --Code changes -Generalize using the running sum output point inside of a variation for all cases: pre, reg and post. -Allow for array variables in variations where the address of each element is stored in m_Params. -Qualify all math functions with std:: -No longer use our own Clamp() in OpenCL, instead use the standard clamp(). -Redesign how functions are used in the variations OpenCL code. -Add tests to EmberTester to verify some of the new functionality. -Place more const and override qualifiers on functions where appropriate. -Add a global rand with a lock to be used very sparingly. -Use a map instead of a vector for bad param names in Xml parsing. -Prefix affine interpolation mode defines with "AFFINE_" to make their purpose more clear. -Allow for variations that change state during iteration by sending a separate copy of the ember to each rendering thread. -Implement this same functionality with a local struct in OpenCL. It's members are the total of all variables that need to change state within an ember. -Add Contains() function to Utils.h. -EmberRender: print names of kernels being printed with --dump_kernel option. -Clean up EmberTester to handle some of the recent changes. -Fix various casts. -Replace % 2 with & 1, even though the compiler was likely doing this already. -Add new file Variations06.h to accommodate new variations. -General cleanup.
2015-11-22 17:15:07 -05:00
T s = std::sin(helper.In.x);
T c = std::cos(helper.In.x);
T sh = std::sinh(absV);
T ch = std::cosh(absV);
06/04/2017 --User changes -Make all fonts be MS Shell Dlg 2. This will require reloading dark.qss if users are already using it. -Limit size of the left side of the palette editor. -Disable create from image buttons in the palette editor when working on a fixed palette. --Bug fixes -The following variations were wrong: coshq, cothq. -During iteration, the color index could become nan if all xform color speeds were negative. This could lead to bad results on the GPU. Fix to check for nan. Minimal speed difference. --Code changes -Make the following variations safer by using Zeps(): sinq, sinhq, secq, sechq, tanq, tanhq, cosq, coshq, cotq, cothq, cscq, cschq, estiq. -Always pass -cl-no-signed-zeros -cl-denorms-are-zero to kernel compiles for both single and double. -Flush all denormals to zero for all executable programs. This will likely lead to a speedup for badly behaving programs.
2017-06-04 20:37:29 -04:00
T d = c * sh / Zeps(absV);
T b = -s * sh / Zeps(absV);
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
T stcv = s * ch;
T nstcv = -stcv;
T ctcv = c * ch;
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
helper.Out.x = (stcv * ctcv + d * b * sysz) * ni;
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
helper.Out.y = -(nstcv * b * helper.In.y + d * helper.In.y * ctcv) * ni;
helper.Out.z = -(nstcv * b * helper.In.z + d * helper.In.z * ctcv) * ni;
}
--Code changes -Better use of const functions. -Ember.pro should not define BUILDING_EMBERCL, instead define BUILDING_EMBER.
2015-06-29 23:13:53 -04:00
virtual string OpenCLString() const override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
ostringstream ss;
0.4.1.3 Beta 10/14/2014 --User Changes Size is no longer fixed to the window size. Size scaling is done differently in the final render dialog. This fixes several bugs. Remove Xml saving size from settings and options dialog, it no longer applies. Final render can be broken into strips. Set default save path to the desktop if none is found in the settings file. Set default output size to 1920x1080 if none is found in the settings file. --Bug Fixes Better memory size reporting in final render dialog. --Code Changes Migrate to C++11, Qt 5.3.1, and Visual Studio 2013. Change most instances of unsigned int to size_t, and int to intmax_t. Add m_OrigPixPerUnit and m_ScaleType to Ember for scaling purposes. Replace some sprintf_s() calls in XmlToEmber with ostringstream. Move more non-templated members into RendererBase. Add CopyVec() overload that takes a per element function pointer. Add vector Memset(). Replace '&' with '+' instead of "&amp;" in XmlToEmber for much faster parsing. Break strips rendering out into EmberCommon and call from EmberRender and Fractorium. Make AddAndWriteBuffer() just call WriteBuffer(). Make AddAndWriteImage() delete the existing image first before replacing it. Add SetOutputTexture() to RendererCL to support making new textures in response to resize events. Remove multiple return statements in RendererCL, and replace with a bool that tracks results. Add ToDouble(), MakeEnd(), ToString() and Exists() wrappers in Fractorium. Add Size() wrapper in EmberFile. Make QString function arguments const QString&, and string with const string&. Make ShowCritical() wrapper for invoking a message box from another thread. Add combo box to TwoButtonWidget and rename.
2014-10-14 11:53:15 -04:00
intmax_t varIndex = IndexInXform();
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
string weight = WeightDefineString();
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
ss << "\t{\n"
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
<< "\t\treal_t sysz = fma(vIn.y, vIn.y, SQR(vIn.z));\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t\treal_t absV = sqrt(sysz);\n"
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
<< "\t\treal_t ni = " << weight << " / Zeps(fma(vIn.x, vIn.x, sysz));\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t\treal_t s = sin(vIn.x);\n"
<< "\t\treal_t c = cos(vIn.x);\n"
<< "\t\treal_t sh = sinh(absV);\n"
<< "\t\treal_t ch = cosh(absV);\n"
06/04/2017 --User changes -Make all fonts be MS Shell Dlg 2. This will require reloading dark.qss if users are already using it. -Limit size of the left side of the palette editor. -Disable create from image buttons in the palette editor when working on a fixed palette. --Bug fixes -The following variations were wrong: coshq, cothq. -During iteration, the color index could become nan if all xform color speeds were negative. This could lead to bad results on the GPU. Fix to check for nan. Minimal speed difference. --Code changes -Make the following variations safer by using Zeps(): sinq, sinhq, secq, sechq, tanq, tanhq, cosq, coshq, cotq, cothq, cscq, cschq, estiq. -Always pass -cl-no-signed-zeros -cl-denorms-are-zero to kernel compiles for both single and double. -Flush all denormals to zero for all executable programs. This will likely lead to a speedup for badly behaving programs.
2017-06-04 20:37:29 -04:00
<< "\t\treal_t d = c * sh / Zeps(absV);\n"
<< "\t\treal_t b = -s * sh / Zeps(absV);\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t\treal_t stcv = s * ch;\n"
<< "\t\treal_t nstcv = -stcv;\n"
<< "\t\treal_t ctcv = c * ch;\n"
<< "\n"
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
<< "\t\tvOut.x = fma(stcv, ctcv, d * b * sysz) * ni;\n"
<< "\t\tvOut.y = -fma(nstcv * b, vIn.y, d * vIn.y * ctcv) * ni;\n"
<< "\t\tvOut.z = -fma(nstcv * b, vIn.z, d * vIn.z * ctcv) * ni;\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t}\n";
return ss.str();
}
06/04/2017 --User changes -Make all fonts be MS Shell Dlg 2. This will require reloading dark.qss if users are already using it. -Limit size of the left side of the palette editor. -Disable create from image buttons in the palette editor when working on a fixed palette. --Bug fixes -The following variations were wrong: coshq, cothq. -During iteration, the color index could become nan if all xform color speeds were negative. This could lead to bad results on the GPU. Fix to check for nan. Minimal speed difference. --Code changes -Make the following variations safer by using Zeps(): sinq, sinhq, secq, sechq, tanq, tanhq, cosq, coshq, cotq, cothq, cscq, cschq, estiq. -Always pass -cl-no-signed-zeros -cl-denorms-are-zero to kernel compiles for both single and double. -Flush all denormals to zero for all executable programs. This will likely lead to a speedup for badly behaving programs.
2017-06-04 20:37:29 -04:00
virtual vector<string> OpenCLGlobalFuncNames() const override
{
return vector<string> { "Zeps" };
}
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
};
/// <summary>
/// cothq.
/// </summary>
template <typename T>
--Code changes -Split VariationList into .h/cpp files. --Make VariationList.cpp the only place where the variations files are included from. This speeds up the build, but variations can no longer be directly instantiated. --Variations are no longer exported.
2016-02-18 21:58:24 -05:00
class CothqVariation : public Variation<T>
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
public:
--Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums.
2016-01-04 19:50:15 -05:00
CothqVariation(T weight = 1.0) : Variation<T>("cothq", eVariationId::VAR_COTHQ, weight) { }
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
VARCOPY(CothqVariation)
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
T sysz = SQR(helper.In.y) + SQR(helper.In.z);
Explicitly namespace calls to sqrt and isnan osx/linux have alternative top-level implementations of these methods, so the calls are ambiguous given the - so these should be used explicitly
2015-04-13 07:32:58 -04:00
T absV = std::sqrt(sysz);
06/04/2017 --User changes -Make all fonts be MS Shell Dlg 2. This will require reloading dark.qss if users are already using it. -Limit size of the left side of the palette editor. -Disable create from image buttons in the palette editor when working on a fixed palette. --Bug fixes -The following variations were wrong: coshq, cothq. -During iteration, the color index could become nan if all xform color speeds were negative. This could lead to bad results on the GPU. Fix to check for nan. Minimal speed difference. --Code changes -Make the following variations safer by using Zeps(): sinq, sinhq, secq, sechq, tanq, tanhq, cosq, coshq, cotq, cothq, cscq, cschq, estiq. -Always pass -cl-no-signed-zeros -cl-denorms-are-zero to kernel compiles for both single and double. -Flush all denormals to zero for all executable programs. This will likely lead to a speedup for badly behaving programs.
2017-06-04 20:37:29 -04:00
T ni = m_Weight / Zeps(Sqr(SQR(helper.In.x) + sysz));
--User changes -Add new variations: bubbleT3D, crob, hexaplay3D, hexcrop, hexes, hexnix3D, loonie2, loonie3, nBlur, octapol and synth. -Allow for pre/post versions of dc_bubble, dc_cylinder and dc_linear whereas before they were omitted. -When saving a file with multiple embers in it, detect if time values are all the same and if so, start them at zero and increment by 1 for each ember. -Allow for numerous quality increases to be coalesced into one. It will pick up at the end of the current render. -Show selection highlight on variations tree in response to mouse hover. This makes it easier to see for which variation or param the current mouse wheel action will apply. -Make default temporal samples be 100, whereas before it was 1000 which was overkill. -Require the shift key to be held with delete for deleting an ember to prevent it from triggering when the user enters delete in the edit box. -This wasn't otherwise fixable without writing a lot more code. --Bug fixes -EmberGenome was crashing when generating a sequence from a source file with more than 2 embers in it. -EmberGenome was improperly handling the first frame of a merge after the last frame of the loop. -These bugs were due to a previous commit. Revert parts of that commit. -Prevent a zoom value of less than 0 when reading from xml. -Slight optimization of the crescents, and mask variations, if the compiler wasn't doing it already. -Unique file naming was broken because it was looking for _# and the default names ended with -#. -Disallow renaming of an ember in the library tree to an empty string. -Severe bug that prevented some variations from being read correctly from params generated outside this program. -Severe OpenCL randomization bug. The first x coordinates of the first points in the first kernel call of the first ember of a render since the OpenCL renderer object was created were not random and were mostly -1. -Severe bug when populating xform selection distributions that could sometimes cause a crash due to roundoff error. Fix by using double. -Limit the max number of variations in a random ember to MAX_CL_VARS, which is 8. This ensures they'll look the same on CPU and GPU. -Prevent user from saving stylesheet to default.qss, it's a special reserved filename. --Code changes -Generalize using the running sum output point inside of a variation for all cases: pre, reg and post. -Allow for array variables in variations where the address of each element is stored in m_Params. -Qualify all math functions with std:: -No longer use our own Clamp() in OpenCL, instead use the standard clamp(). -Redesign how functions are used in the variations OpenCL code. -Add tests to EmberTester to verify some of the new functionality. -Place more const and override qualifiers on functions where appropriate. -Add a global rand with a lock to be used very sparingly. -Use a map instead of a vector for bad param names in Xml parsing. -Prefix affine interpolation mode defines with "AFFINE_" to make their purpose more clear. -Allow for variations that change state during iteration by sending a separate copy of the ember to each rendering thread. -Implement this same functionality with a local struct in OpenCL. It's members are the total of all variables that need to change state within an ember. -Add Contains() function to Utils.h. -EmberRender: print names of kernels being printed with --dump_kernel option. -Clean up EmberTester to handle some of the recent changes. -Fix various casts. -Replace % 2 with & 1, even though the compiler was likely doing this already. -Add new file Variations06.h to accommodate new variations. -General cleanup.
2015-11-22 17:15:07 -05:00
T s = std::sin(absV);
T c = std::cos(absV);
T sh = std::sinh(helper.In.x);
T ch = std::cosh(helper.In.x);
06/04/2017 --User changes -Make all fonts be MS Shell Dlg 2. This will require reloading dark.qss if users are already using it. -Limit size of the left side of the palette editor. -Disable create from image buttons in the palette editor when working on a fixed palette. --Bug fixes -The following variations were wrong: coshq, cothq. -During iteration, the color index could become nan if all xform color speeds were negative. This could lead to bad results on the GPU. Fix to check for nan. Minimal speed difference. --Code changes -Make the following variations safer by using Zeps(): sinq, sinhq, secq, sechq, tanq, tanhq, cosq, coshq, cotq, cothq, cscq, cschq, estiq. -Always pass -cl-no-signed-zeros -cl-denorms-are-zero to kernel compiles for both single and double. -Flush all denormals to zero for all executable programs. This will likely lead to a speedup for badly behaving programs.
2017-06-04 20:37:29 -04:00
T d = ch * s / Zeps(absV);
T b = sh * s / Zeps(absV);
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
T stcv = sh * c;
T nstcv = -stcv;
T ctcv = ch * c;
06/04/2017 --User changes -Make all fonts be MS Shell Dlg 2. This will require reloading dark.qss if users are already using it. -Limit size of the left side of the palette editor. -Disable create from image buttons in the palette editor when working on a fixed palette. --Bug fixes -The following variations were wrong: coshq, cothq. -During iteration, the color index could become nan if all xform color speeds were negative. This could lead to bad results on the GPU. Fix to check for nan. Minimal speed difference. --Code changes -Make the following variations safer by using Zeps(): sinq, sinhq, secq, sechq, tanq, tanhq, cosq, coshq, cotq, cothq, cscq, cschq, estiq. -Always pass -cl-no-signed-zeros -cl-denorms-are-zero to kernel compiles for both single and double. -Flush all denormals to zero for all executable programs. This will likely lead to a speedup for badly behaving programs.
2017-06-04 20:37:29 -04:00
helper.Out.x = (stcv * ctcv + d * b * sysz) * ni;
helper.Out.y = (nstcv * b * helper.In.y + d * helper.In.y * ctcv) * ni;
helper.Out.z = (nstcv * b * helper.In.z + d * helper.In.z * ctcv) * ni;
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
}
--Code changes -Better use of const functions. -Ember.pro should not define BUILDING_EMBERCL, instead define BUILDING_EMBER.
2015-06-29 23:13:53 -04:00
virtual string OpenCLString() const override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
ostringstream ss;
0.4.1.3 Beta 10/14/2014 --User Changes Size is no longer fixed to the window size. Size scaling is done differently in the final render dialog. This fixes several bugs. Remove Xml saving size from settings and options dialog, it no longer applies. Final render can be broken into strips. Set default save path to the desktop if none is found in the settings file. Set default output size to 1920x1080 if none is found in the settings file. --Bug Fixes Better memory size reporting in final render dialog. --Code Changes Migrate to C++11, Qt 5.3.1, and Visual Studio 2013. Change most instances of unsigned int to size_t, and int to intmax_t. Add m_OrigPixPerUnit and m_ScaleType to Ember for scaling purposes. Replace some sprintf_s() calls in XmlToEmber with ostringstream. Move more non-templated members into RendererBase. Add CopyVec() overload that takes a per element function pointer. Add vector Memset(). Replace '&' with '+' instead of "&amp;" in XmlToEmber for much faster parsing. Break strips rendering out into EmberCommon and call from EmberRender and Fractorium. Make AddAndWriteBuffer() just call WriteBuffer(). Make AddAndWriteImage() delete the existing image first before replacing it. Add SetOutputTexture() to RendererCL to support making new textures in response to resize events. Remove multiple return statements in RendererCL, and replace with a bool that tracks results. Add ToDouble(), MakeEnd(), ToString() and Exists() wrappers in Fractorium. Add Size() wrapper in EmberFile. Make QString function arguments const QString&, and string with const string&. Make ShowCritical() wrapper for invoking a message box from another thread. Add combo box to TwoButtonWidget and rename.
2014-10-14 11:53:15 -04:00
intmax_t varIndex = IndexInXform();
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
string weight = WeightDefineString();
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
ss << "\t{\n"
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
<< "\t\treal_t sysz = fma(vIn.y, vIn.y, SQR(vIn.z));\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t\treal_t absV = sqrt(sysz);\n"
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
<< "\t\treal_t ni = " << weight << " / Zeps(Sqr(fma(vIn.x, vIn.x, sysz)));\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t\treal_t s = sin(absV);\n"
<< "\t\treal_t c = cos(absV);\n"
<< "\t\treal_t sh = sinh(vIn.x);\n"
<< "\t\treal_t ch = cosh(vIn.x);\n"
06/04/2017 --User changes -Make all fonts be MS Shell Dlg 2. This will require reloading dark.qss if users are already using it. -Limit size of the left side of the palette editor. -Disable create from image buttons in the palette editor when working on a fixed palette. --Bug fixes -The following variations were wrong: coshq, cothq. -During iteration, the color index could become nan if all xform color speeds were negative. This could lead to bad results on the GPU. Fix to check for nan. Minimal speed difference. --Code changes -Make the following variations safer by using Zeps(): sinq, sinhq, secq, sechq, tanq, tanhq, cosq, coshq, cotq, cothq, cscq, cschq, estiq. -Always pass -cl-no-signed-zeros -cl-denorms-are-zero to kernel compiles for both single and double. -Flush all denormals to zero for all executable programs. This will likely lead to a speedup for badly behaving programs.
2017-06-04 20:37:29 -04:00
<< "\t\treal_t d = ch * s / Zeps(absV);\n"
<< "\t\treal_t b = sh * s / Zeps(absV);\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t\treal_t stcv = sh * c;\n"
<< "\t\treal_t nstcv = -stcv;\n"
<< "\t\treal_t ctcv = ch * c;\n"
<< "\n"
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
<< "\t\tvOut.x = fma(stcv, ctcv, d * b * sysz) * ni;\n"
<< "\t\tvOut.y = fma(nstcv * b, vIn.y, d * vIn.y * ctcv) * ni;\n"
<< "\t\tvOut.z = fma(nstcv * b, vIn.z, d * vIn.z * ctcv) * ni;\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t}\n";
return ss.str();
}
06/04/2017 --User changes -Make all fonts be MS Shell Dlg 2. This will require reloading dark.qss if users are already using it. -Limit size of the left side of the palette editor. -Disable create from image buttons in the palette editor when working on a fixed palette. --Bug fixes -The following variations were wrong: coshq, cothq. -During iteration, the color index could become nan if all xform color speeds were negative. This could lead to bad results on the GPU. Fix to check for nan. Minimal speed difference. --Code changes -Make the following variations safer by using Zeps(): sinq, sinhq, secq, sechq, tanq, tanhq, cosq, coshq, cotq, cothq, cscq, cschq, estiq. -Always pass -cl-no-signed-zeros -cl-denorms-are-zero to kernel compiles for both single and double. -Flush all denormals to zero for all executable programs. This will likely lead to a speedup for badly behaving programs.
2017-06-04 20:37:29 -04:00
virtual vector<string> OpenCLGlobalFuncNames() const override
{
return vector<string> { "Zeps", "Sqr" };
}
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
};
/// <summary>
/// cscq.
/// </summary>
template <typename T>
--Code changes -Split VariationList into .h/cpp files. --Make VariationList.cpp the only place where the variations files are included from. This speeds up the build, but variations can no longer be directly instantiated. --Variations are no longer exported.
2016-02-18 21:58:24 -05:00
class CscqVariation : public Variation<T>
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
public:
--Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums.
2016-01-04 19:50:15 -05:00
CscqVariation(T weight = 1.0) : Variation<T>("cscq", eVariationId::VAR_CSCQ, weight, true) { }
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
VARCOPY(CscqVariation)
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
--Code changes -Make usage of VarFuncs more consistent by moving some global functions there. -Make CircleTrans1 a little safer by calling Zeps() on m_Sc during precalc. -Begin prepping for 0.9.9.5 release.
2016-03-04 21:54:06 -05:00
T absV = VarFuncs<T>::Hypot(helper.In.y, helper.In.z);
06/09/2017 --User changes -dark.qss is now per-OS. -Properly set/reload palette when coming from the palette editor. The latter must be done if they've modified the current palette even if they've clicked cancel. --Bug fixes -Make the following variations safer by using Zeps(): conic, bipolar, edisc, whorl, tan, csc, cot, tanh, sech, csch, coth, auger, bwraps, hypertile3d, hypertile3d1, ortho, poincare, rational3, barycentroid, sschecks, cscq, cschq, scry_3D, splitbrdr, hexcrop, nblur, crob. -Fix bug enabling/disabling overwrite button in palette editor. -Small optimization for gdoffs, use precalcAtanYX. -Properly propagate z through circlesplit, cylinder2 and tile_log variations. -Some values in truchet_fill could've been NaN. --Code changes -Make most installation files read only. -Qualify many calls with std:: to ensure they're not colliding with glm:: -Use auto in more places.
2017-06-09 22:38:06 -04:00
T ni = m_Weight / Zeps(helper.m_PrecalcSumSquares + SQR(helper.In.z));
--User changes -Add new variations: bubbleT3D, crob, hexaplay3D, hexcrop, hexes, hexnix3D, loonie2, loonie3, nBlur, octapol and synth. -Allow for pre/post versions of dc_bubble, dc_cylinder and dc_linear whereas before they were omitted. -When saving a file with multiple embers in it, detect if time values are all the same and if so, start them at zero and increment by 1 for each ember. -Allow for numerous quality increases to be coalesced into one. It will pick up at the end of the current render. -Show selection highlight on variations tree in response to mouse hover. This makes it easier to see for which variation or param the current mouse wheel action will apply. -Make default temporal samples be 100, whereas before it was 1000 which was overkill. -Require the shift key to be held with delete for deleting an ember to prevent it from triggering when the user enters delete in the edit box. -This wasn't otherwise fixable without writing a lot more code. --Bug fixes -EmberGenome was crashing when generating a sequence from a source file with more than 2 embers in it. -EmberGenome was improperly handling the first frame of a merge after the last frame of the loop. -These bugs were due to a previous commit. Revert parts of that commit. -Prevent a zoom value of less than 0 when reading from xml. -Slight optimization of the crescents, and mask variations, if the compiler wasn't doing it already. -Unique file naming was broken because it was looking for _# and the default names ended with -#. -Disallow renaming of an ember in the library tree to an empty string. -Severe bug that prevented some variations from being read correctly from params generated outside this program. -Severe OpenCL randomization bug. The first x coordinates of the first points in the first kernel call of the first ember of a render since the OpenCL renderer object was created were not random and were mostly -1. -Severe bug when populating xform selection distributions that could sometimes cause a crash due to roundoff error. Fix by using double. -Limit the max number of variations in a random ember to MAX_CL_VARS, which is 8. This ensures they'll look the same on CPU and GPU. -Prevent user from saving stylesheet to default.qss, it's a special reserved filename. --Code changes -Generalize using the running sum output point inside of a variation for all cases: pre, reg and post. -Allow for array variables in variations where the address of each element is stored in m_Params. -Qualify all math functions with std:: -No longer use our own Clamp() in OpenCL, instead use the standard clamp(). -Redesign how functions are used in the variations OpenCL code. -Add tests to EmberTester to verify some of the new functionality. -Place more const and override qualifiers on functions where appropriate. -Add a global rand with a lock to be used very sparingly. -Use a map instead of a vector for bad param names in Xml parsing. -Prefix affine interpolation mode defines with "AFFINE_" to make their purpose more clear. -Allow for variations that change state during iteration by sending a separate copy of the ember to each rendering thread. -Implement this same functionality with a local struct in OpenCL. It's members are the total of all variables that need to change state within an ember. -Add Contains() function to Utils.h. -EmberRender: print names of kernels being printed with --dump_kernel option. -Clean up EmberTester to handle some of the recent changes. -Fix various casts. -Replace % 2 with & 1, even though the compiler was likely doing this already. -Add new file Variations06.h to accommodate new variations. -General cleanup.
2015-11-22 17:15:07 -05:00
T s = std::sin(helper.In.x);
T c = std::cos(helper.In.x);
T sh = std::sinh(absV);
T ch = std::cosh(absV);
06/04/2017 --User changes -Make all fonts be MS Shell Dlg 2. This will require reloading dark.qss if users are already using it. -Limit size of the left side of the palette editor. -Disable create from image buttons in the palette editor when working on a fixed palette. --Bug fixes -The following variations were wrong: coshq, cothq. -During iteration, the color index could become nan if all xform color speeds were negative. This could lead to bad results on the GPU. Fix to check for nan. Minimal speed difference. --Code changes -Make the following variations safer by using Zeps(): sinq, sinhq, secq, sechq, tanq, tanhq, cosq, coshq, cotq, cothq, cscq, cschq, estiq. -Always pass -cl-no-signed-zeros -cl-denorms-are-zero to kernel compiles for both single and double. -Flush all denormals to zero for all executable programs. This will likely lead to a speedup for badly behaving programs.
2017-06-04 20:37:29 -04:00
T d = ni * c * sh / Zeps(absV);
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
helper.Out.x = s * ch * ni;
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
helper.Out.y = -(d * helper.In.y);
helper.Out.z = -(d * helper.In.z);
}
--Code changes -Better use of const functions. -Ember.pro should not define BUILDING_EMBERCL, instead define BUILDING_EMBER.
2015-06-29 23:13:53 -04:00
virtual string OpenCLString() const override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
ostringstream ss;
0.4.1.3 Beta 10/14/2014 --User Changes Size is no longer fixed to the window size. Size scaling is done differently in the final render dialog. This fixes several bugs. Remove Xml saving size from settings and options dialog, it no longer applies. Final render can be broken into strips. Set default save path to the desktop if none is found in the settings file. Set default output size to 1920x1080 if none is found in the settings file. --Bug Fixes Better memory size reporting in final render dialog. --Code Changes Migrate to C++11, Qt 5.3.1, and Visual Studio 2013. Change most instances of unsigned int to size_t, and int to intmax_t. Add m_OrigPixPerUnit and m_ScaleType to Ember for scaling purposes. Replace some sprintf_s() calls in XmlToEmber with ostringstream. Move more non-templated members into RendererBase. Add CopyVec() overload that takes a per element function pointer. Add vector Memset(). Replace '&' with '+' instead of "&amp;" in XmlToEmber for much faster parsing. Break strips rendering out into EmberCommon and call from EmberRender and Fractorium. Make AddAndWriteBuffer() just call WriteBuffer(). Make AddAndWriteImage() delete the existing image first before replacing it. Add SetOutputTexture() to RendererCL to support making new textures in response to resize events. Remove multiple return statements in RendererCL, and replace with a bool that tracks results. Add ToDouble(), MakeEnd(), ToString() and Exists() wrappers in Fractorium. Add Size() wrapper in EmberFile. Make QString function arguments const QString&, and string with const string&. Make ShowCritical() wrapper for invoking a message box from another thread. Add combo box to TwoButtonWidget and rename.
2014-10-14 11:53:15 -04:00
intmax_t varIndex = IndexInXform();
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
string weight = WeightDefineString();
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
ss << "\t{\n"
<< "\t\treal_t absV = Hypot(vIn.y, vIn.z);\n"
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
<< "\t\treal_t ni = " << weight << " / Zeps(fma(vIn.z, vIn.z, precalcSumSquares));\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t\treal_t s = sin(vIn.x);\n"
<< "\t\treal_t c = cos(vIn.x);\n"
<< "\t\treal_t sh = sinh(absV);\n"
<< "\t\treal_t ch = cosh(absV);\n"
06/04/2017 --User changes -Make all fonts be MS Shell Dlg 2. This will require reloading dark.qss if users are already using it. -Limit size of the left side of the palette editor. -Disable create from image buttons in the palette editor when working on a fixed palette. --Bug fixes -The following variations were wrong: coshq, cothq. -During iteration, the color index could become nan if all xform color speeds were negative. This could lead to bad results on the GPU. Fix to check for nan. Minimal speed difference. --Code changes -Make the following variations safer by using Zeps(): sinq, sinhq, secq, sechq, tanq, tanhq, cosq, coshq, cotq, cothq, cscq, cschq, estiq. -Always pass -cl-no-signed-zeros -cl-denorms-are-zero to kernel compiles for both single and double. -Flush all denormals to zero for all executable programs. This will likely lead to a speedup for badly behaving programs.
2017-06-04 20:37:29 -04:00
<< "\t\treal_t d = ni * c * sh / Zeps(absV);\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\n"
<< "\t\tvOut.x = s * ch * ni;\n"
<< "\t\tvOut.y = -(d * vIn.y);\n"
<< "\t\tvOut.z = -(d * vIn.z);\n"
<< "\t}\n";
return ss.str();
}
--User changes -Add new variations: bubbleT3D, crob, hexaplay3D, hexcrop, hexes, hexnix3D, loonie2, loonie3, nBlur, octapol and synth. -Allow for pre/post versions of dc_bubble, dc_cylinder and dc_linear whereas before they were omitted. -When saving a file with multiple embers in it, detect if time values are all the same and if so, start them at zero and increment by 1 for each ember. -Allow for numerous quality increases to be coalesced into one. It will pick up at the end of the current render. -Show selection highlight on variations tree in response to mouse hover. This makes it easier to see for which variation or param the current mouse wheel action will apply. -Make default temporal samples be 100, whereas before it was 1000 which was overkill. -Require the shift key to be held with delete for deleting an ember to prevent it from triggering when the user enters delete in the edit box. -This wasn't otherwise fixable without writing a lot more code. --Bug fixes -EmberGenome was crashing when generating a sequence from a source file with more than 2 embers in it. -EmberGenome was improperly handling the first frame of a merge after the last frame of the loop. -These bugs were due to a previous commit. Revert parts of that commit. -Prevent a zoom value of less than 0 when reading from xml. -Slight optimization of the crescents, and mask variations, if the compiler wasn't doing it already. -Unique file naming was broken because it was looking for _# and the default names ended with -#. -Disallow renaming of an ember in the library tree to an empty string. -Severe bug that prevented some variations from being read correctly from params generated outside this program. -Severe OpenCL randomization bug. The first x coordinates of the first points in the first kernel call of the first ember of a render since the OpenCL renderer object was created were not random and were mostly -1. -Severe bug when populating xform selection distributions that could sometimes cause a crash due to roundoff error. Fix by using double. -Limit the max number of variations in a random ember to MAX_CL_VARS, which is 8. This ensures they'll look the same on CPU and GPU. -Prevent user from saving stylesheet to default.qss, it's a special reserved filename. --Code changes -Generalize using the running sum output point inside of a variation for all cases: pre, reg and post. -Allow for array variables in variations where the address of each element is stored in m_Params. -Qualify all math functions with std:: -No longer use our own Clamp() in OpenCL, instead use the standard clamp(). -Redesign how functions are used in the variations OpenCL code. -Add tests to EmberTester to verify some of the new functionality. -Place more const and override qualifiers on functions where appropriate. -Add a global rand with a lock to be used very sparingly. -Use a map instead of a vector for bad param names in Xml parsing. -Prefix affine interpolation mode defines with "AFFINE_" to make their purpose more clear. -Allow for variations that change state during iteration by sending a separate copy of the ember to each rendering thread. -Implement this same functionality with a local struct in OpenCL. It's members are the total of all variables that need to change state within an ember. -Add Contains() function to Utils.h. -EmberRender: print names of kernels being printed with --dump_kernel option. -Clean up EmberTester to handle some of the recent changes. -Fix various casts. -Replace % 2 with & 1, even though the compiler was likely doing this already. -Add new file Variations06.h to accommodate new variations. -General cleanup.
2015-11-22 17:15:07 -05:00
virtual vector<string> OpenCLGlobalFuncNames() const override
{
06/04/2017 --User changes -Make all fonts be MS Shell Dlg 2. This will require reloading dark.qss if users are already using it. -Limit size of the left side of the palette editor. -Disable create from image buttons in the palette editor when working on a fixed palette. --Bug fixes -The following variations were wrong: coshq, cothq. -During iteration, the color index could become nan if all xform color speeds were negative. This could lead to bad results on the GPU. Fix to check for nan. Minimal speed difference. --Code changes -Make the following variations safer by using Zeps(): sinq, sinhq, secq, sechq, tanq, tanhq, cosq, coshq, cotq, cothq, cscq, cschq, estiq. -Always pass -cl-no-signed-zeros -cl-denorms-are-zero to kernel compiles for both single and double. -Flush all denormals to zero for all executable programs. This will likely lead to a speedup for badly behaving programs.
2017-06-04 20:37:29 -04:00
return vector<string> { "Hypot", "Zeps" };
--User changes -Add new variations: bubbleT3D, crob, hexaplay3D, hexcrop, hexes, hexnix3D, loonie2, loonie3, nBlur, octapol and synth. -Allow for pre/post versions of dc_bubble, dc_cylinder and dc_linear whereas before they were omitted. -When saving a file with multiple embers in it, detect if time values are all the same and if so, start them at zero and increment by 1 for each ember. -Allow for numerous quality increases to be coalesced into one. It will pick up at the end of the current render. -Show selection highlight on variations tree in response to mouse hover. This makes it easier to see for which variation or param the current mouse wheel action will apply. -Make default temporal samples be 100, whereas before it was 1000 which was overkill. -Require the shift key to be held with delete for deleting an ember to prevent it from triggering when the user enters delete in the edit box. -This wasn't otherwise fixable without writing a lot more code. --Bug fixes -EmberGenome was crashing when generating a sequence from a source file with more than 2 embers in it. -EmberGenome was improperly handling the first frame of a merge after the last frame of the loop. -These bugs were due to a previous commit. Revert parts of that commit. -Prevent a zoom value of less than 0 when reading from xml. -Slight optimization of the crescents, and mask variations, if the compiler wasn't doing it already. -Unique file naming was broken because it was looking for _# and the default names ended with -#. -Disallow renaming of an ember in the library tree to an empty string. -Severe bug that prevented some variations from being read correctly from params generated outside this program. -Severe OpenCL randomization bug. The first x coordinates of the first points in the first kernel call of the first ember of a render since the OpenCL renderer object was created were not random and were mostly -1. -Severe bug when populating xform selection distributions that could sometimes cause a crash due to roundoff error. Fix by using double. -Limit the max number of variations in a random ember to MAX_CL_VARS, which is 8. This ensures they'll look the same on CPU and GPU. -Prevent user from saving stylesheet to default.qss, it's a special reserved filename. --Code changes -Generalize using the running sum output point inside of a variation for all cases: pre, reg and post. -Allow for array variables in variations where the address of each element is stored in m_Params. -Qualify all math functions with std:: -No longer use our own Clamp() in OpenCL, instead use the standard clamp(). -Redesign how functions are used in the variations OpenCL code. -Add tests to EmberTester to verify some of the new functionality. -Place more const and override qualifiers on functions where appropriate. -Add a global rand with a lock to be used very sparingly. -Use a map instead of a vector for bad param names in Xml parsing. -Prefix affine interpolation mode defines with "AFFINE_" to make their purpose more clear. -Allow for variations that change state during iteration by sending a separate copy of the ember to each rendering thread. -Implement this same functionality with a local struct in OpenCL. It's members are the total of all variables that need to change state within an ember. -Add Contains() function to Utils.h. -EmberRender: print names of kernels being printed with --dump_kernel option. -Clean up EmberTester to handle some of the recent changes. -Fix various casts. -Replace % 2 with & 1, even though the compiler was likely doing this already. -Add new file Variations06.h to accommodate new variations. -General cleanup.
2015-11-22 17:15:07 -05:00
}
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
};
/// <summary>
/// cschq.
/// </summary>
template <typename T>
--Code changes -Split VariationList into .h/cpp files. --Make VariationList.cpp the only place where the variations files are included from. This speeds up the build, but variations can no longer be directly instantiated. --Variations are no longer exported.
2016-02-18 21:58:24 -05:00
class CschqVariation : public Variation<T>
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
public:
--Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums.
2016-01-04 19:50:15 -05:00
CschqVariation(T weight = 1.0) : Variation<T>("cschq", eVariationId::VAR_CSCHQ, weight, true) { }
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
VARCOPY(CschqVariation)
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
--Code changes -Make usage of VarFuncs more consistent by moving some global functions there. -Make CircleTrans1 a little safer by calling Zeps() on m_Sc during precalc. -Begin prepping for 0.9.9.5 release.
2016-03-04 21:54:06 -05:00
T absV = VarFuncs<T>::Hypot(helper.In.y, helper.In.z);
06/09/2017 --User changes -dark.qss is now per-OS. -Properly set/reload palette when coming from the palette editor. The latter must be done if they've modified the current palette even if they've clicked cancel. --Bug fixes -Make the following variations safer by using Zeps(): conic, bipolar, edisc, whorl, tan, csc, cot, tanh, sech, csch, coth, auger, bwraps, hypertile3d, hypertile3d1, ortho, poincare, rational3, barycentroid, sschecks, cscq, cschq, scry_3D, splitbrdr, hexcrop, nblur, crob. -Fix bug enabling/disabling overwrite button in palette editor. -Small optimization for gdoffs, use precalcAtanYX. -Properly propagate z through circlesplit, cylinder2 and tile_log variations. -Some values in truchet_fill could've been NaN. --Code changes -Make most installation files read only. -Qualify many calls with std:: to ensure they're not colliding with glm:: -Use auto in more places.
2017-06-09 22:38:06 -04:00
T ni = m_Weight / Zeps(helper.m_PrecalcSumSquares + SQR(helper.In.z));
--User changes -Add new variations: bubbleT3D, crob, hexaplay3D, hexcrop, hexes, hexnix3D, loonie2, loonie3, nBlur, octapol and synth. -Allow for pre/post versions of dc_bubble, dc_cylinder and dc_linear whereas before they were omitted. -When saving a file with multiple embers in it, detect if time values are all the same and if so, start them at zero and increment by 1 for each ember. -Allow for numerous quality increases to be coalesced into one. It will pick up at the end of the current render. -Show selection highlight on variations tree in response to mouse hover. This makes it easier to see for which variation or param the current mouse wheel action will apply. -Make default temporal samples be 100, whereas before it was 1000 which was overkill. -Require the shift key to be held with delete for deleting an ember to prevent it from triggering when the user enters delete in the edit box. -This wasn't otherwise fixable without writing a lot more code. --Bug fixes -EmberGenome was crashing when generating a sequence from a source file with more than 2 embers in it. -EmberGenome was improperly handling the first frame of a merge after the last frame of the loop. -These bugs were due to a previous commit. Revert parts of that commit. -Prevent a zoom value of less than 0 when reading from xml. -Slight optimization of the crescents, and mask variations, if the compiler wasn't doing it already. -Unique file naming was broken because it was looking for _# and the default names ended with -#. -Disallow renaming of an ember in the library tree to an empty string. -Severe bug that prevented some variations from being read correctly from params generated outside this program. -Severe OpenCL randomization bug. The first x coordinates of the first points in the first kernel call of the first ember of a render since the OpenCL renderer object was created were not random and were mostly -1. -Severe bug when populating xform selection distributions that could sometimes cause a crash due to roundoff error. Fix by using double. -Limit the max number of variations in a random ember to MAX_CL_VARS, which is 8. This ensures they'll look the same on CPU and GPU. -Prevent user from saving stylesheet to default.qss, it's a special reserved filename. --Code changes -Generalize using the running sum output point inside of a variation for all cases: pre, reg and post. -Allow for array variables in variations where the address of each element is stored in m_Params. -Qualify all math functions with std:: -No longer use our own Clamp() in OpenCL, instead use the standard clamp(). -Redesign how functions are used in the variations OpenCL code. -Add tests to EmberTester to verify some of the new functionality. -Place more const and override qualifiers on functions where appropriate. -Add a global rand with a lock to be used very sparingly. -Use a map instead of a vector for bad param names in Xml parsing. -Prefix affine interpolation mode defines with "AFFINE_" to make their purpose more clear. -Allow for variations that change state during iteration by sending a separate copy of the ember to each rendering thread. -Implement this same functionality with a local struct in OpenCL. It's members are the total of all variables that need to change state within an ember. -Add Contains() function to Utils.h. -EmberRender: print names of kernels being printed with --dump_kernel option. -Clean up EmberTester to handle some of the recent changes. -Fix various casts. -Replace % 2 with & 1, even though the compiler was likely doing this already. -Add new file Variations06.h to accommodate new variations. -General cleanup.
2015-11-22 17:15:07 -05:00
T s = std::sin(absV);
T c = std::cos(absV);
T sh = std::sinh(helper.In.x);
T ch = std::cosh(helper.In.x);
06/04/2017 --User changes -Make all fonts be MS Shell Dlg 2. This will require reloading dark.qss if users are already using it. -Limit size of the left side of the palette editor. -Disable create from image buttons in the palette editor when working on a fixed palette. --Bug fixes -The following variations were wrong: coshq, cothq. -During iteration, the color index could become nan if all xform color speeds were negative. This could lead to bad results on the GPU. Fix to check for nan. Minimal speed difference. --Code changes -Make the following variations safer by using Zeps(): sinq, sinhq, secq, sechq, tanq, tanhq, cosq, coshq, cotq, cothq, cscq, cschq, estiq. -Always pass -cl-no-signed-zeros -cl-denorms-are-zero to kernel compiles for both single and double. -Flush all denormals to zero for all executable programs. This will likely lead to a speedup for badly behaving programs.
2017-06-04 20:37:29 -04:00
T d = ni * ch * s / Zeps(absV);
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
helper.Out.x = sh * c * ni;
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
helper.Out.y = -(d * helper.In.y);
helper.Out.z = -(d * helper.In.z);
}
--Code changes -Better use of const functions. -Ember.pro should not define BUILDING_EMBERCL, instead define BUILDING_EMBER.
2015-06-29 23:13:53 -04:00
virtual string OpenCLString() const override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
ostringstream ss;
0.4.1.3 Beta 10/14/2014 --User Changes Size is no longer fixed to the window size. Size scaling is done differently in the final render dialog. This fixes several bugs. Remove Xml saving size from settings and options dialog, it no longer applies. Final render can be broken into strips. Set default save path to the desktop if none is found in the settings file. Set default output size to 1920x1080 if none is found in the settings file. --Bug Fixes Better memory size reporting in final render dialog. --Code Changes Migrate to C++11, Qt 5.3.1, and Visual Studio 2013. Change most instances of unsigned int to size_t, and int to intmax_t. Add m_OrigPixPerUnit and m_ScaleType to Ember for scaling purposes. Replace some sprintf_s() calls in XmlToEmber with ostringstream. Move more non-templated members into RendererBase. Add CopyVec() overload that takes a per element function pointer. Add vector Memset(). Replace '&' with '+' instead of "&amp;" in XmlToEmber for much faster parsing. Break strips rendering out into EmberCommon and call from EmberRender and Fractorium. Make AddAndWriteBuffer() just call WriteBuffer(). Make AddAndWriteImage() delete the existing image first before replacing it. Add SetOutputTexture() to RendererCL to support making new textures in response to resize events. Remove multiple return statements in RendererCL, and replace with a bool that tracks results. Add ToDouble(), MakeEnd(), ToString() and Exists() wrappers in Fractorium. Add Size() wrapper in EmberFile. Make QString function arguments const QString&, and string with const string&. Make ShowCritical() wrapper for invoking a message box from another thread. Add combo box to TwoButtonWidget and rename.
2014-10-14 11:53:15 -04:00
intmax_t varIndex = IndexInXform();
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
string weight = WeightDefineString();
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
ss << "\t{\n"
<< "\t\treal_t absV = Hypot(vIn.y, vIn.z);\n"
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
<< "\t\treal_t ni = " << weight << " / Zeps(fma(vIn.z, vIn.z, precalcSumSquares));\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t\treal_t s = sin(absV);\n"
<< "\t\treal_t c = cos(absV);\n"
<< "\t\treal_t sh = sinh(vIn.x);\n"
<< "\t\treal_t ch = cosh(vIn.x);\n"
06/04/2017 --User changes -Make all fonts be MS Shell Dlg 2. This will require reloading dark.qss if users are already using it. -Limit size of the left side of the palette editor. -Disable create from image buttons in the palette editor when working on a fixed palette. --Bug fixes -The following variations were wrong: coshq, cothq. -During iteration, the color index could become nan if all xform color speeds were negative. This could lead to bad results on the GPU. Fix to check for nan. Minimal speed difference. --Code changes -Make the following variations safer by using Zeps(): sinq, sinhq, secq, sechq, tanq, tanhq, cosq, coshq, cotq, cothq, cscq, cschq, estiq. -Always pass -cl-no-signed-zeros -cl-denorms-are-zero to kernel compiles for both single and double. -Flush all denormals to zero for all executable programs. This will likely lead to a speedup for badly behaving programs.
2017-06-04 20:37:29 -04:00
<< "\t\treal_t d = ni * ch * s / Zeps(absV);\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\n"
<< "\t\tvOut.x = sh * c * ni;\n"
<< "\t\tvOut.y = -(d * vIn.y);\n"
<< "\t\tvOut.z = -(d * vIn.z);\n"
<< "\t}\n";
return ss.str();
}
--User changes -Add new variations: bubbleT3D, crob, hexaplay3D, hexcrop, hexes, hexnix3D, loonie2, loonie3, nBlur, octapol and synth. -Allow for pre/post versions of dc_bubble, dc_cylinder and dc_linear whereas before they were omitted. -When saving a file with multiple embers in it, detect if time values are all the same and if so, start them at zero and increment by 1 for each ember. -Allow for numerous quality increases to be coalesced into one. It will pick up at the end of the current render. -Show selection highlight on variations tree in response to mouse hover. This makes it easier to see for which variation or param the current mouse wheel action will apply. -Make default temporal samples be 100, whereas before it was 1000 which was overkill. -Require the shift key to be held with delete for deleting an ember to prevent it from triggering when the user enters delete in the edit box. -This wasn't otherwise fixable without writing a lot more code. --Bug fixes -EmberGenome was crashing when generating a sequence from a source file with more than 2 embers in it. -EmberGenome was improperly handling the first frame of a merge after the last frame of the loop. -These bugs were due to a previous commit. Revert parts of that commit. -Prevent a zoom value of less than 0 when reading from xml. -Slight optimization of the crescents, and mask variations, if the compiler wasn't doing it already. -Unique file naming was broken because it was looking for _# and the default names ended with -#. -Disallow renaming of an ember in the library tree to an empty string. -Severe bug that prevented some variations from being read correctly from params generated outside this program. -Severe OpenCL randomization bug. The first x coordinates of the first points in the first kernel call of the first ember of a render since the OpenCL renderer object was created were not random and were mostly -1. -Severe bug when populating xform selection distributions that could sometimes cause a crash due to roundoff error. Fix by using double. -Limit the max number of variations in a random ember to MAX_CL_VARS, which is 8. This ensures they'll look the same on CPU and GPU. -Prevent user from saving stylesheet to default.qss, it's a special reserved filename. --Code changes -Generalize using the running sum output point inside of a variation for all cases: pre, reg and post. -Allow for array variables in variations where the address of each element is stored in m_Params. -Qualify all math functions with std:: -No longer use our own Clamp() in OpenCL, instead use the standard clamp(). -Redesign how functions are used in the variations OpenCL code. -Add tests to EmberTester to verify some of the new functionality. -Place more const and override qualifiers on functions where appropriate. -Add a global rand with a lock to be used very sparingly. -Use a map instead of a vector for bad param names in Xml parsing. -Prefix affine interpolation mode defines with "AFFINE_" to make their purpose more clear. -Allow for variations that change state during iteration by sending a separate copy of the ember to each rendering thread. -Implement this same functionality with a local struct in OpenCL. It's members are the total of all variables that need to change state within an ember. -Add Contains() function to Utils.h. -EmberRender: print names of kernels being printed with --dump_kernel option. -Clean up EmberTester to handle some of the recent changes. -Fix various casts. -Replace % 2 with & 1, even though the compiler was likely doing this already. -Add new file Variations06.h to accommodate new variations. -General cleanup.
2015-11-22 17:15:07 -05:00
virtual vector<string> OpenCLGlobalFuncNames() const override
{
06/04/2017 --User changes -Make all fonts be MS Shell Dlg 2. This will require reloading dark.qss if users are already using it. -Limit size of the left side of the palette editor. -Disable create from image buttons in the palette editor when working on a fixed palette. --Bug fixes -The following variations were wrong: coshq, cothq. -During iteration, the color index could become nan if all xform color speeds were negative. This could lead to bad results on the GPU. Fix to check for nan. Minimal speed difference. --Code changes -Make the following variations safer by using Zeps(): sinq, sinhq, secq, sechq, tanq, tanhq, cosq, coshq, cotq, cothq, cscq, cschq, estiq. -Always pass -cl-no-signed-zeros -cl-denorms-are-zero to kernel compiles for both single and double. -Flush all denormals to zero for all executable programs. This will likely lead to a speedup for badly behaving programs.
2017-06-04 20:37:29 -04:00
return vector<string> { "Hypot", "Zeps" };
--User changes -Add new variations: bubbleT3D, crob, hexaplay3D, hexcrop, hexes, hexnix3D, loonie2, loonie3, nBlur, octapol and synth. -Allow for pre/post versions of dc_bubble, dc_cylinder and dc_linear whereas before they were omitted. -When saving a file with multiple embers in it, detect if time values are all the same and if so, start them at zero and increment by 1 for each ember. -Allow for numerous quality increases to be coalesced into one. It will pick up at the end of the current render. -Show selection highlight on variations tree in response to mouse hover. This makes it easier to see for which variation or param the current mouse wheel action will apply. -Make default temporal samples be 100, whereas before it was 1000 which was overkill. -Require the shift key to be held with delete for deleting an ember to prevent it from triggering when the user enters delete in the edit box. -This wasn't otherwise fixable without writing a lot more code. --Bug fixes -EmberGenome was crashing when generating a sequence from a source file with more than 2 embers in it. -EmberGenome was improperly handling the first frame of a merge after the last frame of the loop. -These bugs were due to a previous commit. Revert parts of that commit. -Prevent a zoom value of less than 0 when reading from xml. -Slight optimization of the crescents, and mask variations, if the compiler wasn't doing it already. -Unique file naming was broken because it was looking for _# and the default names ended with -#. -Disallow renaming of an ember in the library tree to an empty string. -Severe bug that prevented some variations from being read correctly from params generated outside this program. -Severe OpenCL randomization bug. The first x coordinates of the first points in the first kernel call of the first ember of a render since the OpenCL renderer object was created were not random and were mostly -1. -Severe bug when populating xform selection distributions that could sometimes cause a crash due to roundoff error. Fix by using double. -Limit the max number of variations in a random ember to MAX_CL_VARS, which is 8. This ensures they'll look the same on CPU and GPU. -Prevent user from saving stylesheet to default.qss, it's a special reserved filename. --Code changes -Generalize using the running sum output point inside of a variation for all cases: pre, reg and post. -Allow for array variables in variations where the address of each element is stored in m_Params. -Qualify all math functions with std:: -No longer use our own Clamp() in OpenCL, instead use the standard clamp(). -Redesign how functions are used in the variations OpenCL code. -Add tests to EmberTester to verify some of the new functionality. -Place more const and override qualifiers on functions where appropriate. -Add a global rand with a lock to be used very sparingly. -Use a map instead of a vector for bad param names in Xml parsing. -Prefix affine interpolation mode defines with "AFFINE_" to make their purpose more clear. -Allow for variations that change state during iteration by sending a separate copy of the ember to each rendering thread. -Implement this same functionality with a local struct in OpenCL. It's members are the total of all variables that need to change state within an ember. -Add Contains() function to Utils.h. -EmberRender: print names of kernels being printed with --dump_kernel option. -Clean up EmberTester to handle some of the recent changes. -Fix various casts. -Replace % 2 with & 1, even though the compiler was likely doing this already. -Add new file Variations06.h to accommodate new variations. -General cleanup.
2015-11-22 17:15:07 -05:00
}
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
};
/// <summary>
/// estiq.
/// </summary>
template <typename T>
--Code changes -Split VariationList into .h/cpp files. --Make VariationList.cpp the only place where the variations files are included from. This speeds up the build, but variations can no longer be directly instantiated. --Variations are no longer exported.
2016-02-18 21:58:24 -05:00
class EstiqVariation : public Variation<T>
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
public:
--Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums.
2016-01-04 19:50:15 -05:00
EstiqVariation(T weight = 1.0) : Variation<T>("estiq", eVariationId::VAR_ESTIQ, weight) { }
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
VARCOPY(EstiqVariation)
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
--Code changes -Make usage of VarFuncs more consistent by moving some global functions there. -Make CircleTrans1 a little safer by calling Zeps() on m_Sc during precalc. -Begin prepping for 0.9.9.5 release.
2016-03-04 21:54:06 -05:00
T absV = VarFuncs<T>::Hypot(helper.In.y, helper.In.z);
--User changes -Add new variations: bubbleT3D, crob, hexaplay3D, hexcrop, hexes, hexnix3D, loonie2, loonie3, nBlur, octapol and synth. -Allow for pre/post versions of dc_bubble, dc_cylinder and dc_linear whereas before they were omitted. -When saving a file with multiple embers in it, detect if time values are all the same and if so, start them at zero and increment by 1 for each ember. -Allow for numerous quality increases to be coalesced into one. It will pick up at the end of the current render. -Show selection highlight on variations tree in response to mouse hover. This makes it easier to see for which variation or param the current mouse wheel action will apply. -Make default temporal samples be 100, whereas before it was 1000 which was overkill. -Require the shift key to be held with delete for deleting an ember to prevent it from triggering when the user enters delete in the edit box. -This wasn't otherwise fixable without writing a lot more code. --Bug fixes -EmberGenome was crashing when generating a sequence from a source file with more than 2 embers in it. -EmberGenome was improperly handling the first frame of a merge after the last frame of the loop. -These bugs were due to a previous commit. Revert parts of that commit. -Prevent a zoom value of less than 0 when reading from xml. -Slight optimization of the crescents, and mask variations, if the compiler wasn't doing it already. -Unique file naming was broken because it was looking for _# and the default names ended with -#. -Disallow renaming of an ember in the library tree to an empty string. -Severe bug that prevented some variations from being read correctly from params generated outside this program. -Severe OpenCL randomization bug. The first x coordinates of the first points in the first kernel call of the first ember of a render since the OpenCL renderer object was created were not random and were mostly -1. -Severe bug when populating xform selection distributions that could sometimes cause a crash due to roundoff error. Fix by using double. -Limit the max number of variations in a random ember to MAX_CL_VARS, which is 8. This ensures they'll look the same on CPU and GPU. -Prevent user from saving stylesheet to default.qss, it's a special reserved filename. --Code changes -Generalize using the running sum output point inside of a variation for all cases: pre, reg and post. -Allow for array variables in variations where the address of each element is stored in m_Params. -Qualify all math functions with std:: -No longer use our own Clamp() in OpenCL, instead use the standard clamp(). -Redesign how functions are used in the variations OpenCL code. -Add tests to EmberTester to verify some of the new functionality. -Place more const and override qualifiers on functions where appropriate. -Add a global rand with a lock to be used very sparingly. -Use a map instead of a vector for bad param names in Xml parsing. -Prefix affine interpolation mode defines with "AFFINE_" to make their purpose more clear. -Allow for variations that change state during iteration by sending a separate copy of the ember to each rendering thread. -Implement this same functionality with a local struct in OpenCL. It's members are the total of all variables that need to change state within an ember. -Add Contains() function to Utils.h. -EmberRender: print names of kernels being printed with --dump_kernel option. -Clean up EmberTester to handle some of the recent changes. -Fix various casts. -Replace % 2 with & 1, even though the compiler was likely doing this already. -Add new file Variations06.h to accommodate new variations. -General cleanup.
2015-11-22 17:15:07 -05:00
T e = std::exp(helper.In.x);
T s = std::sin(absV);
T c = std::cos(absV);
06/04/2017 --User changes -Make all fonts be MS Shell Dlg 2. This will require reloading dark.qss if users are already using it. -Limit size of the left side of the palette editor. -Disable create from image buttons in the palette editor when working on a fixed palette. --Bug fixes -The following variations were wrong: coshq, cothq. -During iteration, the color index could become nan if all xform color speeds were negative. This could lead to bad results on the GPU. Fix to check for nan. Minimal speed difference. --Code changes -Make the following variations safer by using Zeps(): sinq, sinhq, secq, sechq, tanq, tanhq, cosq, coshq, cotq, cothq, cscq, cschq, estiq. -Always pass -cl-no-signed-zeros -cl-denorms-are-zero to kernel compiles for both single and double. -Flush all denormals to zero for all executable programs. This will likely lead to a speedup for badly behaving programs.
2017-06-04 20:37:29 -04:00
T a = e * s / Zeps(absV);
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
helper.Out.x = m_Weight * e * c;
helper.Out.y = m_Weight * a * helper.In.y;
helper.Out.z = m_Weight * a * helper.In.z;
}
--Code changes -Better use of const functions. -Ember.pro should not define BUILDING_EMBERCL, instead define BUILDING_EMBER.
2015-06-29 23:13:53 -04:00
virtual string OpenCLString() const override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
ostringstream ss;
0.4.1.3 Beta 10/14/2014 --User Changes Size is no longer fixed to the window size. Size scaling is done differently in the final render dialog. This fixes several bugs. Remove Xml saving size from settings and options dialog, it no longer applies. Final render can be broken into strips. Set default save path to the desktop if none is found in the settings file. Set default output size to 1920x1080 if none is found in the settings file. --Bug Fixes Better memory size reporting in final render dialog. --Code Changes Migrate to C++11, Qt 5.3.1, and Visual Studio 2013. Change most instances of unsigned int to size_t, and int to intmax_t. Add m_OrigPixPerUnit and m_ScaleType to Ember for scaling purposes. Replace some sprintf_s() calls in XmlToEmber with ostringstream. Move more non-templated members into RendererBase. Add CopyVec() overload that takes a per element function pointer. Add vector Memset(). Replace '&' with '+' instead of "&amp;" in XmlToEmber for much faster parsing. Break strips rendering out into EmberCommon and call from EmberRender and Fractorium. Make AddAndWriteBuffer() just call WriteBuffer(). Make AddAndWriteImage() delete the existing image first before replacing it. Add SetOutputTexture() to RendererCL to support making new textures in response to resize events. Remove multiple return statements in RendererCL, and replace with a bool that tracks results. Add ToDouble(), MakeEnd(), ToString() and Exists() wrappers in Fractorium. Add Size() wrapper in EmberFile. Make QString function arguments const QString&, and string with const string&. Make ShowCritical() wrapper for invoking a message box from another thread. Add combo box to TwoButtonWidget and rename.
2014-10-14 11:53:15 -04:00
intmax_t varIndex = IndexInXform();
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
string weight = WeightDefineString();
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
ss << "\t{\n"
<< "\t\treal_t absV = Hypot(vIn.y, vIn.z);\n"
<< "\t\treal_t e = exp(vIn.x);\n"
<< "\t\treal_t s = sin(absV);\n"
<< "\t\treal_t c = cos(absV);\n"
06/04/2017 --User changes -Make all fonts be MS Shell Dlg 2. This will require reloading dark.qss if users are already using it. -Limit size of the left side of the palette editor. -Disable create from image buttons in the palette editor when working on a fixed palette. --Bug fixes -The following variations were wrong: coshq, cothq. -During iteration, the color index could become nan if all xform color speeds were negative. This could lead to bad results on the GPU. Fix to check for nan. Minimal speed difference. --Code changes -Make the following variations safer by using Zeps(): sinq, sinhq, secq, sechq, tanq, tanhq, cosq, coshq, cotq, cothq, cscq, cschq, estiq. -Always pass -cl-no-signed-zeros -cl-denorms-are-zero to kernel compiles for both single and double. -Flush all denormals to zero for all executable programs. This will likely lead to a speedup for badly behaving programs.
2017-06-04 20:37:29 -04:00
<< "\t\treal_t a = e * s / Zeps(absV);\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\n"
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
<< "\t\tvOut.x = " << weight << " * e * c;\n"
<< "\t\tvOut.y = " << weight << " * a * vIn.y;\n"
<< "\t\tvOut.z = " << weight << " * a * vIn.z;\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t}\n";
return ss.str();
}
--User changes -Add new variations: bubbleT3D, crob, hexaplay3D, hexcrop, hexes, hexnix3D, loonie2, loonie3, nBlur, octapol and synth. -Allow for pre/post versions of dc_bubble, dc_cylinder and dc_linear whereas before they were omitted. -When saving a file with multiple embers in it, detect if time values are all the same and if so, start them at zero and increment by 1 for each ember. -Allow for numerous quality increases to be coalesced into one. It will pick up at the end of the current render. -Show selection highlight on variations tree in response to mouse hover. This makes it easier to see for which variation or param the current mouse wheel action will apply. -Make default temporal samples be 100, whereas before it was 1000 which was overkill. -Require the shift key to be held with delete for deleting an ember to prevent it from triggering when the user enters delete in the edit box. -This wasn't otherwise fixable without writing a lot more code. --Bug fixes -EmberGenome was crashing when generating a sequence from a source file with more than 2 embers in it. -EmberGenome was improperly handling the first frame of a merge after the last frame of the loop. -These bugs were due to a previous commit. Revert parts of that commit. -Prevent a zoom value of less than 0 when reading from xml. -Slight optimization of the crescents, and mask variations, if the compiler wasn't doing it already. -Unique file naming was broken because it was looking for _# and the default names ended with -#. -Disallow renaming of an ember in the library tree to an empty string. -Severe bug that prevented some variations from being read correctly from params generated outside this program. -Severe OpenCL randomization bug. The first x coordinates of the first points in the first kernel call of the first ember of a render since the OpenCL renderer object was created were not random and were mostly -1. -Severe bug when populating xform selection distributions that could sometimes cause a crash due to roundoff error. Fix by using double. -Limit the max number of variations in a random ember to MAX_CL_VARS, which is 8. This ensures they'll look the same on CPU and GPU. -Prevent user from saving stylesheet to default.qss, it's a special reserved filename. --Code changes -Generalize using the running sum output point inside of a variation for all cases: pre, reg and post. -Allow for array variables in variations where the address of each element is stored in m_Params. -Qualify all math functions with std:: -No longer use our own Clamp() in OpenCL, instead use the standard clamp(). -Redesign how functions are used in the variations OpenCL code. -Add tests to EmberTester to verify some of the new functionality. -Place more const and override qualifiers on functions where appropriate. -Add a global rand with a lock to be used very sparingly. -Use a map instead of a vector for bad param names in Xml parsing. -Prefix affine interpolation mode defines with "AFFINE_" to make their purpose more clear. -Allow for variations that change state during iteration by sending a separate copy of the ember to each rendering thread. -Implement this same functionality with a local struct in OpenCL. It's members are the total of all variables that need to change state within an ember. -Add Contains() function to Utils.h. -EmberRender: print names of kernels being printed with --dump_kernel option. -Clean up EmberTester to handle some of the recent changes. -Fix various casts. -Replace % 2 with & 1, even though the compiler was likely doing this already. -Add new file Variations06.h to accommodate new variations. -General cleanup.
2015-11-22 17:15:07 -05:00
virtual vector<string> OpenCLGlobalFuncNames() const override
{
06/04/2017 --User changes -Make all fonts be MS Shell Dlg 2. This will require reloading dark.qss if users are already using it. -Limit size of the left side of the palette editor. -Disable create from image buttons in the palette editor when working on a fixed palette. --Bug fixes -The following variations were wrong: coshq, cothq. -During iteration, the color index could become nan if all xform color speeds were negative. This could lead to bad results on the GPU. Fix to check for nan. Minimal speed difference. --Code changes -Make the following variations safer by using Zeps(): sinq, sinhq, secq, sechq, tanq, tanhq, cosq, coshq, cotq, cothq, cscq, cschq, estiq. -Always pass -cl-no-signed-zeros -cl-denorms-are-zero to kernel compiles for both single and double. -Flush all denormals to zero for all executable programs. This will likely lead to a speedup for badly behaving programs.
2017-06-04 20:37:29 -04:00
return vector<string> { "Hypot", "Zeps" };
--User changes -Add new variations: bubbleT3D, crob, hexaplay3D, hexcrop, hexes, hexnix3D, loonie2, loonie3, nBlur, octapol and synth. -Allow for pre/post versions of dc_bubble, dc_cylinder and dc_linear whereas before they were omitted. -When saving a file with multiple embers in it, detect if time values are all the same and if so, start them at zero and increment by 1 for each ember. -Allow for numerous quality increases to be coalesced into one. It will pick up at the end of the current render. -Show selection highlight on variations tree in response to mouse hover. This makes it easier to see for which variation or param the current mouse wheel action will apply. -Make default temporal samples be 100, whereas before it was 1000 which was overkill. -Require the shift key to be held with delete for deleting an ember to prevent it from triggering when the user enters delete in the edit box. -This wasn't otherwise fixable without writing a lot more code. --Bug fixes -EmberGenome was crashing when generating a sequence from a source file with more than 2 embers in it. -EmberGenome was improperly handling the first frame of a merge after the last frame of the loop. -These bugs were due to a previous commit. Revert parts of that commit. -Prevent a zoom value of less than 0 when reading from xml. -Slight optimization of the crescents, and mask variations, if the compiler wasn't doing it already. -Unique file naming was broken because it was looking for _# and the default names ended with -#. -Disallow renaming of an ember in the library tree to an empty string. -Severe bug that prevented some variations from being read correctly from params generated outside this program. -Severe OpenCL randomization bug. The first x coordinates of the first points in the first kernel call of the first ember of a render since the OpenCL renderer object was created were not random and were mostly -1. -Severe bug when populating xform selection distributions that could sometimes cause a crash due to roundoff error. Fix by using double. -Limit the max number of variations in a random ember to MAX_CL_VARS, which is 8. This ensures they'll look the same on CPU and GPU. -Prevent user from saving stylesheet to default.qss, it's a special reserved filename. --Code changes -Generalize using the running sum output point inside of a variation for all cases: pre, reg and post. -Allow for array variables in variations where the address of each element is stored in m_Params. -Qualify all math functions with std:: -No longer use our own Clamp() in OpenCL, instead use the standard clamp(). -Redesign how functions are used in the variations OpenCL code. -Add tests to EmberTester to verify some of the new functionality. -Place more const and override qualifiers on functions where appropriate. -Add a global rand with a lock to be used very sparingly. -Use a map instead of a vector for bad param names in Xml parsing. -Prefix affine interpolation mode defines with "AFFINE_" to make their purpose more clear. -Allow for variations that change state during iteration by sending a separate copy of the ember to each rendering thread. -Implement this same functionality with a local struct in OpenCL. It's members are the total of all variables that need to change state within an ember. -Add Contains() function to Utils.h. -EmberRender: print names of kernels being printed with --dump_kernel option. -Clean up EmberTester to handle some of the recent changes. -Fix various casts. -Replace % 2 with & 1, even though the compiler was likely doing this already. -Add new file Variations06.h to accommodate new variations. -General cleanup.
2015-11-22 17:15:07 -05:00
}
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
};
/// <summary>
/// loq.
/// </summary>
template <typename T>
--Code changes -Split VariationList into .h/cpp files. --Make VariationList.cpp the only place where the variations files are included from. This speeds up the build, but variations can no longer be directly instantiated. --Variations are no longer exported.
2016-02-18 21:58:24 -05:00
class LoqVariation : public ParametricVariation<T>
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
public:
--Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums.
2016-01-04 19:50:15 -05:00
LoqVariation(T weight = 1.0) : ParametricVariation<T>("loq", eVariationId::VAR_LOQ, weight)
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
Init();
}
PARVARCOPY(LoqVariation)
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
--Code changes -Make usage of VarFuncs more consistent by moving some global functions there. -Make CircleTrans1 a little safer by calling Zeps() on m_Sc during precalc. -Begin prepping for 0.9.9.5 release.
2016-03-04 21:54:06 -05:00
T absV = VarFuncs<T>::Hypot(helper.In.y, helper.In.z);
06/04/2017 --User changes -Make all fonts be MS Shell Dlg 2. This will require reloading dark.qss if users are already using it. -Limit size of the left side of the palette editor. -Disable create from image buttons in the palette editor when working on a fixed palette. --Bug fixes -The following variations were wrong: coshq, cothq. -During iteration, the color index could become nan if all xform color speeds were negative. This could lead to bad results on the GPU. Fix to check for nan. Minimal speed difference. --Code changes -Make the following variations safer by using Zeps(): sinq, sinhq, secq, sechq, tanq, tanhq, cosq, coshq, cotq, cothq, cscq, cschq, estiq. -Always pass -cl-no-signed-zeros -cl-denorms-are-zero to kernel compiles for both single and double. -Flush all denormals to zero for all executable programs. This will likely lead to a speedup for badly behaving programs.
2017-06-04 20:37:29 -04:00
T c = m_Weight * std::atan2(absV, helper.In.x) / Zeps(absV);
--User changes -Add new variations: bubbleT3D, crob, hexaplay3D, hexcrop, hexes, hexnix3D, loonie2, loonie3, nBlur, octapol and synth. -Allow for pre/post versions of dc_bubble, dc_cylinder and dc_linear whereas before they were omitted. -When saving a file with multiple embers in it, detect if time values are all the same and if so, start them at zero and increment by 1 for each ember. -Allow for numerous quality increases to be coalesced into one. It will pick up at the end of the current render. -Show selection highlight on variations tree in response to mouse hover. This makes it easier to see for which variation or param the current mouse wheel action will apply. -Make default temporal samples be 100, whereas before it was 1000 which was overkill. -Require the shift key to be held with delete for deleting an ember to prevent it from triggering when the user enters delete in the edit box. -This wasn't otherwise fixable without writing a lot more code. --Bug fixes -EmberGenome was crashing when generating a sequence from a source file with more than 2 embers in it. -EmberGenome was improperly handling the first frame of a merge after the last frame of the loop. -These bugs were due to a previous commit. Revert parts of that commit. -Prevent a zoom value of less than 0 when reading from xml. -Slight optimization of the crescents, and mask variations, if the compiler wasn't doing it already. -Unique file naming was broken because it was looking for _# and the default names ended with -#. -Disallow renaming of an ember in the library tree to an empty string. -Severe bug that prevented some variations from being read correctly from params generated outside this program. -Severe OpenCL randomization bug. The first x coordinates of the first points in the first kernel call of the first ember of a render since the OpenCL renderer object was created were not random and were mostly -1. -Severe bug when populating xform selection distributions that could sometimes cause a crash due to roundoff error. Fix by using double. -Limit the max number of variations in a random ember to MAX_CL_VARS, which is 8. This ensures they'll look the same on CPU and GPU. -Prevent user from saving stylesheet to default.qss, it's a special reserved filename. --Code changes -Generalize using the running sum output point inside of a variation for all cases: pre, reg and post. -Allow for array variables in variations where the address of each element is stored in m_Params. -Qualify all math functions with std:: -No longer use our own Clamp() in OpenCL, instead use the standard clamp(). -Redesign how functions are used in the variations OpenCL code. -Add tests to EmberTester to verify some of the new functionality. -Place more const and override qualifiers on functions where appropriate. -Add a global rand with a lock to be used very sparingly. -Use a map instead of a vector for bad param names in Xml parsing. -Prefix affine interpolation mode defines with "AFFINE_" to make their purpose more clear. -Allow for variations that change state during iteration by sending a separate copy of the ember to each rendering thread. -Implement this same functionality with a local struct in OpenCL. It's members are the total of all variables that need to change state within an ember. -Add Contains() function to Utils.h. -EmberRender: print names of kernels being printed with --dump_kernel option. -Clean up EmberTester to handle some of the recent changes. -Fix various casts. -Replace % 2 with & 1, even though the compiler was likely doing this already. -Add new file Variations06.h to accommodate new variations. -General cleanup.
2015-11-22 17:15:07 -05:00
helper.Out.x = std::log(SQR(helper.In.x) + SQR(absV)) * m_Denom;
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
helper.Out.y = c * helper.In.y;
helper.Out.z = c * helper.In.z;
}
--Code changes -Better use of const functions. -Ember.pro should not define BUILDING_EMBERCL, instead define BUILDING_EMBER.
2015-06-29 23:13:53 -04:00
virtual string OpenCLString() const override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
ostringstream ss, ss2;
0.4.1.3 Beta 10/14/2014 --User Changes Size is no longer fixed to the window size. Size scaling is done differently in the final render dialog. This fixes several bugs. Remove Xml saving size from settings and options dialog, it no longer applies. Final render can be broken into strips. Set default save path to the desktop if none is found in the settings file. Set default output size to 1920x1080 if none is found in the settings file. --Bug Fixes Better memory size reporting in final render dialog. --Code Changes Migrate to C++11, Qt 5.3.1, and Visual Studio 2013. Change most instances of unsigned int to size_t, and int to intmax_t. Add m_OrigPixPerUnit and m_ScaleType to Ember for scaling purposes. Replace some sprintf_s() calls in XmlToEmber with ostringstream. Move more non-templated members into RendererBase. Add CopyVec() overload that takes a per element function pointer. Add vector Memset(). Replace '&' with '+' instead of "&amp;" in XmlToEmber for much faster parsing. Break strips rendering out into EmberCommon and call from EmberRender and Fractorium. Make AddAndWriteBuffer() just call WriteBuffer(). Make AddAndWriteImage() delete the existing image first before replacing it. Add SetOutputTexture() to RendererCL to support making new textures in response to resize events. Remove multiple return statements in RendererCL, and replace with a bool that tracks results. Add ToDouble(), MakeEnd(), ToString() and Exists() wrappers in Fractorium. Add Size() wrapper in EmberFile. Make QString function arguments const QString&, and string with const string&. Make ShowCritical() wrapper for invoking a message box from another thread. Add combo box to TwoButtonWidget and rename.
2014-10-14 11:53:15 -04:00
intmax_t i = 0, varIndex = IndexInXform();
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
ss2 << "_" << XformIndexInEmber() << "]";
string index = ss2.str();
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
string weight = WeightDefineString();
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
string base = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
string denom = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
ss << "\t{\n"
<< "\t\treal_t absV = Hypot(vIn.y, vIn.z);\n"
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
<< "\t\treal_t c = " << weight << " * atan2(absV, vIn.x) / Zeps(absV);\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\n"
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
<< "\t\tvOut.x = log(fma(vIn.x, vIn.x, SQR(absV))) * " << denom << ";\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t\tvOut.y = c * vIn.y;\n"
<< "\t\tvOut.z = c * vIn.z;\n"
<< "\t}\n";
return ss.str();
}
--User changes -Add post_smartcrop. --Bug fixes -Fix bug in crackle. -Wrong point assignment in hexaplay3D, hexnix3D. -Improper Z assignment in rblur. -Fix inconsistency with original in circlecrop. -Put EMBER_ROOT bakc to ./../../../ in default.pri. This is TBD. --Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums. -Optimize crob, nBlur. -Fix weird assignment statement in falloff3. -Cleanup in VarFuncs::SimplexNoise3D(). -Replace fabs() with std::abs(). -General cleanup.
2016-01-12 23:42:12 -05:00
--User changes -Add new variations: bubbleT3D, crob, hexaplay3D, hexcrop, hexes, hexnix3D, loonie2, loonie3, nBlur, octapol and synth. -Allow for pre/post versions of dc_bubble, dc_cylinder and dc_linear whereas before they were omitted. -When saving a file with multiple embers in it, detect if time values are all the same and if so, start them at zero and increment by 1 for each ember. -Allow for numerous quality increases to be coalesced into one. It will pick up at the end of the current render. -Show selection highlight on variations tree in response to mouse hover. This makes it easier to see for which variation or param the current mouse wheel action will apply. -Make default temporal samples be 100, whereas before it was 1000 which was overkill. -Require the shift key to be held with delete for deleting an ember to prevent it from triggering when the user enters delete in the edit box. -This wasn't otherwise fixable without writing a lot more code. --Bug fixes -EmberGenome was crashing when generating a sequence from a source file with more than 2 embers in it. -EmberGenome was improperly handling the first frame of a merge after the last frame of the loop. -These bugs were due to a previous commit. Revert parts of that commit. -Prevent a zoom value of less than 0 when reading from xml. -Slight optimization of the crescents, and mask variations, if the compiler wasn't doing it already. -Unique file naming was broken because it was looking for _# and the default names ended with -#. -Disallow renaming of an ember in the library tree to an empty string. -Severe bug that prevented some variations from being read correctly from params generated outside this program. -Severe OpenCL randomization bug. The first x coordinates of the first points in the first kernel call of the first ember of a render since the OpenCL renderer object was created were not random and were mostly -1. -Severe bug when populating xform selection distributions that could sometimes cause a crash due to roundoff error. Fix by using double. -Limit the max number of variations in a random ember to MAX_CL_VARS, which is 8. This ensures they'll look the same on CPU and GPU. -Prevent user from saving stylesheet to default.qss, it's a special reserved filename. --Code changes -Generalize using the running sum output point inside of a variation for all cases: pre, reg and post. -Allow for array variables in variations where the address of each element is stored in m_Params. -Qualify all math functions with std:: -No longer use our own Clamp() in OpenCL, instead use the standard clamp(). -Redesign how functions are used in the variations OpenCL code. -Add tests to EmberTester to verify some of the new functionality. -Place more const and override qualifiers on functions where appropriate. -Add a global rand with a lock to be used very sparingly. -Use a map instead of a vector for bad param names in Xml parsing. -Prefix affine interpolation mode defines with "AFFINE_" to make their purpose more clear. -Allow for variations that change state during iteration by sending a separate copy of the ember to each rendering thread. -Implement this same functionality with a local struct in OpenCL. It's members are the total of all variables that need to change state within an ember. -Add Contains() function to Utils.h. -EmberRender: print names of kernels being printed with --dump_kernel option. -Clean up EmberTester to handle some of the recent changes. -Fix various casts. -Replace % 2 with & 1, even though the compiler was likely doing this already. -Add new file Variations06.h to accommodate new variations. -General cleanup.
2015-11-22 17:15:07 -05:00
virtual vector<string> OpenCLGlobalFuncNames() const override
{
06/04/2017 --User changes -Make all fonts be MS Shell Dlg 2. This will require reloading dark.qss if users are already using it. -Limit size of the left side of the palette editor. -Disable create from image buttons in the palette editor when working on a fixed palette. --Bug fixes -The following variations were wrong: coshq, cothq. -During iteration, the color index could become nan if all xform color speeds were negative. This could lead to bad results on the GPU. Fix to check for nan. Minimal speed difference. --Code changes -Make the following variations safer by using Zeps(): sinq, sinhq, secq, sechq, tanq, tanhq, cosq, coshq, cotq, cothq, cscq, cschq, estiq. -Always pass -cl-no-signed-zeros -cl-denorms-are-zero to kernel compiles for both single and double. -Flush all denormals to zero for all executable programs. This will likely lead to a speedup for badly behaving programs.
2017-06-04 20:37:29 -04:00
return vector<string> { "Hypot", "Zeps" };
--User changes -Add new variations: bubbleT3D, crob, hexaplay3D, hexcrop, hexes, hexnix3D, loonie2, loonie3, nBlur, octapol and synth. -Allow for pre/post versions of dc_bubble, dc_cylinder and dc_linear whereas before they were omitted. -When saving a file with multiple embers in it, detect if time values are all the same and if so, start them at zero and increment by 1 for each ember. -Allow for numerous quality increases to be coalesced into one. It will pick up at the end of the current render. -Show selection highlight on variations tree in response to mouse hover. This makes it easier to see for which variation or param the current mouse wheel action will apply. -Make default temporal samples be 100, whereas before it was 1000 which was overkill. -Require the shift key to be held with delete for deleting an ember to prevent it from triggering when the user enters delete in the edit box. -This wasn't otherwise fixable without writing a lot more code. --Bug fixes -EmberGenome was crashing when generating a sequence from a source file with more than 2 embers in it. -EmberGenome was improperly handling the first frame of a merge after the last frame of the loop. -These bugs were due to a previous commit. Revert parts of that commit. -Prevent a zoom value of less than 0 when reading from xml. -Slight optimization of the crescents, and mask variations, if the compiler wasn't doing it already. -Unique file naming was broken because it was looking for _# and the default names ended with -#. -Disallow renaming of an ember in the library tree to an empty string. -Severe bug that prevented some variations from being read correctly from params generated outside this program. -Severe OpenCL randomization bug. The first x coordinates of the first points in the first kernel call of the first ember of a render since the OpenCL renderer object was created were not random and were mostly -1. -Severe bug when populating xform selection distributions that could sometimes cause a crash due to roundoff error. Fix by using double. -Limit the max number of variations in a random ember to MAX_CL_VARS, which is 8. This ensures they'll look the same on CPU and GPU. -Prevent user from saving stylesheet to default.qss, it's a special reserved filename. --Code changes -Generalize using the running sum output point inside of a variation for all cases: pre, reg and post. -Allow for array variables in variations where the address of each element is stored in m_Params. -Qualify all math functions with std:: -No longer use our own Clamp() in OpenCL, instead use the standard clamp(). -Redesign how functions are used in the variations OpenCL code. -Add tests to EmberTester to verify some of the new functionality. -Place more const and override qualifiers on functions where appropriate. -Add a global rand with a lock to be used very sparingly. -Use a map instead of a vector for bad param names in Xml parsing. -Prefix affine interpolation mode defines with "AFFINE_" to make their purpose more clear. -Allow for variations that change state during iteration by sending a separate copy of the ember to each rendering thread. -Implement this same functionality with a local struct in OpenCL. It's members are the total of all variables that need to change state within an ember. -Add Contains() function to Utils.h. -EmberRender: print names of kernels being printed with --dump_kernel option. -Clean up EmberTester to handle some of the recent changes. -Fix various casts. -Replace % 2 with & 1, even though the compiler was likely doing this already. -Add new file Variations06.h to accommodate new variations. -General cleanup.
2015-11-22 17:15:07 -05:00
}
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
virtual void Precalc() override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
--User changes -Add new variations: bubbleT3D, crob, hexaplay3D, hexcrop, hexes, hexnix3D, loonie2, loonie3, nBlur, octapol and synth. -Allow for pre/post versions of dc_bubble, dc_cylinder and dc_linear whereas before they were omitted. -When saving a file with multiple embers in it, detect if time values are all the same and if so, start them at zero and increment by 1 for each ember. -Allow for numerous quality increases to be coalesced into one. It will pick up at the end of the current render. -Show selection highlight on variations tree in response to mouse hover. This makes it easier to see for which variation or param the current mouse wheel action will apply. -Make default temporal samples be 100, whereas before it was 1000 which was overkill. -Require the shift key to be held with delete for deleting an ember to prevent it from triggering when the user enters delete in the edit box. -This wasn't otherwise fixable without writing a lot more code. --Bug fixes -EmberGenome was crashing when generating a sequence from a source file with more than 2 embers in it. -EmberGenome was improperly handling the first frame of a merge after the last frame of the loop. -These bugs were due to a previous commit. Revert parts of that commit. -Prevent a zoom value of less than 0 when reading from xml. -Slight optimization of the crescents, and mask variations, if the compiler wasn't doing it already. -Unique file naming was broken because it was looking for _# and the default names ended with -#. -Disallow renaming of an ember in the library tree to an empty string. -Severe bug that prevented some variations from being read correctly from params generated outside this program. -Severe OpenCL randomization bug. The first x coordinates of the first points in the first kernel call of the first ember of a render since the OpenCL renderer object was created were not random and were mostly -1. -Severe bug when populating xform selection distributions that could sometimes cause a crash due to roundoff error. Fix by using double. -Limit the max number of variations in a random ember to MAX_CL_VARS, which is 8. This ensures they'll look the same on CPU and GPU. -Prevent user from saving stylesheet to default.qss, it's a special reserved filename. --Code changes -Generalize using the running sum output point inside of a variation for all cases: pre, reg and post. -Allow for array variables in variations where the address of each element is stored in m_Params. -Qualify all math functions with std:: -No longer use our own Clamp() in OpenCL, instead use the standard clamp(). -Redesign how functions are used in the variations OpenCL code. -Add tests to EmberTester to verify some of the new functionality. -Place more const and override qualifiers on functions where appropriate. -Add a global rand with a lock to be used very sparingly. -Use a map instead of a vector for bad param names in Xml parsing. -Prefix affine interpolation mode defines with "AFFINE_" to make their purpose more clear. -Allow for variations that change state during iteration by sending a separate copy of the ember to each rendering thread. -Implement this same functionality with a local struct in OpenCL. It's members are the total of all variables that need to change state within an ember. -Add Contains() function to Utils.h. -EmberRender: print names of kernels being printed with --dump_kernel option. -Clean up EmberTester to handle some of the recent changes. -Fix various casts. -Replace % 2 with & 1, even though the compiler was likely doing this already. -Add new file Variations06.h to accommodate new variations. -General cleanup.
2015-11-22 17:15:07 -05:00
m_Denom = T(0.5) / std::log(m_Base);
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
}
protected:
void Init()
{
string prefix = Prefix();
m_Params.clear();
--Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums.
2016-01-04 19:50:15 -05:00
m_Params.push_back(ParamWithName<T>(&m_Base, prefix + "loq_base", T(M_E), eParamType::REAL, EPS, TMAX));
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
m_Params.push_back(ParamWithName<T>(true, &m_Denom, prefix + "loq_denom"));//Precalc.
}
private:
T m_Base;
T m_Denom;//Precalc.
};
/// <summary>
/// curvature.
/// </summary>
template <typename T>
--Code changes -Split VariationList into .h/cpp files. --Make VariationList.cpp the only place where the variations files are included from. This speeds up the build, but variations can no longer be directly instantiated. --Variations are no longer exported.
2016-02-18 21:58:24 -05:00
class CurvatureVariation : public Variation<T>
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
public:
--Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums.
2016-01-04 19:50:15 -05:00
CurvatureVariation(T weight = 1.0) : Variation<T>("curvature", eVariationId::VAR_CURVATURE, weight, true, true, false, false, true) { }
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
VARCOPY(CurvatureVariation)
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
Numerous fixes 0.4.0.5 Beta 07/18/2014 --User Changes Allow for vibrancy values > 1. Add flatten and unflatten menu items. Automatically flatten like Apophysis does. Add plugin and new_linear tags to Xml to be compatible with Apophysis. --Bug Fixes Fix blur, blur3d, bubble, cropn, cross, curl, curl3d, epispiral, ho, julia3d, julia3dz, loonie, mirror_x, mirror_y, mirror_z, rotate_x, sinusoidal, spherical, spherical3d, stripes. Unique filename on final render was completely broken. Two severe OpenCL bugs. Random seeds were biased and fusing was being reset too often leading to results that differ from the CPU. Subtle, but sometimes severe bug in the setup of the xaos weights. Use properly defined epsilon by getting the value from std::numeric_limits, rather than hard coding 1e-6 or 1e-10. Omit incorrect usage of epsilon everywhere. It should not be automatically added to denominators. Rather, it should only be used if the denominator is zero. Force final render progress bars to 100 on completion. Sometimes they didn't seem to make it there. Make variation name and params comparisons be case insensitive. --Code Changes Make ForEach and FindIf wrappers around std::for_each and std::find_if.
2014-07-19 02:33:18 -04:00
helper.Out.x = m_Weight / Zeps(helper.m_PrecalcSqrtSumSquares);
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
helper.Out.y = helper.m_PrecalcAtanyx;
--User changes -Allow for pausing the renderer in the main window. This makes is more efficient when entering many parameters, such as when following a tutorial. -Add support for new variations: erf, gamma, jac_cn, jac_dn, jac_sn, logDB, pressure_wave, pRose3D, splits3D, w, waves2b, x, xerf, y, z. -Inform user of the start and stop of file parsing in EmberAnimate because the files could potentially be very large. -Move the follwing fields to a new table called Animation: Interpolation, Affine Interpolation, Temporal Samples, Temporal Filter Width, Temporal Filter Type. -These currently have no effect on the interactive renderer and instead are used when running flames through EmberGenome to generate sequences, and then animating them in Fractorium or EmberAnimate. -Add new parameter overrides for EmberRender and EmberAnimate which directly assign values to all flames being rendered, rather than scale: --quality --demin --demax --Bug fixes -Left pad instead of right pad names of sequence outputs from EmberGenome. -Unique file naming was broken for files which already had an underscore in them. -Properly report that png is the default format of EmberRender and EmberAnimate output instead of erroneously claiming it was jpg. -Make command line programs search these folders in this order for the palette file: ./ ~/.fractorium ~/.config/fractorium /usr/share/fractorium /usr/local/share/fractorium -Fix possible bad values in hexes. -Better assignment of Z variables. -Fix boarders due to use of poorly implemented rint() function from flam3. Use std::rint() now. -wedge_sph was completely wrong due to having accidentally swapped the mapping of two parameters. -Make juliascope juliascope_power parameter be of type REAL_NONZERO since it's used as a denominator. -Make Z assignment compatible with the originals in: -arch, bcircle, bCollide, bent, bent2, bisplit, blob, blur_linear, blur_square, bMod, boarders, boarders2, bSwirl, bTransform, butterfly, cardioid, cell, circleblur, circlize, circlize2, circus, collideoscope, cos, cosine, cosh, coth, cpow, cpow2, crescents, cropn, csc, csch, curl, curve, dc_gridout, deltaa, diamond, disc2, eclipse, eCollide, edisc, eJulia, elliptic, eMod, eMotion, ennepers, epispiral, ePush, eRotate, eScale, eSwirl, ex, exp, expo, exponential, fan, fdisc, fibonacci, fibonacci2, fisheye, flipcircle, flipy, flower, flux, funnel, glynnia, GlynnSim1, GlynnSim2, GlynnSim3, gridout, handkerchief, heart, hole, idisc, julia, julian2, juliaNab, kaleidoscope, lazyTravis, Lissajous, mask, MobiusN, mobius_strip, modulus, murl, murl2, npolar, ortho, oscilloscope, parabola, perspective, petal, phoenix_julia, pie (was also inconsistent between cpu and gpu), poincare, popcorn, popcorn2, power, pow_block, rational3, rays, rblur, rings, rippled, roundspher, sec, secant2, sigmoid, sin, sineblur, sinh, sinusgrid, sphericaln, spiralwing, spirograph, split, squarize, squirrel, squish, sschecks, starblur, stripes, stwin, super_shape, tan, tancos, tangent, tanh, TwinTrian, twoface, unpolar, waves, wavesn, wedge_julia, whorl, xheart, zblur, zscale. --Code changes -Generalize Variation::PrecalcHelper() and rename to PrePostPrecalcHelper(). --Do the same for the OpenCL version and rename it PrePostPrecalcOpenCLString(). -Rename Variation::m_AssignType to m_PrePostAssignType since it's only relevant to pre/post variations.
2016-01-29 20:02:15 -05:00
helper.Out.z = DefaultZ(helper);
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
}
--Code changes -Better use of const functions. -Ember.pro should not define BUILDING_EMBERCL, instead define BUILDING_EMBER.
2015-06-29 23:13:53 -04:00
virtual string OpenCLString() const override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
ostringstream ss;
0.4.1.3 Beta 10/14/2014 --User Changes Size is no longer fixed to the window size. Size scaling is done differently in the final render dialog. This fixes several bugs. Remove Xml saving size from settings and options dialog, it no longer applies. Final render can be broken into strips. Set default save path to the desktop if none is found in the settings file. Set default output size to 1920x1080 if none is found in the settings file. --Bug Fixes Better memory size reporting in final render dialog. --Code Changes Migrate to C++11, Qt 5.3.1, and Visual Studio 2013. Change most instances of unsigned int to size_t, and int to intmax_t. Add m_OrigPixPerUnit and m_ScaleType to Ember for scaling purposes. Replace some sprintf_s() calls in XmlToEmber with ostringstream. Move more non-templated members into RendererBase. Add CopyVec() overload that takes a per element function pointer. Add vector Memset(). Replace '&' with '+' instead of "&amp;" in XmlToEmber for much faster parsing. Break strips rendering out into EmberCommon and call from EmberRender and Fractorium. Make AddAndWriteBuffer() just call WriteBuffer(). Make AddAndWriteImage() delete the existing image first before replacing it. Add SetOutputTexture() to RendererCL to support making new textures in response to resize events. Remove multiple return statements in RendererCL, and replace with a bool that tracks results. Add ToDouble(), MakeEnd(), ToString() and Exists() wrappers in Fractorium. Add Size() wrapper in EmberFile. Make QString function arguments const QString&, and string with const string&. Make ShowCritical() wrapper for invoking a message box from another thread. Add combo box to TwoButtonWidget and rename.
2014-10-14 11:53:15 -04:00
intmax_t varIndex = IndexInXform();
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
string weight = WeightDefineString();
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
ss << "\t{\n"
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
<< "\t\tvOut.x = " << weight << " / Zeps(precalcSqrtSumSquares);\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t\tvOut.y = precalcAtanyx;\n"
--User changes -Allow for pausing the renderer in the main window. This makes is more efficient when entering many parameters, such as when following a tutorial. -Add support for new variations: erf, gamma, jac_cn, jac_dn, jac_sn, logDB, pressure_wave, pRose3D, splits3D, w, waves2b, x, xerf, y, z. -Inform user of the start and stop of file parsing in EmberAnimate because the files could potentially be very large. -Move the follwing fields to a new table called Animation: Interpolation, Affine Interpolation, Temporal Samples, Temporal Filter Width, Temporal Filter Type. -These currently have no effect on the interactive renderer and instead are used when running flames through EmberGenome to generate sequences, and then animating them in Fractorium or EmberAnimate. -Add new parameter overrides for EmberRender and EmberAnimate which directly assign values to all flames being rendered, rather than scale: --quality --demin --demax --Bug fixes -Left pad instead of right pad names of sequence outputs from EmberGenome. -Unique file naming was broken for files which already had an underscore in them. -Properly report that png is the default format of EmberRender and EmberAnimate output instead of erroneously claiming it was jpg. -Make command line programs search these folders in this order for the palette file: ./ ~/.fractorium ~/.config/fractorium /usr/share/fractorium /usr/local/share/fractorium -Fix possible bad values in hexes. -Better assignment of Z variables. -Fix boarders due to use of poorly implemented rint() function from flam3. Use std::rint() now. -wedge_sph was completely wrong due to having accidentally swapped the mapping of two parameters. -Make juliascope juliascope_power parameter be of type REAL_NONZERO since it's used as a denominator. -Make Z assignment compatible with the originals in: -arch, bcircle, bCollide, bent, bent2, bisplit, blob, blur_linear, blur_square, bMod, boarders, boarders2, bSwirl, bTransform, butterfly, cardioid, cell, circleblur, circlize, circlize2, circus, collideoscope, cos, cosine, cosh, coth, cpow, cpow2, crescents, cropn, csc, csch, curl, curve, dc_gridout, deltaa, diamond, disc2, eclipse, eCollide, edisc, eJulia, elliptic, eMod, eMotion, ennepers, epispiral, ePush, eRotate, eScale, eSwirl, ex, exp, expo, exponential, fan, fdisc, fibonacci, fibonacci2, fisheye, flipcircle, flipy, flower, flux, funnel, glynnia, GlynnSim1, GlynnSim2, GlynnSim3, gridout, handkerchief, heart, hole, idisc, julia, julian2, juliaNab, kaleidoscope, lazyTravis, Lissajous, mask, MobiusN, mobius_strip, modulus, murl, murl2, npolar, ortho, oscilloscope, parabola, perspective, petal, phoenix_julia, pie (was also inconsistent between cpu and gpu), poincare, popcorn, popcorn2, power, pow_block, rational3, rays, rblur, rings, rippled, roundspher, sec, secant2, sigmoid, sin, sineblur, sinh, sinusgrid, sphericaln, spiralwing, spirograph, split, squarize, squirrel, squish, sschecks, starblur, stripes, stwin, super_shape, tan, tancos, tangent, tanh, TwinTrian, twoface, unpolar, waves, wavesn, wedge_julia, whorl, xheart, zblur, zscale. --Code changes -Generalize Variation::PrecalcHelper() and rename to PrePostPrecalcHelper(). --Do the same for the OpenCL version and rename it PrePostPrecalcOpenCLString(). -Rename Variation::m_AssignType to m_PrePostAssignType since it's only relevant to pre/post variations.
2016-01-29 20:02:15 -05:00
<< "\t\tvOut.z = " << DefaultZCl()
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t}\n";
return ss.str();
}
--User changes -Add new variations: bubbleT3D, crob, hexaplay3D, hexcrop, hexes, hexnix3D, loonie2, loonie3, nBlur, octapol and synth. -Allow for pre/post versions of dc_bubble, dc_cylinder and dc_linear whereas before they were omitted. -When saving a file with multiple embers in it, detect if time values are all the same and if so, start them at zero and increment by 1 for each ember. -Allow for numerous quality increases to be coalesced into one. It will pick up at the end of the current render. -Show selection highlight on variations tree in response to mouse hover. This makes it easier to see for which variation or param the current mouse wheel action will apply. -Make default temporal samples be 100, whereas before it was 1000 which was overkill. -Require the shift key to be held with delete for deleting an ember to prevent it from triggering when the user enters delete in the edit box. -This wasn't otherwise fixable without writing a lot more code. --Bug fixes -EmberGenome was crashing when generating a sequence from a source file with more than 2 embers in it. -EmberGenome was improperly handling the first frame of a merge after the last frame of the loop. -These bugs were due to a previous commit. Revert parts of that commit. -Prevent a zoom value of less than 0 when reading from xml. -Slight optimization of the crescents, and mask variations, if the compiler wasn't doing it already. -Unique file naming was broken because it was looking for _# and the default names ended with -#. -Disallow renaming of an ember in the library tree to an empty string. -Severe bug that prevented some variations from being read correctly from params generated outside this program. -Severe OpenCL randomization bug. The first x coordinates of the first points in the first kernel call of the first ember of a render since the OpenCL renderer object was created were not random and were mostly -1. -Severe bug when populating xform selection distributions that could sometimes cause a crash due to roundoff error. Fix by using double. -Limit the max number of variations in a random ember to MAX_CL_VARS, which is 8. This ensures they'll look the same on CPU and GPU. -Prevent user from saving stylesheet to default.qss, it's a special reserved filename. --Code changes -Generalize using the running sum output point inside of a variation for all cases: pre, reg and post. -Allow for array variables in variations where the address of each element is stored in m_Params. -Qualify all math functions with std:: -No longer use our own Clamp() in OpenCL, instead use the standard clamp(). -Redesign how functions are used in the variations OpenCL code. -Add tests to EmberTester to verify some of the new functionality. -Place more const and override qualifiers on functions where appropriate. -Add a global rand with a lock to be used very sparingly. -Use a map instead of a vector for bad param names in Xml parsing. -Prefix affine interpolation mode defines with "AFFINE_" to make their purpose more clear. -Allow for variations that change state during iteration by sending a separate copy of the ember to each rendering thread. -Implement this same functionality with a local struct in OpenCL. It's members are the total of all variables that need to change state within an ember. -Add Contains() function to Utils.h. -EmberRender: print names of kernels being printed with --dump_kernel option. -Clean up EmberTester to handle some of the recent changes. -Fix various casts. -Replace % 2 with & 1, even though the compiler was likely doing this already. -Add new file Variations06.h to accommodate new variations. -General cleanup.
2015-11-22 17:15:07 -05:00
virtual vector<string> OpenCLGlobalFuncNames() const override
{
return vector<string> { "Zeps" };
}
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
};
/// <summary>
/// q_ode.
/// </summary>
template <typename T>
--Code changes -Split VariationList into .h/cpp files. --Make VariationList.cpp the only place where the variations files are included from. This speeds up the build, but variations can no longer be directly instantiated. --Variations are no longer exported.
2016-02-18 21:58:24 -05:00
class QodeVariation : public ParametricVariation<T>
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
public:
--Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums.
2016-01-04 19:50:15 -05:00
QodeVariation(T weight = 1.0) : ParametricVariation<T>("q_ode", eVariationId::VAR_Q_ODE, weight)
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
Init();
}
PARVARCOPY(QodeVariation)
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
T sqx = SQR(helper.In.x);
T sqy = SQR(helper.In.y);
T xy = helper.In.x * helper.In.y;
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
helper.Out.x = (m_Q01 + m_Weight * m_Q02 * helper.In.x + m_Q03 * sqx) +
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
(m_Q04 * xy + m_Q05 * helper.In.y + m_Q06 * sqy);
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
helper.Out.y = (m_Q07 + m_Q08 * helper.In.x + m_Q09 * sqx) +
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
(m_Q10 * xy + m_Weight * m_Q11 * helper.In.y + m_Q12 * sqy);
--User changes -Allow for pausing the renderer in the main window. This makes is more efficient when entering many parameters, such as when following a tutorial. -Add support for new variations: erf, gamma, jac_cn, jac_dn, jac_sn, logDB, pressure_wave, pRose3D, splits3D, w, waves2b, x, xerf, y, z. -Inform user of the start and stop of file parsing in EmberAnimate because the files could potentially be very large. -Move the follwing fields to a new table called Animation: Interpolation, Affine Interpolation, Temporal Samples, Temporal Filter Width, Temporal Filter Type. -These currently have no effect on the interactive renderer and instead are used when running flames through EmberGenome to generate sequences, and then animating them in Fractorium or EmberAnimate. -Add new parameter overrides for EmberRender and EmberAnimate which directly assign values to all flames being rendered, rather than scale: --quality --demin --demax --Bug fixes -Left pad instead of right pad names of sequence outputs from EmberGenome. -Unique file naming was broken for files which already had an underscore in them. -Properly report that png is the default format of EmberRender and EmberAnimate output instead of erroneously claiming it was jpg. -Make command line programs search these folders in this order for the palette file: ./ ~/.fractorium ~/.config/fractorium /usr/share/fractorium /usr/local/share/fractorium -Fix possible bad values in hexes. -Better assignment of Z variables. -Fix boarders due to use of poorly implemented rint() function from flam3. Use std::rint() now. -wedge_sph was completely wrong due to having accidentally swapped the mapping of two parameters. -Make juliascope juliascope_power parameter be of type REAL_NONZERO since it's used as a denominator. -Make Z assignment compatible with the originals in: -arch, bcircle, bCollide, bent, bent2, bisplit, blob, blur_linear, blur_square, bMod, boarders, boarders2, bSwirl, bTransform, butterfly, cardioid, cell, circleblur, circlize, circlize2, circus, collideoscope, cos, cosine, cosh, coth, cpow, cpow2, crescents, cropn, csc, csch, curl, curve, dc_gridout, deltaa, diamond, disc2, eclipse, eCollide, edisc, eJulia, elliptic, eMod, eMotion, ennepers, epispiral, ePush, eRotate, eScale, eSwirl, ex, exp, expo, exponential, fan, fdisc, fibonacci, fibonacci2, fisheye, flipcircle, flipy, flower, flux, funnel, glynnia, GlynnSim1, GlynnSim2, GlynnSim3, gridout, handkerchief, heart, hole, idisc, julia, julian2, juliaNab, kaleidoscope, lazyTravis, Lissajous, mask, MobiusN, mobius_strip, modulus, murl, murl2, npolar, ortho, oscilloscope, parabola, perspective, petal, phoenix_julia, pie (was also inconsistent between cpu and gpu), poincare, popcorn, popcorn2, power, pow_block, rational3, rays, rblur, rings, rippled, roundspher, sec, secant2, sigmoid, sin, sineblur, sinh, sinusgrid, sphericaln, spiralwing, spirograph, split, squarize, squirrel, squish, sschecks, starblur, stripes, stwin, super_shape, tan, tancos, tangent, tanh, TwinTrian, twoface, unpolar, waves, wavesn, wedge_julia, whorl, xheart, zblur, zscale. --Code changes -Generalize Variation::PrecalcHelper() and rename to PrePostPrecalcHelper(). --Do the same for the OpenCL version and rename it PrePostPrecalcOpenCLString(). -Rename Variation::m_AssignType to m_PrePostAssignType since it's only relevant to pre/post variations.
2016-01-29 20:02:15 -05:00
helper.Out.z = DefaultZ(helper);
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
}
--Code changes -Better use of const functions. -Ember.pro should not define BUILDING_EMBERCL, instead define BUILDING_EMBER.
2015-06-29 23:13:53 -04:00
virtual string OpenCLString() const override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
ostringstream ss, ss2;
0.4.1.3 Beta 10/14/2014 --User Changes Size is no longer fixed to the window size. Size scaling is done differently in the final render dialog. This fixes several bugs. Remove Xml saving size from settings and options dialog, it no longer applies. Final render can be broken into strips. Set default save path to the desktop if none is found in the settings file. Set default output size to 1920x1080 if none is found in the settings file. --Bug Fixes Better memory size reporting in final render dialog. --Code Changes Migrate to C++11, Qt 5.3.1, and Visual Studio 2013. Change most instances of unsigned int to size_t, and int to intmax_t. Add m_OrigPixPerUnit and m_ScaleType to Ember for scaling purposes. Replace some sprintf_s() calls in XmlToEmber with ostringstream. Move more non-templated members into RendererBase. Add CopyVec() overload that takes a per element function pointer. Add vector Memset(). Replace '&' with '+' instead of "&amp;" in XmlToEmber for much faster parsing. Break strips rendering out into EmberCommon and call from EmberRender and Fractorium. Make AddAndWriteBuffer() just call WriteBuffer(). Make AddAndWriteImage() delete the existing image first before replacing it. Add SetOutputTexture() to RendererCL to support making new textures in response to resize events. Remove multiple return statements in RendererCL, and replace with a bool that tracks results. Add ToDouble(), MakeEnd(), ToString() and Exists() wrappers in Fractorium. Add Size() wrapper in EmberFile. Make QString function arguments const QString&, and string with const string&. Make ShowCritical() wrapper for invoking a message box from another thread. Add combo box to TwoButtonWidget and rename.
2014-10-14 11:53:15 -04:00
intmax_t i = 0, varIndex = IndexInXform();
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
ss2 << "_" << XformIndexInEmber() << "]";
string index = ss2.str();
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
string weight = WeightDefineString();
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
string q01 = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string q02 = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string q03 = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string q04 = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string q05 = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string q06 = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string q07 = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string q08 = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string q09 = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string q10 = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string q11 = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string q12 = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
ss << "\t{\n"
<< "\t\treal_t sqx = SQR(vIn.x);\n"
<< "\t\treal_t sqy = SQR(vIn.y);\n"
<< "\t\treal_t xy = vIn.x * vIn.y;\n"
<< "\n"
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
<< "\t\tvOut.x = (" << q01 << " + fma(" << weight << ", " << q02 << " * vIn.x, " << q03 << " * sqx)) + \n"
<< "\t\t fma(" << q04 << ", xy, fma(" << q05 << ", vIn.y, " << q06 << " * sqy));\n"
<< "\t\tvOut.y = (" << q07 << " + fma(" << q08 << ", vIn.x, " << q09 << " * sqx)) + \n"
<< "\t\t fma(" << q10 << ", xy, fma(" << weight << ", " << q11 << " * vIn.y, " << q12 << " * sqy));\n"
--User changes -Allow for pausing the renderer in the main window. This makes is more efficient when entering many parameters, such as when following a tutorial. -Add support for new variations: erf, gamma, jac_cn, jac_dn, jac_sn, logDB, pressure_wave, pRose3D, splits3D, w, waves2b, x, xerf, y, z. -Inform user of the start and stop of file parsing in EmberAnimate because the files could potentially be very large. -Move the follwing fields to a new table called Animation: Interpolation, Affine Interpolation, Temporal Samples, Temporal Filter Width, Temporal Filter Type. -These currently have no effect on the interactive renderer and instead are used when running flames through EmberGenome to generate sequences, and then animating them in Fractorium or EmberAnimate. -Add new parameter overrides for EmberRender and EmberAnimate which directly assign values to all flames being rendered, rather than scale: --quality --demin --demax --Bug fixes -Left pad instead of right pad names of sequence outputs from EmberGenome. -Unique file naming was broken for files which already had an underscore in them. -Properly report that png is the default format of EmberRender and EmberAnimate output instead of erroneously claiming it was jpg. -Make command line programs search these folders in this order for the palette file: ./ ~/.fractorium ~/.config/fractorium /usr/share/fractorium /usr/local/share/fractorium -Fix possible bad values in hexes. -Better assignment of Z variables. -Fix boarders due to use of poorly implemented rint() function from flam3. Use std::rint() now. -wedge_sph was completely wrong due to having accidentally swapped the mapping of two parameters. -Make juliascope juliascope_power parameter be of type REAL_NONZERO since it's used as a denominator. -Make Z assignment compatible with the originals in: -arch, bcircle, bCollide, bent, bent2, bisplit, blob, blur_linear, blur_square, bMod, boarders, boarders2, bSwirl, bTransform, butterfly, cardioid, cell, circleblur, circlize, circlize2, circus, collideoscope, cos, cosine, cosh, coth, cpow, cpow2, crescents, cropn, csc, csch, curl, curve, dc_gridout, deltaa, diamond, disc2, eclipse, eCollide, edisc, eJulia, elliptic, eMod, eMotion, ennepers, epispiral, ePush, eRotate, eScale, eSwirl, ex, exp, expo, exponential, fan, fdisc, fibonacci, fibonacci2, fisheye, flipcircle, flipy, flower, flux, funnel, glynnia, GlynnSim1, GlynnSim2, GlynnSim3, gridout, handkerchief, heart, hole, idisc, julia, julian2, juliaNab, kaleidoscope, lazyTravis, Lissajous, mask, MobiusN, mobius_strip, modulus, murl, murl2, npolar, ortho, oscilloscope, parabola, perspective, petal, phoenix_julia, pie (was also inconsistent between cpu and gpu), poincare, popcorn, popcorn2, power, pow_block, rational3, rays, rblur, rings, rippled, roundspher, sec, secant2, sigmoid, sin, sineblur, sinh, sinusgrid, sphericaln, spiralwing, spirograph, split, squarize, squirrel, squish, sschecks, starblur, stripes, stwin, super_shape, tan, tancos, tangent, tanh, TwinTrian, twoface, unpolar, waves, wavesn, wedge_julia, whorl, xheart, zblur, zscale. --Code changes -Generalize Variation::PrecalcHelper() and rename to PrePostPrecalcHelper(). --Do the same for the OpenCL version and rename it PrePostPrecalcOpenCLString(). -Rename Variation::m_AssignType to m_PrePostAssignType since it's only relevant to pre/post variations.
2016-01-29 20:02:15 -05:00
<< "\t\tvOut.z = " << DefaultZCl()
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t}\n";
return ss.str();
}
protected:
void Init()
{
string prefix = Prefix();
m_Params.clear();
m_Params.push_back(ParamWithName<T>(&m_Q01, prefix + "q_ode01", 1));
m_Params.push_back(ParamWithName<T>(&m_Q02, prefix + "q_ode02", -1));
m_Params.push_back(ParamWithName<T>(&m_Q03, prefix + "q_ode03"));
m_Params.push_back(ParamWithName<T>(&m_Q04, prefix + "q_ode04"));
m_Params.push_back(ParamWithName<T>(&m_Q05, prefix + "q_ode05"));
m_Params.push_back(ParamWithName<T>(&m_Q06, prefix + "q_ode06"));
m_Params.push_back(ParamWithName<T>(&m_Q07, prefix + "q_ode07", 1));
m_Params.push_back(ParamWithName<T>(&m_Q08, prefix + "q_ode08"));
m_Params.push_back(ParamWithName<T>(&m_Q09, prefix + "q_ode09"));
m_Params.push_back(ParamWithName<T>(&m_Q10, prefix + "q_ode10"));
m_Params.push_back(ParamWithName<T>(&m_Q11, prefix + "q_ode11"));
m_Params.push_back(ParamWithName<T>(&m_Q12, prefix + "q_ode12"));
}
private:
T m_Q01;
T m_Q02;
T m_Q03;
T m_Q04;
T m_Q05;
T m_Q06;
T m_Q07;
T m_Q08;
T m_Q09;
T m_Q10;
T m_Q11;
T m_Q12;
};
/// <summary>
/// blur_heart.
/// </summary>
template <typename T>
--Code changes -Split VariationList into .h/cpp files. --Make VariationList.cpp the only place where the variations files are included from. This speeds up the build, but variations can no longer be directly instantiated. --Variations are no longer exported.
2016-02-18 21:58:24 -05:00
class BlurHeartVariation : public ParametricVariation<T>
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
public:
--Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums.
2016-01-04 19:50:15 -05:00
BlurHeartVariation(T weight = 1.0) : ParametricVariation<T>("blur_heart", eVariationId::VAR_BLUR_HEART, weight)
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
Init();
}
PARVARCOPY(BlurHeartVariation)
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
T xx = (rand.Frand01<T>() - T(0.5)) * 2;
T yy = (rand.Frand01<T>() - T(0.5)) * 2;
--Code changes -Make usage of VarFuncs more consistent by moving some global functions there. -Make CircleTrans1 a little safer by calling Zeps() on m_Sc during precalc. -Begin prepping for 0.9.9.5 release.
2016-03-04 21:54:06 -05:00
T k = VarFuncs<T>::SignNz(yy);
--User changes -Add post_smartcrop. --Bug fixes -Fix bug in crackle. -Wrong point assignment in hexaplay3D, hexnix3D. -Improper Z assignment in rblur. -Fix inconsistency with original in circlecrop. -Put EMBER_ROOT bakc to ./../../../ in default.pri. This is TBD. --Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums. -Optimize crob, nBlur. -Fix weird assignment statement in falloff3. -Cleanup in VarFuncs::SimplexNoise3D(). -Replace fabs() with std::abs(). -General cleanup.
2016-01-12 23:42:12 -05:00
T yymax = ((m_A * std::pow(std::abs(xx), m_P) + k * m_B * std::sqrt(std::abs(1 - SQR(xx)))) - m_A);
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
//The function must be in a range 0-1 to work properly.
--User changes -Add post_smartcrop. --Bug fixes -Fix bug in crackle. -Wrong point assignment in hexaplay3D, hexnix3D. -Improper Z assignment in rblur. -Fix inconsistency with original in circlecrop. -Put EMBER_ROOT bakc to ./../../../ in default.pri. This is TBD. --Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums. -Optimize crob, nBlur. -Fix weird assignment statement in falloff3. -Cleanup in VarFuncs::SimplexNoise3D(). -Replace fabs() with std::abs(). -General cleanup.
2016-01-12 23:42:12 -05:00
yymax /= Zeps(std::abs(m_A) + std::abs(m_B));
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
//Quick and dirty way to force y to be in range without altering the density.
if (k > 0)
{
if (yy > yymax)
yy = yymax;
}
else
{
if (yy < yymax)
yy = yymax;
}
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
helper.Out.x = xx * m_Weight;
helper.Out.y = yy * m_Weight;
--User changes -Allow for pausing the renderer in the main window. This makes is more efficient when entering many parameters, such as when following a tutorial. -Add support for new variations: erf, gamma, jac_cn, jac_dn, jac_sn, logDB, pressure_wave, pRose3D, splits3D, w, waves2b, x, xerf, y, z. -Inform user of the start and stop of file parsing in EmberAnimate because the files could potentially be very large. -Move the follwing fields to a new table called Animation: Interpolation, Affine Interpolation, Temporal Samples, Temporal Filter Width, Temporal Filter Type. -These currently have no effect on the interactive renderer and instead are used when running flames through EmberGenome to generate sequences, and then animating them in Fractorium or EmberAnimate. -Add new parameter overrides for EmberRender and EmberAnimate which directly assign values to all flames being rendered, rather than scale: --quality --demin --demax --Bug fixes -Left pad instead of right pad names of sequence outputs from EmberGenome. -Unique file naming was broken for files which already had an underscore in them. -Properly report that png is the default format of EmberRender and EmberAnimate output instead of erroneously claiming it was jpg. -Make command line programs search these folders in this order for the palette file: ./ ~/.fractorium ~/.config/fractorium /usr/share/fractorium /usr/local/share/fractorium -Fix possible bad values in hexes. -Better assignment of Z variables. -Fix boarders due to use of poorly implemented rint() function from flam3. Use std::rint() now. -wedge_sph was completely wrong due to having accidentally swapped the mapping of two parameters. -Make juliascope juliascope_power parameter be of type REAL_NONZERO since it's used as a denominator. -Make Z assignment compatible with the originals in: -arch, bcircle, bCollide, bent, bent2, bisplit, blob, blur_linear, blur_square, bMod, boarders, boarders2, bSwirl, bTransform, butterfly, cardioid, cell, circleblur, circlize, circlize2, circus, collideoscope, cos, cosine, cosh, coth, cpow, cpow2, crescents, cropn, csc, csch, curl, curve, dc_gridout, deltaa, diamond, disc2, eclipse, eCollide, edisc, eJulia, elliptic, eMod, eMotion, ennepers, epispiral, ePush, eRotate, eScale, eSwirl, ex, exp, expo, exponential, fan, fdisc, fibonacci, fibonacci2, fisheye, flipcircle, flipy, flower, flux, funnel, glynnia, GlynnSim1, GlynnSim2, GlynnSim3, gridout, handkerchief, heart, hole, idisc, julia, julian2, juliaNab, kaleidoscope, lazyTravis, Lissajous, mask, MobiusN, mobius_strip, modulus, murl, murl2, npolar, ortho, oscilloscope, parabola, perspective, petal, phoenix_julia, pie (was also inconsistent between cpu and gpu), poincare, popcorn, popcorn2, power, pow_block, rational3, rays, rblur, rings, rippled, roundspher, sec, secant2, sigmoid, sin, sineblur, sinh, sinusgrid, sphericaln, spiralwing, spirograph, split, squarize, squirrel, squish, sschecks, starblur, stripes, stwin, super_shape, tan, tancos, tangent, tanh, TwinTrian, twoface, unpolar, waves, wavesn, wedge_julia, whorl, xheart, zblur, zscale. --Code changes -Generalize Variation::PrecalcHelper() and rename to PrePostPrecalcHelper(). --Do the same for the OpenCL version and rename it PrePostPrecalcOpenCLString(). -Rename Variation::m_AssignType to m_PrePostAssignType since it's only relevant to pre/post variations.
2016-01-29 20:02:15 -05:00
helper.Out.z = DefaultZ(helper);
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
}
--Code changes -Better use of const functions. -Ember.pro should not define BUILDING_EMBERCL, instead define BUILDING_EMBER.
2015-06-29 23:13:53 -04:00
virtual string OpenCLString() const override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
ostringstream ss, ss2;
0.4.1.3 Beta 10/14/2014 --User Changes Size is no longer fixed to the window size. Size scaling is done differently in the final render dialog. This fixes several bugs. Remove Xml saving size from settings and options dialog, it no longer applies. Final render can be broken into strips. Set default save path to the desktop if none is found in the settings file. Set default output size to 1920x1080 if none is found in the settings file. --Bug Fixes Better memory size reporting in final render dialog. --Code Changes Migrate to C++11, Qt 5.3.1, and Visual Studio 2013. Change most instances of unsigned int to size_t, and int to intmax_t. Add m_OrigPixPerUnit and m_ScaleType to Ember for scaling purposes. Replace some sprintf_s() calls in XmlToEmber with ostringstream. Move more non-templated members into RendererBase. Add CopyVec() overload that takes a per element function pointer. Add vector Memset(). Replace '&' with '+' instead of "&amp;" in XmlToEmber for much faster parsing. Break strips rendering out into EmberCommon and call from EmberRender and Fractorium. Make AddAndWriteBuffer() just call WriteBuffer(). Make AddAndWriteImage() delete the existing image first before replacing it. Add SetOutputTexture() to RendererCL to support making new textures in response to resize events. Remove multiple return statements in RendererCL, and replace with a bool that tracks results. Add ToDouble(), MakeEnd(), ToString() and Exists() wrappers in Fractorium. Add Size() wrapper in EmberFile. Make QString function arguments const QString&, and string with const string&. Make ShowCritical() wrapper for invoking a message box from another thread. Add combo box to TwoButtonWidget and rename.
2014-10-14 11:53:15 -04:00
intmax_t i = 0, varIndex = IndexInXform();
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
ss2 << "_" << XformIndexInEmber() << "]";
string index = ss2.str();
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
string weight = WeightDefineString();
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
string p = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string a = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string b = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
ss << "\t{\n"
Ditch T(f) define in opencl programs in preference to already-defined real_t. Also removed some erroneously-added double-'T(' statements
2015-02-26 16:18:55 -05:00
<< "\t\treal_t xx = (MwcNext01(mwc) - (real_t)(0.5)) * 2;\n"
<< "\t\treal_t yy = (MwcNext01(mwc) - (real_t)(0.5)) * 2;\n"
Numerous fixes 0.4.0.5 Beta 07/18/2014 --User Changes Allow for vibrancy values > 1. Add flatten and unflatten menu items. Automatically flatten like Apophysis does. Add plugin and new_linear tags to Xml to be compatible with Apophysis. --Bug Fixes Fix blur, blur3d, bubble, cropn, cross, curl, curl3d, epispiral, ho, julia3d, julia3dz, loonie, mirror_x, mirror_y, mirror_z, rotate_x, sinusoidal, spherical, spherical3d, stripes. Unique filename on final render was completely broken. Two severe OpenCL bugs. Random seeds were biased and fusing was being reset too often leading to results that differ from the CPU. Subtle, but sometimes severe bug in the setup of the xaos weights. Use properly defined epsilon by getting the value from std::numeric_limits, rather than hard coding 1e-6 or 1e-10. Omit incorrect usage of epsilon everywhere. It should not be automatically added to denominators. Rather, it should only be used if the denominator is zero. Force final render progress bars to 100 on completion. Sometimes they didn't seem to make it there. Make variation name and params comparisons be case insensitive. --Code Changes Make ForEach and FindIf wrappers around std::for_each and std::find_if.
2014-07-19 02:33:18 -04:00
<< "\t\treal_t k = SignNz(yy);\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t\treal_t yymax = ((" << a << " * pow(fabs(xx), " << p << ") + k * " << b << " * sqrt(fabs(1 - SQR(xx)))) - " << a << ");\n"
<< "\n"
Numerous fixes 0.4.0.5 Beta 07/18/2014 --User Changes Allow for vibrancy values > 1. Add flatten and unflatten menu items. Automatically flatten like Apophysis does. Add plugin and new_linear tags to Xml to be compatible with Apophysis. --Bug Fixes Fix blur, blur3d, bubble, cropn, cross, curl, curl3d, epispiral, ho, julia3d, julia3dz, loonie, mirror_x, mirror_y, mirror_z, rotate_x, sinusoidal, spherical, spherical3d, stripes. Unique filename on final render was completely broken. Two severe OpenCL bugs. Random seeds were biased and fusing was being reset too often leading to results that differ from the CPU. Subtle, but sometimes severe bug in the setup of the xaos weights. Use properly defined epsilon by getting the value from std::numeric_limits, rather than hard coding 1e-6 or 1e-10. Omit incorrect usage of epsilon everywhere. It should not be automatically added to denominators. Rather, it should only be used if the denominator is zero. Force final render progress bars to 100 on completion. Sometimes they didn't seem to make it there. Make variation name and params comparisons be case insensitive. --Code Changes Make ForEach and FindIf wrappers around std::for_each and std::find_if.
2014-07-19 02:33:18 -04:00
<< "\t\tyymax /= Zeps(fabs(" << a << ") + fabs(" << b << "));\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\n"
<< "\t\tif (k > 0)\n"
<< "\t\t{\n"
<< "\t\t if (yy > yymax)\n"
<< "\t\t yy = yymax;\n"
<< "\t\t}\n"
<< "\t\telse\n"
<< "\t\t{\n"
<< "\t\t if (yy < yymax)\n"
<< "\t\t yy = yymax;\n"
<< "\t\t}\n"
<< "\n"
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
<< "\t\tvOut.x = xx * " << weight << ";\n"
<< "\t\tvOut.y = yy * " << weight << ";\n"
--User changes -Allow for pausing the renderer in the main window. This makes is more efficient when entering many parameters, such as when following a tutorial. -Add support for new variations: erf, gamma, jac_cn, jac_dn, jac_sn, logDB, pressure_wave, pRose3D, splits3D, w, waves2b, x, xerf, y, z. -Inform user of the start and stop of file parsing in EmberAnimate because the files could potentially be very large. -Move the follwing fields to a new table called Animation: Interpolation, Affine Interpolation, Temporal Samples, Temporal Filter Width, Temporal Filter Type. -These currently have no effect on the interactive renderer and instead are used when running flames through EmberGenome to generate sequences, and then animating them in Fractorium or EmberAnimate. -Add new parameter overrides for EmberRender and EmberAnimate which directly assign values to all flames being rendered, rather than scale: --quality --demin --demax --Bug fixes -Left pad instead of right pad names of sequence outputs from EmberGenome. -Unique file naming was broken for files which already had an underscore in them. -Properly report that png is the default format of EmberRender and EmberAnimate output instead of erroneously claiming it was jpg. -Make command line programs search these folders in this order for the palette file: ./ ~/.fractorium ~/.config/fractorium /usr/share/fractorium /usr/local/share/fractorium -Fix possible bad values in hexes. -Better assignment of Z variables. -Fix boarders due to use of poorly implemented rint() function from flam3. Use std::rint() now. -wedge_sph was completely wrong due to having accidentally swapped the mapping of two parameters. -Make juliascope juliascope_power parameter be of type REAL_NONZERO since it's used as a denominator. -Make Z assignment compatible with the originals in: -arch, bcircle, bCollide, bent, bent2, bisplit, blob, blur_linear, blur_square, bMod, boarders, boarders2, bSwirl, bTransform, butterfly, cardioid, cell, circleblur, circlize, circlize2, circus, collideoscope, cos, cosine, cosh, coth, cpow, cpow2, crescents, cropn, csc, csch, curl, curve, dc_gridout, deltaa, diamond, disc2, eclipse, eCollide, edisc, eJulia, elliptic, eMod, eMotion, ennepers, epispiral, ePush, eRotate, eScale, eSwirl, ex, exp, expo, exponential, fan, fdisc, fibonacci, fibonacci2, fisheye, flipcircle, flipy, flower, flux, funnel, glynnia, GlynnSim1, GlynnSim2, GlynnSim3, gridout, handkerchief, heart, hole, idisc, julia, julian2, juliaNab, kaleidoscope, lazyTravis, Lissajous, mask, MobiusN, mobius_strip, modulus, murl, murl2, npolar, ortho, oscilloscope, parabola, perspective, petal, phoenix_julia, pie (was also inconsistent between cpu and gpu), poincare, popcorn, popcorn2, power, pow_block, rational3, rays, rblur, rings, rippled, roundspher, sec, secant2, sigmoid, sin, sineblur, sinh, sinusgrid, sphericaln, spiralwing, spirograph, split, squarize, squirrel, squish, sschecks, starblur, stripes, stwin, super_shape, tan, tancos, tangent, tanh, TwinTrian, twoface, unpolar, waves, wavesn, wedge_julia, whorl, xheart, zblur, zscale. --Code changes -Generalize Variation::PrecalcHelper() and rename to PrePostPrecalcHelper(). --Do the same for the OpenCL version and rename it PrePostPrecalcOpenCLString(). -Rename Variation::m_AssignType to m_PrePostAssignType since it's only relevant to pre/post variations.
2016-01-29 20:02:15 -05:00
<< "\t\tvOut.z = " << DefaultZCl()
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t}\n";
return ss.str();
}
--User changes -Add new variations: bubbleT3D, crob, hexaplay3D, hexcrop, hexes, hexnix3D, loonie2, loonie3, nBlur, octapol and synth. -Allow for pre/post versions of dc_bubble, dc_cylinder and dc_linear whereas before they were omitted. -When saving a file with multiple embers in it, detect if time values are all the same and if so, start them at zero and increment by 1 for each ember. -Allow for numerous quality increases to be coalesced into one. It will pick up at the end of the current render. -Show selection highlight on variations tree in response to mouse hover. This makes it easier to see for which variation or param the current mouse wheel action will apply. -Make default temporal samples be 100, whereas before it was 1000 which was overkill. -Require the shift key to be held with delete for deleting an ember to prevent it from triggering when the user enters delete in the edit box. -This wasn't otherwise fixable without writing a lot more code. --Bug fixes -EmberGenome was crashing when generating a sequence from a source file with more than 2 embers in it. -EmberGenome was improperly handling the first frame of a merge after the last frame of the loop. -These bugs were due to a previous commit. Revert parts of that commit. -Prevent a zoom value of less than 0 when reading from xml. -Slight optimization of the crescents, and mask variations, if the compiler wasn't doing it already. -Unique file naming was broken because it was looking for _# and the default names ended with -#. -Disallow renaming of an ember in the library tree to an empty string. -Severe bug that prevented some variations from being read correctly from params generated outside this program. -Severe OpenCL randomization bug. The first x coordinates of the first points in the first kernel call of the first ember of a render since the OpenCL renderer object was created were not random and were mostly -1. -Severe bug when populating xform selection distributions that could sometimes cause a crash due to roundoff error. Fix by using double. -Limit the max number of variations in a random ember to MAX_CL_VARS, which is 8. This ensures they'll look the same on CPU and GPU. -Prevent user from saving stylesheet to default.qss, it's a special reserved filename. --Code changes -Generalize using the running sum output point inside of a variation for all cases: pre, reg and post. -Allow for array variables in variations where the address of each element is stored in m_Params. -Qualify all math functions with std:: -No longer use our own Clamp() in OpenCL, instead use the standard clamp(). -Redesign how functions are used in the variations OpenCL code. -Add tests to EmberTester to verify some of the new functionality. -Place more const and override qualifiers on functions where appropriate. -Add a global rand with a lock to be used very sparingly. -Use a map instead of a vector for bad param names in Xml parsing. -Prefix affine interpolation mode defines with "AFFINE_" to make their purpose more clear. -Allow for variations that change state during iteration by sending a separate copy of the ember to each rendering thread. -Implement this same functionality with a local struct in OpenCL. It's members are the total of all variables that need to change state within an ember. -Add Contains() function to Utils.h. -EmberRender: print names of kernels being printed with --dump_kernel option. -Clean up EmberTester to handle some of the recent changes. -Fix various casts. -Replace % 2 with & 1, even though the compiler was likely doing this already. -Add new file Variations06.h to accommodate new variations. -General cleanup.
2015-11-22 17:15:07 -05:00
virtual vector<string> OpenCLGlobalFuncNames() const override
{
return vector<string> { "SignNz", "Zeps" };
}
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
protected:
void Init()
{
string prefix = Prefix();
m_Params.clear();
m_Params.push_back(ParamWithName<T>(&m_P, prefix + "blur_heart_p", T(0.5)));
Force all float literals in opencl variations to correct precision This avoids 'ambiguous call' errors on mac. It's probably not necessary to have changed every single instance, only the ones in functions with multiple signatures. However, it was easier and safer to do this automatically than manually.
2015-02-25 08:02:50 -05:00
m_Params.push_back(ParamWithName<T>(&m_A, prefix + "blur_heart_a", T(-T(0.6))));
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
m_Params.push_back(ParamWithName<T>(&m_B, prefix + "blur_heart_b", T(0.7)));
}
private:
T m_P;
T m_A;
T m_B;
};
/// <summary>
/// Truchet.
/// </summary>
template <typename T>
--Code changes -Split VariationList into .h/cpp files. --Make VariationList.cpp the only place where the variations files are included from. This speeds up the build, but variations can no longer be directly instantiated. --Variations are no longer exported.
2016-02-18 21:58:24 -05:00
class TruchetVariation : public ParametricVariation<T>
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
public:
--Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums.
2016-01-04 19:50:15 -05:00
TruchetVariation(T weight = 1.0) : ParametricVariation<T>("Truchet", eVariationId::VAR_TRUCHET, weight)
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
Init();
}
PARVARCOPY(TruchetVariation)
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
More Linux work. Convert all casts to new style, away from legacy.
2014-12-07 02:51:44 -05:00
int extended = int(m_Extended);
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
T seed = m_AbsSeed;
T r = -m_Rotation;
T r0 = 0;
T r1 = 0;
T tileType = 0;
T randInt = 0;
T modBase = 65535;
T multiplier = 32747;
T offset = 12345;
T niter = 0;
T x = helper.In.x * m_Scale;
T y = helper.In.y * m_Scale;
More Linux work. Convert all casts to new style, away from legacy.
2014-12-07 02:51:44 -05:00
int intx = int(Round(x));
int inty = int(Round(y));
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
int randiter;
r = x - intx;
if (r < 0)
x = 1 + r;
else
x = r;
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
r = y - inty;
if (r < 0)
y = 1 + r;
else
y = r;
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
//Calculate the tile type.
if (seed == 0)
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
tileType = 0;
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
else if (seed == 1)
tileType = 1;
else
{
if (extended == 0)
{
T xrand = Round(helper.In.x);
T yrand = Round(helper.In.y);
xrand = xrand * m_Seed2;
yrand = yrand * m_Seed2;
--User changes -Add post_smartcrop. --Bug fixes -Fix bug in crackle. -Wrong point assignment in hexaplay3D, hexnix3D. -Improper Z assignment in rblur. -Fix inconsistency with original in circlecrop. -Put EMBER_ROOT bakc to ./../../../ in default.pri. This is TBD. --Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums. -Optimize crob, nBlur. -Fix weird assignment statement in falloff3. -Cleanup in VarFuncs::SimplexNoise3D(). -Replace fabs() with std::abs(). -General cleanup.
2016-01-12 23:42:12 -05:00
niter = xrand + yrand + xrand * yrand;
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
randInt = (niter + seed) * m_Seed2 / 2;
randInt = fmod((randInt * multiplier + offset), modBase);
}
else
{
More Linux work. Convert all casts to new style, away from legacy.
2014-12-07 02:51:44 -05:00
int xrand = int(Round(helper.In.x));
int yrand = int(Round(helper.In.y));
seed = T(Floor<T>(seed));
niter = T(abs(xrand + yrand + xrand * yrand));
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
randInt = seed + niter;
randiter = 0;
Numerous fixes 0.4.0.5 Beta 07/18/2014 --User Changes Allow for vibrancy values > 1. Add flatten and unflatten menu items. Automatically flatten like Apophysis does. Add plugin and new_linear tags to Xml to be compatible with Apophysis. --Bug Fixes Fix blur, blur3d, bubble, cropn, cross, curl, curl3d, epispiral, ho, julia3d, julia3dz, loonie, mirror_x, mirror_y, mirror_z, rotate_x, sinusoidal, spherical, spherical3d, stripes. Unique filename on final render was completely broken. Two severe OpenCL bugs. Random seeds were biased and fusing was being reset too often leading to results that differ from the CPU. Subtle, but sometimes severe bug in the setup of the xaos weights. Use properly defined epsilon by getting the value from std::numeric_limits, rather than hard coding 1e-6 or 1e-10. Omit incorrect usage of epsilon everywhere. It should not be automatically added to denominators. Rather, it should only be used if the denominator is zero. Force final render progress bars to 100 on completion. Sometimes they didn't seem to make it there. Make variation name and params comparisons be case insensitive. --Code Changes Make ForEach and FindIf wrappers around std::for_each and std::find_if.
2014-07-19 02:33:18 -04:00
while (randiter < niter && randiter < 20)//Allow it to escape.
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
Numerous fixes 0.4.0.5 Beta 07/18/2014 --User Changes Allow for vibrancy values > 1. Add flatten and unflatten menu items. Automatically flatten like Apophysis does. Add plugin and new_linear tags to Xml to be compatible with Apophysis. --Bug Fixes Fix blur, blur3d, bubble, cropn, cross, curl, curl3d, epispiral, ho, julia3d, julia3dz, loonie, mirror_x, mirror_y, mirror_z, rotate_x, sinusoidal, spherical, spherical3d, stripes. Unique filename on final render was completely broken. Two severe OpenCL bugs. Random seeds were biased and fusing was being reset too often leading to results that differ from the CPU. Subtle, but sometimes severe bug in the setup of the xaos weights. Use properly defined epsilon by getting the value from std::numeric_limits, rather than hard coding 1e-6 or 1e-10. Omit incorrect usage of epsilon everywhere. It should not be automatically added to denominators. Rather, it should only be used if the denominator is zero. Force final render progress bars to 100 on completion. Sometimes they didn't seem to make it there. Make variation name and params comparisons be case insensitive. --Code Changes Make ForEach and FindIf wrappers around std::for_each and std::find_if.
2014-07-19 02:33:18 -04:00
randiter++;
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
randInt = fmod((randInt * multiplier + offset), modBase);
}
}
tileType = fmod(randInt, T(2));
}
//Drawing the points.
if (extended == 0)//Fast drawmode
{
if (tileType < 1)
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
{
--User changes -Add post_smartcrop. --Bug fixes -Fix bug in crackle. -Wrong point assignment in hexaplay3D, hexnix3D. -Improper Z assignment in rblur. -Fix inconsistency with original in circlecrop. -Put EMBER_ROOT bakc to ./../../../ in default.pri. This is TBD. --Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums. -Optimize crob, nBlur. -Fix weird assignment statement in falloff3. -Cleanup in VarFuncs::SimplexNoise3D(). -Replace fabs() with std::abs(). -General cleanup.
2016-01-12 23:42:12 -05:00
r0 = std::pow((pow(std::abs(x), m_Exponent) + std::pow(std::abs(y), m_Exponent)), m_OneOverEx);
r1 = std::pow((pow(std::abs(x - 1), m_Exponent) + std::pow(std::abs(y - 1), m_Exponent)), m_OneOverEx);
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
}
else
{
--User changes -Add post_smartcrop. --Bug fixes -Fix bug in crackle. -Wrong point assignment in hexaplay3D, hexnix3D. -Improper Z assignment in rblur. -Fix inconsistency with original in circlecrop. -Put EMBER_ROOT bakc to ./../../../ in default.pri. This is TBD. --Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums. -Optimize crob, nBlur. -Fix weird assignment statement in falloff3. -Cleanup in VarFuncs::SimplexNoise3D(). -Replace fabs() with std::abs(). -General cleanup.
2016-01-12 23:42:12 -05:00
r0 = std::pow((pow(std::abs(x - 1), m_Exponent) + std::pow(std::abs(y), m_Exponent)), m_OneOverEx);
r1 = std::pow((pow(std::abs(x), m_Exponent) + std::pow(std::abs(y - 1), m_Exponent)), m_OneOverEx);
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
}
}
else//Slow drawmode
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
if (tileType == 1)
{
--User changes -Add post_smartcrop. --Bug fixes -Fix bug in crackle. -Wrong point assignment in hexaplay3D, hexnix3D. -Improper Z assignment in rblur. -Fix inconsistency with original in circlecrop. -Put EMBER_ROOT bakc to ./../../../ in default.pri. This is TBD. --Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums. -Optimize crob, nBlur. -Fix weird assignment statement in falloff3. -Cleanup in VarFuncs::SimplexNoise3D(). -Replace fabs() with std::abs(). -General cleanup.
2016-01-12 23:42:12 -05:00
r0 = std::pow((std::pow(std::abs(x), m_Exponent) + std::pow(std::abs(y), m_Exponent)), m_OneOverEx);
r1 = std::pow((std::pow(std::abs(x - 1), m_Exponent) + std::pow(std::abs(y - 1), m_Exponent)), m_OneOverEx);
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
}
else
{
--User changes -Add post_smartcrop. --Bug fixes -Fix bug in crackle. -Wrong point assignment in hexaplay3D, hexnix3D. -Improper Z assignment in rblur. -Fix inconsistency with original in circlecrop. -Put EMBER_ROOT bakc to ./../../../ in default.pri. This is TBD. --Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums. -Optimize crob, nBlur. -Fix weird assignment statement in falloff3. -Cleanup in VarFuncs::SimplexNoise3D(). -Replace fabs() with std::abs(). -General cleanup.
2016-01-12 23:42:12 -05:00
r0 = std::pow((std::pow(std::abs(x - 1), m_Exponent) + std::pow(std::abs(y), m_Exponent)), m_OneOverEx);
r1 = std::pow((std::pow(std::abs(x), m_Exponent) + std::pow(std::abs(y - 1), m_Exponent)), m_OneOverEx);
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
}
}
--User changes -Add post_smartcrop. --Bug fixes -Fix bug in crackle. -Wrong point assignment in hexaplay3D, hexnix3D. -Improper Z assignment in rblur. -Fix inconsistency with original in circlecrop. -Put EMBER_ROOT bakc to ./../../../ in default.pri. This is TBD. --Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums. -Optimize crob, nBlur. -Fix weird assignment statement in falloff3. -Cleanup in VarFuncs::SimplexNoise3D(). -Replace fabs() with std::abs(). -General cleanup.
2016-01-12 23:42:12 -05:00
r = std::abs(r0 - T(0.5)) * m_OneOverRmax;
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
if (r < 1)
{
helper.Out.x = m_Size * (x + Floor<T>(helper.In.x));
helper.Out.y = m_Size * (y + Floor<T>(helper.In.y));
}
0.4.0.8 Beta 07/26/2014 0.4.0.8 Beta 07/26/2014 --Bug Fixes Fix falloff, falloff2, falloff3.
2014-07-26 20:26:15 -04:00
else
{
helper.Out.x = 0;//Needed because of possible sum below.
helper.Out.y = 0;
}
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
--User changes -Add post_smartcrop. --Bug fixes -Fix bug in crackle. -Wrong point assignment in hexaplay3D, hexnix3D. -Improper Z assignment in rblur. -Fix inconsistency with original in circlecrop. -Put EMBER_ROOT bakc to ./../../../ in default.pri. This is TBD. --Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums. -Optimize crob, nBlur. -Fix weird assignment statement in falloff3. -Cleanup in VarFuncs::SimplexNoise3D(). -Replace fabs() with std::abs(). -General cleanup.
2016-01-12 23:42:12 -05:00
r = std::abs(r1 - T(0.5)) * m_OneOverRmax;
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
if (r < 1)
{
helper.Out.x += m_Size * (x + Floor<T>(helper.In.x));//The += is intended here.
helper.Out.y += m_Size * (y + Floor<T>(helper.In.y));
}
Numerous fixes 0.4.0.5 Beta 07/18/2014 --User Changes Allow for vibrancy values > 1. Add flatten and unflatten menu items. Automatically flatten like Apophysis does. Add plugin and new_linear tags to Xml to be compatible with Apophysis. --Bug Fixes Fix blur, blur3d, bubble, cropn, cross, curl, curl3d, epispiral, ho, julia3d, julia3dz, loonie, mirror_x, mirror_y, mirror_z, rotate_x, sinusoidal, spherical, spherical3d, stripes. Unique filename on final render was completely broken. Two severe OpenCL bugs. Random seeds were biased and fusing was being reset too often leading to results that differ from the CPU. Subtle, but sometimes severe bug in the setup of the xaos weights. Use properly defined epsilon by getting the value from std::numeric_limits, rather than hard coding 1e-6 or 1e-10. Omit incorrect usage of epsilon everywhere. It should not be automatically added to denominators. Rather, it should only be used if the denominator is zero. Force final render progress bars to 100 on completion. Sometimes they didn't seem to make it there. Make variation name and params comparisons be case insensitive. --Code Changes Make ForEach and FindIf wrappers around std::for_each and std::find_if.
2014-07-19 02:33:18 -04:00
--User changes -Allow for pausing the renderer in the main window. This makes is more efficient when entering many parameters, such as when following a tutorial. -Add support for new variations: erf, gamma, jac_cn, jac_dn, jac_sn, logDB, pressure_wave, pRose3D, splits3D, w, waves2b, x, xerf, y, z. -Inform user of the start and stop of file parsing in EmberAnimate because the files could potentially be very large. -Move the follwing fields to a new table called Animation: Interpolation, Affine Interpolation, Temporal Samples, Temporal Filter Width, Temporal Filter Type. -These currently have no effect on the interactive renderer and instead are used when running flames through EmberGenome to generate sequences, and then animating them in Fractorium or EmberAnimate. -Add new parameter overrides for EmberRender and EmberAnimate which directly assign values to all flames being rendered, rather than scale: --quality --demin --demax --Bug fixes -Left pad instead of right pad names of sequence outputs from EmberGenome. -Unique file naming was broken for files which already had an underscore in them. -Properly report that png is the default format of EmberRender and EmberAnimate output instead of erroneously claiming it was jpg. -Make command line programs search these folders in this order for the palette file: ./ ~/.fractorium ~/.config/fractorium /usr/share/fractorium /usr/local/share/fractorium -Fix possible bad values in hexes. -Better assignment of Z variables. -Fix boarders due to use of poorly implemented rint() function from flam3. Use std::rint() now. -wedge_sph was completely wrong due to having accidentally swapped the mapping of two parameters. -Make juliascope juliascope_power parameter be of type REAL_NONZERO since it's used as a denominator. -Make Z assignment compatible with the originals in: -arch, bcircle, bCollide, bent, bent2, bisplit, blob, blur_linear, blur_square, bMod, boarders, boarders2, bSwirl, bTransform, butterfly, cardioid, cell, circleblur, circlize, circlize2, circus, collideoscope, cos, cosine, cosh, coth, cpow, cpow2, crescents, cropn, csc, csch, curl, curve, dc_gridout, deltaa, diamond, disc2, eclipse, eCollide, edisc, eJulia, elliptic, eMod, eMotion, ennepers, epispiral, ePush, eRotate, eScale, eSwirl, ex, exp, expo, exponential, fan, fdisc, fibonacci, fibonacci2, fisheye, flipcircle, flipy, flower, flux, funnel, glynnia, GlynnSim1, GlynnSim2, GlynnSim3, gridout, handkerchief, heart, hole, idisc, julia, julian2, juliaNab, kaleidoscope, lazyTravis, Lissajous, mask, MobiusN, mobius_strip, modulus, murl, murl2, npolar, ortho, oscilloscope, parabola, perspective, petal, phoenix_julia, pie (was also inconsistent between cpu and gpu), poincare, popcorn, popcorn2, power, pow_block, rational3, rays, rblur, rings, rippled, roundspher, sec, secant2, sigmoid, sin, sineblur, sinh, sinusgrid, sphericaln, spiralwing, spirograph, split, squarize, squirrel, squish, sschecks, starblur, stripes, stwin, super_shape, tan, tancos, tangent, tanh, TwinTrian, twoface, unpolar, waves, wavesn, wedge_julia, whorl, xheart, zblur, zscale. --Code changes -Generalize Variation::PrecalcHelper() and rename to PrePostPrecalcHelper(). --Do the same for the OpenCL version and rename it PrePostPrecalcOpenCLString(). -Rename Variation::m_AssignType to m_PrePostAssignType since it's only relevant to pre/post variations.
2016-01-29 20:02:15 -05:00
helper.Out.z = DefaultZ(helper);
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
}
--Code changes -Better use of const functions. -Ember.pro should not define BUILDING_EMBERCL, instead define BUILDING_EMBER.
2015-06-29 23:13:53 -04:00
virtual string OpenCLString() const override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
ostringstream ss, ss2;
More C++11
2014-09-10 01:41:26 -04:00
int i = 0;
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
ss2 << "_" << XformIndexInEmber() << "]";
string index = ss2.str();
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
string weight = WeightDefineString();
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
string extended = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string exponent = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string arcWidth = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string rotation = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string size = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string seed = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string oneOverEx = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string absSeed = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string seed2 = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
string oneOverRmax = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
string scale = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
ss << "\t{\n"
<< "\t\tint extended = (int)" << extended << ";\n"
<< "\t\treal_t seed = " << absSeed << ";\n"
<< "\t\treal_t r = -" << rotation << ";\n"
<< "\t\treal_t r0 = 0;\n"
<< "\t\treal_t r1 = 0;\n"
<< "\t\treal_t tileType = 0;\n"
<< "\t\treal_t randInt = 0;\n"
<< "\t\treal_t modBase = 65535;\n"
<< "\t\treal_t multiplier = 32747;\n"
<< "\t\treal_t offset = 12345;\n"
<< "\t\treal_t niter = 0;\n"
<< "\t\treal_t x = vIn.x * " << scale << ";\n"
<< "\t\treal_t y = vIn.y * " << scale << ";\n"
<< "\t\tint intx = (int)Round(x);\n"
<< "\t\tint inty = (int)Round(y);\n"
<< "\t\tint randiter;\n"
<< "\n"
<< "\t\tr = x - intx;\n"
<< "\n"
<< "\t\tif (r < 0)\n"
<< "\t\t x = 1 + r;\n"
<< "\t\telse\n"
<< "\t\t x = r;\n"
<< "\n"
<< "\t\tr = y - inty;\n"
<< "\n"
<< "\t\tif (r < 0)\n"
<< "\t\t y = 1 + r;\n"
<< "\t\telse\n"
<< "\t\t y = r;\n"
<< "\n"
<< "\t\tif (seed == 0)\n"
<< "\t\t tileType = 0;\n"
<< "\t\telse if (seed == 1)\n"
<< "\t\t tileType = 1;\n"
<< "\t\telse\n"
<< "\t\t{\n"
<< "\t\t if (extended == 0)\n"
<< "\t\t {\n"
<< "\t\t real_t xrand = Round(vIn.x);\n"
<< "\t\t real_t yrand = Round(vIn.y);\n"
<< "\n"
<< "\t\t xrand = xrand * " << seed2 << ";\n"
<< "\t\t yrand = yrand * " << seed2 << ";\n"
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
<< "\t\t niter = fma(xrand, yrand, xrand + yrand);\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t\t randInt = (niter + seed) * " << seed2 << " / 2;\n"
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
<< "\t\t randInt = fmod(fma(randInt, multiplier, offset), modBase);\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t\t }\n"
<< "\t\t else\n"
<< "\t\t {\n"
<< "\t\t int xrand = (int)Round(vIn.x);\n"
<< "\t\t int yrand = (int)Round(vIn.y);\n"
<< "\n"
<< "\t\t seed = floor(seed);\n"
<< "\t\t niter = (real_t)abs(xrand + yrand + xrand * yrand);\n"
<< "\t\t randInt = seed + niter;\n"
<< "\t\t randiter = 0;\n"
<< "\n"
Numerous fixes 0.4.0.5 Beta 07/18/2014 --User Changes Allow for vibrancy values > 1. Add flatten and unflatten menu items. Automatically flatten like Apophysis does. Add plugin and new_linear tags to Xml to be compatible with Apophysis. --Bug Fixes Fix blur, blur3d, bubble, cropn, cross, curl, curl3d, epispiral, ho, julia3d, julia3dz, loonie, mirror_x, mirror_y, mirror_z, rotate_x, sinusoidal, spherical, spherical3d, stripes. Unique filename on final render was completely broken. Two severe OpenCL bugs. Random seeds were biased and fusing was being reset too often leading to results that differ from the CPU. Subtle, but sometimes severe bug in the setup of the xaos weights. Use properly defined epsilon by getting the value from std::numeric_limits, rather than hard coding 1e-6 or 1e-10. Omit incorrect usage of epsilon everywhere. It should not be automatically added to denominators. Rather, it should only be used if the denominator is zero. Force final render progress bars to 100 on completion. Sometimes they didn't seem to make it there. Make variation name and params comparisons be case insensitive. --Code Changes Make ForEach and FindIf wrappers around std::for_each and std::find_if.
2014-07-19 02:33:18 -04:00
<< "\t\t while (randiter < niter && randiter < 20)\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t\t {\n"
Numerous fixes 0.4.0.5 Beta 07/18/2014 --User Changes Allow for vibrancy values > 1. Add flatten and unflatten menu items. Automatically flatten like Apophysis does. Add plugin and new_linear tags to Xml to be compatible with Apophysis. --Bug Fixes Fix blur, blur3d, bubble, cropn, cross, curl, curl3d, epispiral, ho, julia3d, julia3dz, loonie, mirror_x, mirror_y, mirror_z, rotate_x, sinusoidal, spherical, spherical3d, stripes. Unique filename on final render was completely broken. Two severe OpenCL bugs. Random seeds were biased and fusing was being reset too often leading to results that differ from the CPU. Subtle, but sometimes severe bug in the setup of the xaos weights. Use properly defined epsilon by getting the value from std::numeric_limits, rather than hard coding 1e-6 or 1e-10. Omit incorrect usage of epsilon everywhere. It should not be automatically added to denominators. Rather, it should only be used if the denominator is zero. Force final render progress bars to 100 on completion. Sometimes they didn't seem to make it there. Make variation name and params comparisons be case insensitive. --Code Changes Make ForEach and FindIf wrappers around std::for_each and std::find_if.
2014-07-19 02:33:18 -04:00
<< "\t\t randiter++;\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t\t randInt = fmod((randInt * multiplier + offset), modBase);\n"
<< "\t\t }\n"
<< "\t\t }\n"
<< "\n"
<< "\t\t tileType = fmod(randInt, 2);\n"
<< "\t\t}\n"
<< "\n"
<< "\t\tif (extended == 0)\n"
<< "\t\t{\n"
<< "\t\t if (tileType < 1)\n"
<< "\t\t {\n"
<< "\t\t r0 = pow((pow(fabs(x ), " << exponent << ") + pow(fabs(y ), " << exponent << ")), " << oneOverEx << ");\n"
<< "\t\t r1 = pow((pow(fabs(x - 1), " << exponent << ") + pow(fabs(y - 1), " << exponent << ")), " << oneOverEx << ");\n"
<< "\t\t }\n"
<< "\t\t else\n"
<< "\t\t {\n"
<< "\t\t r0 = pow((pow(fabs(x - 1), " << exponent << ") + pow(fabs(y ), " << exponent << ")), " << oneOverEx << ");\n"
<< "\t\t r1 = pow((pow(fabs(x ), " << exponent << ") + pow(fabs(y - 1), " << exponent << ")), " << oneOverEx << ");\n"
<< "\t\t }\n"
<< "\t\t}\n"
<< "\t\telse\n"
<< "\t\t{\n"
<< "\t\t if (tileType == 1)\n"
<< "\t\t {\n"
<< "\t\t r0 = pow((pow(fabs(x ), " << exponent << ") + pow(fabs(y ), " << exponent << ")), " << oneOverEx << ");\n"
<< "\t\t r1 = pow((pow(fabs(x - 1), " << exponent << ") + pow(fabs(y - 1), " << exponent << ")), " << oneOverEx << ");\n"
<< "\t\t }\n"
<< "\t\t else\n"
<< "\t\t {\n"
<< "\t\t r0 = pow((pow(fabs(x - 1), " << exponent << ") + pow(fabs(y ), " << exponent << ")), " << oneOverEx << ");\n"
<< "\t\t r1 = pow((pow(fabs(x ), " << exponent << ") + pow(fabs(y - 1), " << exponent << ")), " << oneOverEx << ");\n"
<< "\t\t }\n"
<< "\t\t}\n"
<< "\n"
Ditch T(f) define in opencl programs in preference to already-defined real_t. Also removed some erroneously-added double-'T(' statements
2015-02-26 16:18:55 -05:00
<< "\t\tr = fabs(r0 - (real_t)(0.5)) * " << oneOverRmax << ";\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\n"
<< "\t\tif (r < 1)\n"
<< "\t\t{\n"
<< "\t\t vOut.x = " << size << " * (x + floor(vIn.x));\n"
<< "\t\t vOut.y = " << size << " * (y + floor(vIn.y));\n"
<< "\t\t}\n"
0.4.0.8 Beta 07/26/2014 0.4.0.8 Beta 07/26/2014 --Bug Fixes Fix falloff, falloff2, falloff3.
2014-07-26 20:26:15 -04:00
<< "\t\telse\n"
<< "\t\t{\n"
Ditch T(f) define in opencl programs in preference to already-defined real_t. Also removed some erroneously-added double-'T(' statements
2015-02-26 16:18:55 -05:00
<< "\t\t vOut.x = (real_t)(0.0);\n"
<< "\t\t vOut.y = (real_t)(0.0);\n"
0.4.0.8 Beta 07/26/2014 0.4.0.8 Beta 07/26/2014 --Bug Fixes Fix falloff, falloff2, falloff3.
2014-07-26 20:26:15 -04:00
<< "\t\t}\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\n"
Ditch T(f) define in opencl programs in preference to already-defined real_t. Also removed some erroneously-added double-'T(' statements
2015-02-26 16:18:55 -05:00
<< "\t\tr = fabs(r1 - (real_t)(0.5)) * " << oneOverRmax << ";\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\n"
<< "\t\tif (r < 1)\n"
<< "\t\t{\n"
<< "\t\t vOut.x += " << size << " * (x + floor(vIn.x));\n"
<< "\t\t vOut.y += " << size << " * (y + floor(vIn.y));\n"
<< "\t\t}\n"
Numerous fixes 0.4.0.5 Beta 07/18/2014 --User Changes Allow for vibrancy values > 1. Add flatten and unflatten menu items. Automatically flatten like Apophysis does. Add plugin and new_linear tags to Xml to be compatible with Apophysis. --Bug Fixes Fix blur, blur3d, bubble, cropn, cross, curl, curl3d, epispiral, ho, julia3d, julia3dz, loonie, mirror_x, mirror_y, mirror_z, rotate_x, sinusoidal, spherical, spherical3d, stripes. Unique filename on final render was completely broken. Two severe OpenCL bugs. Random seeds were biased and fusing was being reset too often leading to results that differ from the CPU. Subtle, but sometimes severe bug in the setup of the xaos weights. Use properly defined epsilon by getting the value from std::numeric_limits, rather than hard coding 1e-6 or 1e-10. Omit incorrect usage of epsilon everywhere. It should not be automatically added to denominators. Rather, it should only be used if the denominator is zero. Force final render progress bars to 100 on completion. Sometimes they didn't seem to make it there. Make variation name and params comparisons be case insensitive. --Code Changes Make ForEach and FindIf wrappers around std::for_each and std::find_if.
2014-07-19 02:33:18 -04:00
<< "\n"
--User changes -Allow for pausing the renderer in the main window. This makes is more efficient when entering many parameters, such as when following a tutorial. -Add support for new variations: erf, gamma, jac_cn, jac_dn, jac_sn, logDB, pressure_wave, pRose3D, splits3D, w, waves2b, x, xerf, y, z. -Inform user of the start and stop of file parsing in EmberAnimate because the files could potentially be very large. -Move the follwing fields to a new table called Animation: Interpolation, Affine Interpolation, Temporal Samples, Temporal Filter Width, Temporal Filter Type. -These currently have no effect on the interactive renderer and instead are used when running flames through EmberGenome to generate sequences, and then animating them in Fractorium or EmberAnimate. -Add new parameter overrides for EmberRender and EmberAnimate which directly assign values to all flames being rendered, rather than scale: --quality --demin --demax --Bug fixes -Left pad instead of right pad names of sequence outputs from EmberGenome. -Unique file naming was broken for files which already had an underscore in them. -Properly report that png is the default format of EmberRender and EmberAnimate output instead of erroneously claiming it was jpg. -Make command line programs search these folders in this order for the palette file: ./ ~/.fractorium ~/.config/fractorium /usr/share/fractorium /usr/local/share/fractorium -Fix possible bad values in hexes. -Better assignment of Z variables. -Fix boarders due to use of poorly implemented rint() function from flam3. Use std::rint() now. -wedge_sph was completely wrong due to having accidentally swapped the mapping of two parameters. -Make juliascope juliascope_power parameter be of type REAL_NONZERO since it's used as a denominator. -Make Z assignment compatible with the originals in: -arch, bcircle, bCollide, bent, bent2, bisplit, blob, blur_linear, blur_square, bMod, boarders, boarders2, bSwirl, bTransform, butterfly, cardioid, cell, circleblur, circlize, circlize2, circus, collideoscope, cos, cosine, cosh, coth, cpow, cpow2, crescents, cropn, csc, csch, curl, curve, dc_gridout, deltaa, diamond, disc2, eclipse, eCollide, edisc, eJulia, elliptic, eMod, eMotion, ennepers, epispiral, ePush, eRotate, eScale, eSwirl, ex, exp, expo, exponential, fan, fdisc, fibonacci, fibonacci2, fisheye, flipcircle, flipy, flower, flux, funnel, glynnia, GlynnSim1, GlynnSim2, GlynnSim3, gridout, handkerchief, heart, hole, idisc, julia, julian2, juliaNab, kaleidoscope, lazyTravis, Lissajous, mask, MobiusN, mobius_strip, modulus, murl, murl2, npolar, ortho, oscilloscope, parabola, perspective, petal, phoenix_julia, pie (was also inconsistent between cpu and gpu), poincare, popcorn, popcorn2, power, pow_block, rational3, rays, rblur, rings, rippled, roundspher, sec, secant2, sigmoid, sin, sineblur, sinh, sinusgrid, sphericaln, spiralwing, spirograph, split, squarize, squirrel, squish, sschecks, starblur, stripes, stwin, super_shape, tan, tancos, tangent, tanh, TwinTrian, twoface, unpolar, waves, wavesn, wedge_julia, whorl, xheart, zblur, zscale. --Code changes -Generalize Variation::PrecalcHelper() and rename to PrePostPrecalcHelper(). --Do the same for the OpenCL version and rename it PrePostPrecalcOpenCLString(). -Rename Variation::m_AssignType to m_PrePostAssignType since it's only relevant to pre/post variations.
2016-01-29 20:02:15 -05:00
<< "\t\tvOut.z = " << DefaultZCl()
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t}\n";
return ss.str();
}
--User changes -Add new variations: bubbleT3D, crob, hexaplay3D, hexcrop, hexes, hexnix3D, loonie2, loonie3, nBlur, octapol and synth. -Allow for pre/post versions of dc_bubble, dc_cylinder and dc_linear whereas before they were omitted. -When saving a file with multiple embers in it, detect if time values are all the same and if so, start them at zero and increment by 1 for each ember. -Allow for numerous quality increases to be coalesced into one. It will pick up at the end of the current render. -Show selection highlight on variations tree in response to mouse hover. This makes it easier to see for which variation or param the current mouse wheel action will apply. -Make default temporal samples be 100, whereas before it was 1000 which was overkill. -Require the shift key to be held with delete for deleting an ember to prevent it from triggering when the user enters delete in the edit box. -This wasn't otherwise fixable without writing a lot more code. --Bug fixes -EmberGenome was crashing when generating a sequence from a source file with more than 2 embers in it. -EmberGenome was improperly handling the first frame of a merge after the last frame of the loop. -These bugs were due to a previous commit. Revert parts of that commit. -Prevent a zoom value of less than 0 when reading from xml. -Slight optimization of the crescents, and mask variations, if the compiler wasn't doing it already. -Unique file naming was broken because it was looking for _# and the default names ended with -#. -Disallow renaming of an ember in the library tree to an empty string. -Severe bug that prevented some variations from being read correctly from params generated outside this program. -Severe OpenCL randomization bug. The first x coordinates of the first points in the first kernel call of the first ember of a render since the OpenCL renderer object was created were not random and were mostly -1. -Severe bug when populating xform selection distributions that could sometimes cause a crash due to roundoff error. Fix by using double. -Limit the max number of variations in a random ember to MAX_CL_VARS, which is 8. This ensures they'll look the same on CPU and GPU. -Prevent user from saving stylesheet to default.qss, it's a special reserved filename. --Code changes -Generalize using the running sum output point inside of a variation for all cases: pre, reg and post. -Allow for array variables in variations where the address of each element is stored in m_Params. -Qualify all math functions with std:: -No longer use our own Clamp() in OpenCL, instead use the standard clamp(). -Redesign how functions are used in the variations OpenCL code. -Add tests to EmberTester to verify some of the new functionality. -Place more const and override qualifiers on functions where appropriate. -Add a global rand with a lock to be used very sparingly. -Use a map instead of a vector for bad param names in Xml parsing. -Prefix affine interpolation mode defines with "AFFINE_" to make their purpose more clear. -Allow for variations that change state during iteration by sending a separate copy of the ember to each rendering thread. -Implement this same functionality with a local struct in OpenCL. It's members are the total of all variables that need to change state within an ember. -Add Contains() function to Utils.h. -EmberRender: print names of kernels being printed with --dump_kernel option. -Clean up EmberTester to handle some of the recent changes. -Fix various casts. -Replace % 2 with & 1, even though the compiler was likely doing this already. -Add new file Variations06.h to accommodate new variations. -General cleanup.
2015-11-22 17:15:07 -05:00
virtual vector<string> OpenCLGlobalFuncNames() const override
{
return vector<string> { "Round" };
}
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
virtual void Precalc() override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
m_OneOverEx = 1 / m_Exponent;
--User changes -Add post_smartcrop. --Bug fixes -Fix bug in crackle. -Wrong point assignment in hexaplay3D, hexnix3D. -Improper Z assignment in rblur. -Fix inconsistency with original in circlecrop. -Put EMBER_ROOT bakc to ./../../../ in default.pri. This is TBD. --Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums. -Optimize crob, nBlur. -Fix weird assignment statement in falloff3. -Cleanup in VarFuncs::SimplexNoise3D(). -Replace fabs() with std::abs(). -General cleanup.
2016-01-12 23:42:12 -05:00
m_AbsSeed = std::abs(m_Seed);
Explicitly namespace calls to sqrt and isnan osx/linux have alternative top-level implementations of these methods, so the calls are ambiguous given the - so these should be used explicitly
2015-04-13 07:32:58 -04:00
m_Seed2 = std::sqrt(Zeps(m_AbsSeed + (m_AbsSeed / 2))) / Zeps((m_AbsSeed * T(0.5))) * T(0.25);
--User changes -Add new variations: bubbleT3D, crob, hexaplay3D, hexcrop, hexes, hexnix3D, loonie2, loonie3, nBlur, octapol and synth. -Allow for pre/post versions of dc_bubble, dc_cylinder and dc_linear whereas before they were omitted. -When saving a file with multiple embers in it, detect if time values are all the same and if so, start them at zero and increment by 1 for each ember. -Allow for numerous quality increases to be coalesced into one. It will pick up at the end of the current render. -Show selection highlight on variations tree in response to mouse hover. This makes it easier to see for which variation or param the current mouse wheel action will apply. -Make default temporal samples be 100, whereas before it was 1000 which was overkill. -Require the shift key to be held with delete for deleting an ember to prevent it from triggering when the user enters delete in the edit box. -This wasn't otherwise fixable without writing a lot more code. --Bug fixes -EmberGenome was crashing when generating a sequence from a source file with more than 2 embers in it. -EmberGenome was improperly handling the first frame of a merge after the last frame of the loop. -These bugs were due to a previous commit. Revert parts of that commit. -Prevent a zoom value of less than 0 when reading from xml. -Slight optimization of the crescents, and mask variations, if the compiler wasn't doing it already. -Unique file naming was broken because it was looking for _# and the default names ended with -#. -Disallow renaming of an ember in the library tree to an empty string. -Severe bug that prevented some variations from being read correctly from params generated outside this program. -Severe OpenCL randomization bug. The first x coordinates of the first points in the first kernel call of the first ember of a render since the OpenCL renderer object was created were not random and were mostly -1. -Severe bug when populating xform selection distributions that could sometimes cause a crash due to roundoff error. Fix by using double. -Limit the max number of variations in a random ember to MAX_CL_VARS, which is 8. This ensures they'll look the same on CPU and GPU. -Prevent user from saving stylesheet to default.qss, it's a special reserved filename. --Code changes -Generalize using the running sum output point inside of a variation for all cases: pre, reg and post. -Allow for array variables in variations where the address of each element is stored in m_Params. -Qualify all math functions with std:: -No longer use our own Clamp() in OpenCL, instead use the standard clamp(). -Redesign how functions are used in the variations OpenCL code. -Add tests to EmberTester to verify some of the new functionality. -Place more const and override qualifiers on functions where appropriate. -Add a global rand with a lock to be used very sparingly. -Use a map instead of a vector for bad param names in Xml parsing. -Prefix affine interpolation mode defines with "AFFINE_" to make their purpose more clear. -Allow for variations that change state during iteration by sending a separate copy of the ember to each rendering thread. -Implement this same functionality with a local struct in OpenCL. It's members are the total of all variables that need to change state within an ember. -Add Contains() function to Utils.h. -EmberRender: print names of kernels being printed with --dump_kernel option. -Clean up EmberTester to handle some of the recent changes. -Fix various casts. -Replace % 2 with & 1, even though the compiler was likely doing this already. -Add new file Variations06.h to accommodate new variations. -General cleanup.
2015-11-22 17:15:07 -05:00
m_OneOverRmax = 1 / (T(0.5) * (std::pow(T(2), 1 / m_Exponent) - 1) * m_ArcWidth);
m_Scale = (std::cos(-m_Rotation) - std::sin(-m_Rotation)) / m_Weight;
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
}
protected:
void Init()
{
string prefix = Prefix();
m_Params.clear();
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
m_Params.push_back(ParamWithName<T>(&m_Extended, prefix + "Truchet_extended", 0, eParamType::INTEGER, 0, 1));
m_Params.push_back(ParamWithName<T>(&m_Exponent, prefix + "Truchet_exponent", 2, eParamType::REAL_CYCLIC, T(0.001), 2));
--Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums.
2016-01-04 19:50:15 -05:00
m_Params.push_back(ParamWithName<T>(&m_ArcWidth, prefix + "Truchet_arc_width", T(0.5), eParamType::REAL_CYCLIC, T(0.001), 1));
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
m_Params.push_back(ParamWithName<T>(&m_Rotation, prefix + "Truchet_rotation"));
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
m_Params.push_back(ParamWithName<T>(&m_Size, prefix + "Truchet_size", 1, eParamType::REAL_CYCLIC, T(0.001), 10));
m_Params.push_back(ParamWithName<T>(&m_Seed, prefix + "Truchet_seed", 50));
m_Params.push_back(ParamWithName<T>(true, &m_OneOverEx, prefix + "Truchet_one_over_ex"));//Precalc.
m_Params.push_back(ParamWithName<T>(true, &m_AbsSeed, prefix + "Truchet_abs_seed"));
m_Params.push_back(ParamWithName<T>(true, &m_Seed2, prefix + "Truchet_seed2"));
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
m_Params.push_back(ParamWithName<T>(true, &m_OneOverRmax, prefix + "Truchet_one_over_rmax"));
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
m_Params.push_back(ParamWithName<T>(true, &m_Scale, prefix + "Truchet_scale"));
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
}
private:
T m_Extended;
T m_Exponent;
T m_ArcWidth;
T m_Rotation;
T m_Size;
T m_Seed;
T m_OneOverEx;//Precalc.
T m_AbsSeed;
T m_Seed2;
T m_OneOverRmax;
T m_Scale;
};
--User changes: -Show common folder locations such as documents, downloads, pictures in the sidebar in all file dialogs. -Warning message about exceeding memory in final render dialog now suggests strips as the solution to the problem. -Strips now has a tooltip explaining what it does. -Allow more digits in the spinners on the color section the flame tab. -Add manually adjustable size spinners in the final render dialog. Percentage scale and absolute size are fully synced. -Default prefix in final render is now the filename when doing animations (coming from sequence section of the library tab). -Changed the elliptic variation back to using a less precise version for float, and a more precise version for double. The last release had it always using double. -New applied xaos table that shows a read-only view of actual weights by taking the base xform weights and multiplying them by the xaos values. -New table in the xaos tab that gives a graphical representation of the probability that each xform is chosen, with and without xaos. -Add button to transpose the xaos rows and columns. -Add support for importing .chaos files from Chaotica. --Pasting back to Chaotica will work for most, but not all, variations due to incompatible parameter names in some. -Curves are now splines instead of Bezier. This adds compatibility with Chaotica, but breaks it for Apophysis. Xmls are still pastable, but the color curves will look different. --The curve editor on the palette tab can now add points by clicking on the lines and remove points by clicking on the points themselves, just like Chaotica. --Splines are saved in four new xml fields: overall_curve, red_curve, green_curve and blue_curve. -Allow for specifying the percentage of a sub batch each thread should iterate through per kernel call when running with OpenCL. This gives a roughly 1% performance increase due to having to make less kernel calls while iterating. --This field is present for interactive editing (where it's not very useful) and in the final render dialog. --On the command line, this is specified as --sbpctth for EmberRender and EmberAnimate. -Allow double clicking to toggle the supersample field in the flame tab between 1 and 2 for easily checking the effect of the field. -When showing affine values as polar coordinates, show angles normalized to 360 to match Chaotica. -Fuse Count spinner now toggles between 15 and 100 when double clicking for easily checking the effect of the field. -Added field for limiting the range in the x and y direction that the initial points are chosen from. -Added a field called K2 which is an alternative way to set brightness, ignored when zero. --This has no effect for many variations, but hs a noticeable effect for some. -Added new variations: arcsech arcsech2 arcsinh arctanh asteria block bwraps_rand circlecrop2 coth_spiral crackle2 depth_blur depth_blur2 depth_gaussian depth_gaussian2 depth_ngon depth_ngon2 depth_sine depth_sine2 dragonfire dspherical dust excinis exp2 flipx flowerdb foci_p gaussian glynnia2 glynnsim4 glynnsim5 henon henon hex_rand hex_truchet hypershift lazyjess lens lozi lozi modulusx modulusy oscilloscope2 point_symmetry pointsymmetry projective pulse rotate scry2 shift smartshape spher squares starblur2 swirl3 swirl3r tanh_spiral target0 target2 tile_hlp truchet_glyph truchet_inv truchet_knot unicorngaloshen vibration vibration2 --hex_truchet, hex_rand should always use double. They are extremely sensitive. --Bug fixes: -Bounds sign was flipped for x coordinate of world space when center was not zero. -Right clicking and dragging spinner showed menu on mouse up, even if it was very far away. -Text boxes for size in final render dialog were hard to type in. Same bug as xform weight used to be so fix the same way. -Fix spelling to be plural in toggle color speed box. -Stop using the blank user palette to generate flames. Either put colored palettes in it, or exclude it from randoms. -Clicking the random palette button for a palette file with only one palette in it would freeze the program. -Clicking none scale in final render did not re-render the preview. -Use less precision on random xaos. No need for 12 decimal places. -The term sub batch is overloaded in the options dialog. Change the naming and tooltip of those settings for cpu and opencl. --Also made clear in the tooltip for the default opencl quality setting that the value is per device. -The arrows spinner in palette editor appears like a read-only label. Made it look like a spinner. -Fix border colors for various spin boxes and table headers in the style sheet. Requires reload. -Fix a bug in the bwraps variation which would produce different results than Chaotica and Apophysis. -Synth was allowed to be selected for random flame generation when using an Nvidia card but it shouldn't have been because Nvidia has a hard time compiling synth. -A casting bug in the OpenCL kernels for log scaling and density filtering was preventing successful compilations on Intel iGPUs. Fixed even though we don't support anything other than AMD and Nvidia. -Palette rotation (click and drag) position was not being reset when loading a new flame. -When the xform circles were hidden, opening and closing the options dialog would improperly reshow them. -Double click toggle was broken on integer spin boxes. -Fixed tab order of some controls. -Creating a palette from a jpg in the palette editor only produced a single color. --Needed to package imageformats/qjpeg.dll with the Windows installer. -The basic memory benchmark test flame was not really testing memory. Make it more spread out. -Remove the temporal samples field from the flame tab, it was never used because it's only an animation parameter which is specified in the final render dialog or on the command line with EmberAnimate. --Code changes: -Add IsEmpty() to Palette to determine if a palette is all black. -Attempt to avoid selecting a blank palette in PaletteList::GetRandomPalette(). -Add function ScanForChaosNodes() and some associated helper functions in XmlToEmber. -Make variation param name correction be case insensitive in XmlToEmber. -Report error when assigning a variation param value in XmlToEmber. -Add SubBatchPercentPerThread() method to RendererCL. -Override enterEvent() and leaveEvent() in DoubleSpinBox and SpinBox to prevent the context menu from showing up on right mouse up after already leaving the spinner. -Filtering the mouse wheel event in TableWidget no longer appears to be needed. It was probably an old Qt bug that has been fixed. -Gui/ember syncing code in the final render dialog needed to be reworked to accommodate absolute sizes.
2019-04-13 22:00:46 -04:00
/// <summary>
/// truchet_knot.
/// </summary>
template <typename T>
class TruchetKnotVariation : public Variation<T>
{
public:
TruchetKnotVariation(T weight = 1.0) : Variation<T>("truchet_knot", eVariationId::VAR_TRUCHET_KNOT, weight) { }
VARCOPY(TruchetKnotVariation)
virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
{
T wd = T(0.5);
T space = T(0.1);
T cellx = T(Floor<T>(helper.In.x));
T celly = T(Floor<T>(helper.In.y));
T xy0x = (rand.Frand01<T>() - T(0.5)) * wd;
T xy0y = (rand.Frand01<T>() * 2 - 1) * (1 - space - wd * T(0.5));
T dir0 = std::abs(cellx + celly);
T dir1 = dir0 - 2 * Floor<T>(dir0 * T(0.5));
T xyx, xyy;
if (dir1 < 0.5)
{
xyx = xy0x;
xyy = xy0y;
}
else
{
xyx = -xy0y;//y and x intentionally flipped.
xyy = xy0x;
}
helper.Out.x = m_Weight * (cellx + xyx);
helper.Out.y = m_Weight * (celly + xyy);
helper.Out.z = DefaultZ(helper);
}
virtual string OpenCLString() const override
{
ostringstream ss;
intmax_t varIndex = IndexInXform();
string weight = WeightDefineString();
ss << "\t{\n"
<< "\t\treal_t wd = 0.5;\n"
<< "\t\treal_t space = 0.1;\n"
<< "\t\treal_t cellx = floor(vIn.x);\n"
<< "\t\treal_t celly = floor(vIn.y);\n"
<< "\t\treal_t xy0x = (MwcNext01(mwc) - 0.5) * wd;\n"
--Bug fixes -Ensure all OpenCL functions that take multiple arguments cast literal values to real_t.
2019-05-23 22:44:27 -04:00
<< "\t\treal_t xy0y = fma(MwcNext01(mwc), (real_t)(2.0), (real_t)(-1.0)) * (1.0 - space - wd * 0.5);\n"
--User changes: -Show common folder locations such as documents, downloads, pictures in the sidebar in all file dialogs. -Warning message about exceeding memory in final render dialog now suggests strips as the solution to the problem. -Strips now has a tooltip explaining what it does. -Allow more digits in the spinners on the color section the flame tab. -Add manually adjustable size spinners in the final render dialog. Percentage scale and absolute size are fully synced. -Default prefix in final render is now the filename when doing animations (coming from sequence section of the library tab). -Changed the elliptic variation back to using a less precise version for float, and a more precise version for double. The last release had it always using double. -New applied xaos table that shows a read-only view of actual weights by taking the base xform weights and multiplying them by the xaos values. -New table in the xaos tab that gives a graphical representation of the probability that each xform is chosen, with and without xaos. -Add button to transpose the xaos rows and columns. -Add support for importing .chaos files from Chaotica. --Pasting back to Chaotica will work for most, but not all, variations due to incompatible parameter names in some. -Curves are now splines instead of Bezier. This adds compatibility with Chaotica, but breaks it for Apophysis. Xmls are still pastable, but the color curves will look different. --The curve editor on the palette tab can now add points by clicking on the lines and remove points by clicking on the points themselves, just like Chaotica. --Splines are saved in four new xml fields: overall_curve, red_curve, green_curve and blue_curve. -Allow for specifying the percentage of a sub batch each thread should iterate through per kernel call when running with OpenCL. This gives a roughly 1% performance increase due to having to make less kernel calls while iterating. --This field is present for interactive editing (where it's not very useful) and in the final render dialog. --On the command line, this is specified as --sbpctth for EmberRender and EmberAnimate. -Allow double clicking to toggle the supersample field in the flame tab between 1 and 2 for easily checking the effect of the field. -When showing affine values as polar coordinates, show angles normalized to 360 to match Chaotica. -Fuse Count spinner now toggles between 15 and 100 when double clicking for easily checking the effect of the field. -Added field for limiting the range in the x and y direction that the initial points are chosen from. -Added a field called K2 which is an alternative way to set brightness, ignored when zero. --This has no effect for many variations, but hs a noticeable effect for some. -Added new variations: arcsech arcsech2 arcsinh arctanh asteria block bwraps_rand circlecrop2 coth_spiral crackle2 depth_blur depth_blur2 depth_gaussian depth_gaussian2 depth_ngon depth_ngon2 depth_sine depth_sine2 dragonfire dspherical dust excinis exp2 flipx flowerdb foci_p gaussian glynnia2 glynnsim4 glynnsim5 henon henon hex_rand hex_truchet hypershift lazyjess lens lozi lozi modulusx modulusy oscilloscope2 point_symmetry pointsymmetry projective pulse rotate scry2 shift smartshape spher squares starblur2 swirl3 swirl3r tanh_spiral target0 target2 tile_hlp truchet_glyph truchet_inv truchet_knot unicorngaloshen vibration vibration2 --hex_truchet, hex_rand should always use double. They are extremely sensitive. --Bug fixes: -Bounds sign was flipped for x coordinate of world space when center was not zero. -Right clicking and dragging spinner showed menu on mouse up, even if it was very far away. -Text boxes for size in final render dialog were hard to type in. Same bug as xform weight used to be so fix the same way. -Fix spelling to be plural in toggle color speed box. -Stop using the blank user palette to generate flames. Either put colored palettes in it, or exclude it from randoms. -Clicking the random palette button for a palette file with only one palette in it would freeze the program. -Clicking none scale in final render did not re-render the preview. -Use less precision on random xaos. No need for 12 decimal places. -The term sub batch is overloaded in the options dialog. Change the naming and tooltip of those settings for cpu and opencl. --Also made clear in the tooltip for the default opencl quality setting that the value is per device. -The arrows spinner in palette editor appears like a read-only label. Made it look like a spinner. -Fix border colors for various spin boxes and table headers in the style sheet. Requires reload. -Fix a bug in the bwraps variation which would produce different results than Chaotica and Apophysis. -Synth was allowed to be selected for random flame generation when using an Nvidia card but it shouldn't have been because Nvidia has a hard time compiling synth. -A casting bug in the OpenCL kernels for log scaling and density filtering was preventing successful compilations on Intel iGPUs. Fixed even though we don't support anything other than AMD and Nvidia. -Palette rotation (click and drag) position was not being reset when loading a new flame. -When the xform circles were hidden, opening and closing the options dialog would improperly reshow them. -Double click toggle was broken on integer spin boxes. -Fixed tab order of some controls. -Creating a palette from a jpg in the palette editor only produced a single color. --Needed to package imageformats/qjpeg.dll with the Windows installer. -The basic memory benchmark test flame was not really testing memory. Make it more spread out. -Remove the temporal samples field from the flame tab, it was never used because it's only an animation parameter which is specified in the final render dialog or on the command line with EmberAnimate. --Code changes: -Add IsEmpty() to Palette to determine if a palette is all black. -Attempt to avoid selecting a blank palette in PaletteList::GetRandomPalette(). -Add function ScanForChaosNodes() and some associated helper functions in XmlToEmber. -Make variation param name correction be case insensitive in XmlToEmber. -Report error when assigning a variation param value in XmlToEmber. -Add SubBatchPercentPerThread() method to RendererCL. -Override enterEvent() and leaveEvent() in DoubleSpinBox and SpinBox to prevent the context menu from showing up on right mouse up after already leaving the spinner. -Filtering the mouse wheel event in TableWidget no longer appears to be needed. It was probably an old Qt bug that has been fixed. -Gui/ember syncing code in the final render dialog needed to be reworked to accommodate absolute sizes.
2019-04-13 22:00:46 -04:00
<< "\t\treal_t dir0 = fabs(cellx + celly);\n"
<< "\t\treal_t dir1 = dir0 - 2.0 * floor(dir0 / 2.0);\n"
<< "\t\treal_t xyx, xyy;\n"
<< "\n"
<< "\t\tif (dir1 < 0.5)\n"
<< "\t\t{\n"
<< "\t\t xyx = xy0x;\n"
<< "\t\t xyy = xy0y;\n"
<< "\t\t}\n"
<< "\t\telse\n"
<< "\t\t{\n"
<< "\t\t xyx = -xy0y;//y and x intentionally flipped.\n"
<< "\t\t xyy = xy0x;\n"
<< "\t\t}\n"
<< "\n"
<< "\t\tvOut.x = " << weight << " * (cellx + xyx);\n"
<< "\t\tvOut.y = " << weight << " * (celly + xyy);\n"
<< "\t\tvOut.z = " << DefaultZCl()
<< "\t}\n";
return ss.str();
}
};
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
/// <summary>
/// gdoffs.
/// </summary>
template <typename T>
--Code changes -Split VariationList into .h/cpp files. --Make VariationList.cpp the only place where the variations files are included from. This speeds up the build, but variations can no longer be directly instantiated. --Variations are no longer exported.
2016-02-18 21:58:24 -05:00
class GdoffsVariation : public ParametricVariation<T>
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
public:
--Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums.
2016-01-04 19:50:15 -05:00
GdoffsVariation(T weight = 1.0) : ParametricVariation<T>("gdoffs", eVariationId::VAR_GDOFFS, weight)
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
Init();
}
PARVARCOPY(GdoffsVariation)
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
T oscX = GdoffsFosc(m_Dx, 1);
T oscY = GdoffsFosc(m_Dy, 1);
T inX = helper.In.x + m_Cx;
T inY = helper.In.y + m_Cy;
T outX;
T outY;
if (m_Square != 0)
{
outX = GdoffsFlip(GdoffsFlip(inX, GdoffsFosc(inX, 4), oscX), GdoffsFosc(GdoffsFclp(m_B * inX), 4), oscX);
outY = GdoffsFlip(GdoffsFlip(inY, GdoffsFosc(inY, 4), oscX), GdoffsFosc(GdoffsFclp(m_B * inY), 4), oscX);
}
else
{
outX = GdoffsFlip(GdoffsFlip(inX, GdoffsFosc(inX, 4), oscX), GdoffsFosc(GdoffsFclp(m_B * inX), 4), oscX);
outY = GdoffsFlip(GdoffsFlip(inY, GdoffsFosc(inY, 4), oscY), GdoffsFosc(GdoffsFclp(m_B * inY), 4), oscY);
}
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
helper.Out.x = m_Weight * outX;
helper.Out.y = m_Weight * outY;
helper.Out.z = m_Weight * helper.In.z;
}
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
--Code changes -Better use of const functions. -Ember.pro should not define BUILDING_EMBERCL, instead define BUILDING_EMBER.
2015-06-29 23:13:53 -04:00
virtual string OpenCLString() const override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
ostringstream ss, ss2;
0.4.1.3 Beta 10/14/2014 --User Changes Size is no longer fixed to the window size. Size scaling is done differently in the final render dialog. This fixes several bugs. Remove Xml saving size from settings and options dialog, it no longer applies. Final render can be broken into strips. Set default save path to the desktop if none is found in the settings file. Set default output size to 1920x1080 if none is found in the settings file. --Bug Fixes Better memory size reporting in final render dialog. --Code Changes Migrate to C++11, Qt 5.3.1, and Visual Studio 2013. Change most instances of unsigned int to size_t, and int to intmax_t. Add m_OrigPixPerUnit and m_ScaleType to Ember for scaling purposes. Replace some sprintf_s() calls in XmlToEmber with ostringstream. Move more non-templated members into RendererBase. Add CopyVec() overload that takes a per element function pointer. Add vector Memset(). Replace '&' with '+' instead of "&amp;" in XmlToEmber for much faster parsing. Break strips rendering out into EmberCommon and call from EmberRender and Fractorium. Make AddAndWriteBuffer() just call WriteBuffer(). Make AddAndWriteImage() delete the existing image first before replacing it. Add SetOutputTexture() to RendererCL to support making new textures in response to resize events. Remove multiple return statements in RendererCL, and replace with a bool that tracks results. Add ToDouble(), MakeEnd(), ToString() and Exists() wrappers in Fractorium. Add Size() wrapper in EmberFile. Make QString function arguments const QString&, and string with const string&. Make ShowCritical() wrapper for invoking a message box from another thread. Add combo box to TwoButtonWidget and rename.
2014-10-14 11:53:15 -04:00
intmax_t i = 0, varIndex = IndexInXform();
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
ss2 << "_" << XformIndexInEmber() << "]";
string index = ss2.str();
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
string weight = WeightDefineString();
string deltaX = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string deltaY = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string areaX = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string areaY = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
string centerX = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string centerY = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
string gamma = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string square = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string dx = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string ax = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string cx = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string dy = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string ay = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string cy = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string b = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
ss << "\t{\n"
--User changes -Further work on the About box. --Change qss files to make the text box in the About box be the same color as the one in the style dialog, so the link text is easier to see. --This requires reloading the qss file. --Bug fixes -Pre/post assign method was wrong for Sphereblur, Concentric, RandCubes and PixelFlow. --Code changes -Make Hypershift2 use MwcNextRange() instead of MwcNext() %. -Add m_HasPre to Xform similar to m_HasPost, to skip application of the affine if it's the ID matrix (in this case just assign the input x and y points). This gives a 5% speedup.
2018-09-18 23:49:38 -04:00
<< "\t\treal_t oscX = GdoffsFosc_(" << dx << ", 1);\n"
<< "\t\treal_t oscY = GdoffsFosc_(" << dy << ", 1);\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t\treal_t inX = vIn.x + " << cx << ";\n"
<< "\t\treal_t inY = vIn.y + " << cy << ";\n"
<< "\t\treal_t outX;\n"
<< "\t\treal_t outY;\n"
<< "\n"
<< "\t\tif (" << square << " != 0)\n"
<< "\t\t{\n"
--User changes -Further work on the About box. --Change qss files to make the text box in the About box be the same color as the one in the style dialog, so the link text is easier to see. --This requires reloading the qss file. --Bug fixes -Pre/post assign method was wrong for Sphereblur, Concentric, RandCubes and PixelFlow. --Code changes -Make Hypershift2 use MwcNextRange() instead of MwcNext() %. -Add m_HasPre to Xform similar to m_HasPost, to skip application of the affine if it's the ID matrix (in this case just assign the input x and y points). This gives a 5% speedup.
2018-09-18 23:49:38 -04:00
<< "\t\t outX = GdoffsFlip(GdoffsFlip(inX, GdoffsFosc_(inX, 4), oscX), GdoffsFosc_(GdoffsFclp(" << b << " * inX), 4), oscX);\n"
<< "\t\t outY = GdoffsFlip(GdoffsFlip(inY, GdoffsFosc_(inY, 4), oscX), GdoffsFosc_(GdoffsFclp(" << b << " * inY), 4), oscX);\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t\t}\n"
<< "\t\telse\n"
<< "\t\t{\n"
--User changes -Further work on the About box. --Change qss files to make the text box in the About box be the same color as the one in the style dialog, so the link text is easier to see. --This requires reloading the qss file. --Bug fixes -Pre/post assign method was wrong for Sphereblur, Concentric, RandCubes and PixelFlow. --Code changes -Make Hypershift2 use MwcNextRange() instead of MwcNext() %. -Add m_HasPre to Xform similar to m_HasPost, to skip application of the affine if it's the ID matrix (in this case just assign the input x and y points). This gives a 5% speedup.
2018-09-18 23:49:38 -04:00
<< "\t\t outX = GdoffsFlip(GdoffsFlip(inX, GdoffsFosc_(inX, 4), oscX), GdoffsFosc_(GdoffsFclp(" << b << " * inX), 4), oscX);\n"
<< "\t\t outY = GdoffsFlip(GdoffsFlip(inY, GdoffsFosc_(inY, 4), oscY), GdoffsFosc_(GdoffsFclp(" << b << " * inY), 4), oscY);\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t\t}\n"
<< "\n"
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
<< "\t\tvOut.x = " << weight << " * outX;\n"
<< "\t\tvOut.y = " << weight << " * outY;\n"
<< "\t\tvOut.z = " << weight << " * vIn.z;\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t}\n";
return ss.str();
}
--User changes -Add new variations: bubbleT3D, crob, hexaplay3D, hexcrop, hexes, hexnix3D, loonie2, loonie3, nBlur, octapol and synth. -Allow for pre/post versions of dc_bubble, dc_cylinder and dc_linear whereas before they were omitted. -When saving a file with multiple embers in it, detect if time values are all the same and if so, start them at zero and increment by 1 for each ember. -Allow for numerous quality increases to be coalesced into one. It will pick up at the end of the current render. -Show selection highlight on variations tree in response to mouse hover. This makes it easier to see for which variation or param the current mouse wheel action will apply. -Make default temporal samples be 100, whereas before it was 1000 which was overkill. -Require the shift key to be held with delete for deleting an ember to prevent it from triggering when the user enters delete in the edit box. -This wasn't otherwise fixable without writing a lot more code. --Bug fixes -EmberGenome was crashing when generating a sequence from a source file with more than 2 embers in it. -EmberGenome was improperly handling the first frame of a merge after the last frame of the loop. -These bugs were due to a previous commit. Revert parts of that commit. -Prevent a zoom value of less than 0 when reading from xml. -Slight optimization of the crescents, and mask variations, if the compiler wasn't doing it already. -Unique file naming was broken because it was looking for _# and the default names ended with -#. -Disallow renaming of an ember in the library tree to an empty string. -Severe bug that prevented some variations from being read correctly from params generated outside this program. -Severe OpenCL randomization bug. The first x coordinates of the first points in the first kernel call of the first ember of a render since the OpenCL renderer object was created were not random and were mostly -1. -Severe bug when populating xform selection distributions that could sometimes cause a crash due to roundoff error. Fix by using double. -Limit the max number of variations in a random ember to MAX_CL_VARS, which is 8. This ensures they'll look the same on CPU and GPU. -Prevent user from saving stylesheet to default.qss, it's a special reserved filename. --Code changes -Generalize using the running sum output point inside of a variation for all cases: pre, reg and post. -Allow for array variables in variations where the address of each element is stored in m_Params. -Qualify all math functions with std:: -No longer use our own Clamp() in OpenCL, instead use the standard clamp(). -Redesign how functions are used in the variations OpenCL code. -Add tests to EmberTester to verify some of the new functionality. -Place more const and override qualifiers on functions where appropriate. -Add a global rand with a lock to be used very sparingly. -Use a map instead of a vector for bad param names in Xml parsing. -Prefix affine interpolation mode defines with "AFFINE_" to make their purpose more clear. -Allow for variations that change state during iteration by sending a separate copy of the ember to each rendering thread. -Implement this same functionality with a local struct in OpenCL. It's members are the total of all variables that need to change state within an ember. -Add Contains() function to Utils.h. -EmberRender: print names of kernels being printed with --dump_kernel option. -Clean up EmberTester to handle some of the recent changes. -Fix various casts. -Replace % 2 with & 1, even though the compiler was likely doing this already. -Add new file Variations06.h to accommodate new variations. -General cleanup.
2015-11-22 17:15:07 -05:00
virtual string OpenCLFuncsString() const override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
return
"inline real_t GdoffsFcip(real_t a) { return (real_t)((a < 0) ? -((int)(fabs(a)) + 1) : 0) + ((a > 1) ? ((int)(a)) : 0); }\n"
"inline real_t GdoffsFclp(real_t a) { return ((a < 0) ? -(fmod(fabs(a), 1)) : fmod(fabs(a), 1)); }\n"
"inline real_t GdoffsFscl(real_t a) { return GdoffsFclp((a + 1) / 2); }\n"
--User changes -Further work on the About box. --Change qss files to make the text box in the About box be the same color as the one in the style dialog, so the link text is easier to see. --This requires reloading the qss file. --Bug fixes -Pre/post assign method was wrong for Sphereblur, Concentric, RandCubes and PixelFlow. --Code changes -Make Hypershift2 use MwcNextRange() instead of MwcNext() %. -Add m_HasPre to Xform similar to m_HasPost, to skip application of the affine if it's the ID matrix (in this case just assign the input x and y points). This gives a 5% speedup.
2018-09-18 23:49:38 -04:00
"inline real_t GdoffsFosc_(real_t p, real_t a) { return GdoffsFscl(-1 * cos(p * a * M_2PI)); }\n"//Underscore added to name to prevent false detection with the global function Fosc() in TestGlobalFuncs() in EmberTester.
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
"inline real_t GdoffsFlip(real_t a, real_t b, real_t c) { return fma(c, (b - a), a); }\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
"\n";
}
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
virtual void Precalc() override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
const T agdod = T(0.1);
const T agdoa = 2;
const T agdoc = 1;
m_Dx = m_DeltaX * agdod;
m_Dy = m_DeltaY * agdod;
--User changes -Add post_smartcrop. --Bug fixes -Fix bug in crackle. -Wrong point assignment in hexaplay3D, hexnix3D. -Improper Z assignment in rblur. -Fix inconsistency with original in circlecrop. -Put EMBER_ROOT bakc to ./../../../ in default.pri. This is TBD. --Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums. -Optimize crob, nBlur. -Fix weird assignment statement in falloff3. -Cleanup in VarFuncs::SimplexNoise3D(). -Replace fabs() with std::abs(). -General cleanup.
2016-01-12 23:42:12 -05:00
m_Ax = ((std::abs(m_AreaX) < 0.1) ? T(0.1) : std::abs(m_AreaX)) * agdoa;
m_Ay = ((std::abs(m_AreaY) < 0.1) ? T(0.1) : std::abs(m_AreaY)) * agdoa;
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
m_Cx = m_CenterX * agdoc;
m_Cy = m_CenterY * agdoc;
Remove ReadMe.txt from all project files. Add Curves.h, and CurvesGraphicsView.h/cpp to support bezier color curves. Add Curves member to Ember. Add curves capability to EmberCL. Remove some unused variables in the kernel created in RendererCL::CreateFinalAccumKernelString(). Use glm namespace for vec classes if GLM_VERSION >= 96, else use glm::detail. As a result of using glm namespace, all instances of min and max had to be qualified with std:: Split ComputeCamera into that and ComputeQuality(). Reduce the amount of ComputeCamera() and MakeDmap() calls on each incremental iter that doesn't use temporal samples. Fix clamping bug with DE filter widths. Provide functions to return the kernels from RendererCL to assist with diagnostics and debugging. Prevent extra newline in EmberRender when only rendering a single image. Add the ability to delete an ember at a given index in EmberFile. Allow deleting/focusing ember in library tab with delete and enter keys. Reorder some code in Fractorium.h to match the tabs order. Add and call ClearFinalImages() to clear buffers in controller to fix bug where previous CPU render would be shown for a split second when switching from OpenCL back to CPU. Refactor ember library pointer syncing to a function SyncPointers(). Add the ability to save ember Xmls to an unique automatically generated name after the first time the user has specified a name.
2015-03-21 18:27:37 -04:00
m_B = m_Gamma * agdoa / (std::max(m_Ax, m_Ay));
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
}
protected:
void Init()
{
string prefix = Prefix();
m_Params.clear();
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
m_Params.push_back(ParamWithName<T>(&m_DeltaX, prefix + "gdoffs_delta_x", 0, eParamType::REAL, 0, 16));
m_Params.push_back(ParamWithName<T>(&m_DeltaY, prefix + "gdoffs_delta_y", 0, eParamType::REAL, 0, 16));
m_Params.push_back(ParamWithName<T>(&m_AreaX, prefix + "gdoffs_area_x", 2));
m_Params.push_back(ParamWithName<T>(&m_AreaY, prefix + "gdoffs_area_y", 2));
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
m_Params.push_back(ParamWithName<T>(&m_CenterX, prefix + "gdoffs_center_x"));
m_Params.push_back(ParamWithName<T>(&m_CenterY, prefix + "gdoffs_center_y"));
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
m_Params.push_back(ParamWithName<T>(&m_Gamma, prefix + "gdoffs_gamma", 1, eParamType::INTEGER, 1, 6));
m_Params.push_back(ParamWithName<T>(&m_Square, prefix + "gdoffs_square", 0, eParamType::INTEGER, 0, 1));
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
m_Params.push_back(ParamWithName<T>(true, &m_Dx, prefix + "gdoffs_dx"));
m_Params.push_back(ParamWithName<T>(true, &m_Ax, prefix + "gdoffs_ax"));
m_Params.push_back(ParamWithName<T>(true, &m_Cx, prefix + "gdoffs_cx"));
m_Params.push_back(ParamWithName<T>(true, &m_Dy, prefix + "gdoffs_dyd"));
m_Params.push_back(ParamWithName<T>(true, &m_Ay, prefix + "gdoffs_ay"));
m_Params.push_back(ParamWithName<T>(true, &m_Cy, prefix + "gdoffs_cy"));
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
m_Params.push_back(ParamWithName<T>(true, &m_B, prefix + "gdoffs_b"));
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
}
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
private:
--User changes -Add post_smartcrop. --Bug fixes -Fix bug in crackle. -Wrong point assignment in hexaplay3D, hexnix3D. -Improper Z assignment in rblur. -Fix inconsistency with original in circlecrop. -Put EMBER_ROOT bakc to ./../../../ in default.pri. This is TBD. --Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums. -Optimize crob, nBlur. -Fix weird assignment statement in falloff3. -Cleanup in VarFuncs::SimplexNoise3D(). -Replace fabs() with std::abs(). -General cleanup.
2016-01-12 23:42:12 -05:00
static inline T GdoffsFcip(T a) { return T((a < 0) ? -(int(std::abs(a)) + 1) : 0) + ((a > 1) ? (int(a)) : 0); }
static inline T GdoffsFclp(T a) { return ((a < 0) ? -(fmod(std::abs(a), T(1))) : fmod(std::abs(a), T(1))); }
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
static inline T GdoffsFscl(T a) { return GdoffsFclp((a + 1) / 2); }
--User changes -Add new variations: bubbleT3D, crob, hexaplay3D, hexcrop, hexes, hexnix3D, loonie2, loonie3, nBlur, octapol and synth. -Allow for pre/post versions of dc_bubble, dc_cylinder and dc_linear whereas before they were omitted. -When saving a file with multiple embers in it, detect if time values are all the same and if so, start them at zero and increment by 1 for each ember. -Allow for numerous quality increases to be coalesced into one. It will pick up at the end of the current render. -Show selection highlight on variations tree in response to mouse hover. This makes it easier to see for which variation or param the current mouse wheel action will apply. -Make default temporal samples be 100, whereas before it was 1000 which was overkill. -Require the shift key to be held with delete for deleting an ember to prevent it from triggering when the user enters delete in the edit box. -This wasn't otherwise fixable without writing a lot more code. --Bug fixes -EmberGenome was crashing when generating a sequence from a source file with more than 2 embers in it. -EmberGenome was improperly handling the first frame of a merge after the last frame of the loop. -These bugs were due to a previous commit. Revert parts of that commit. -Prevent a zoom value of less than 0 when reading from xml. -Slight optimization of the crescents, and mask variations, if the compiler wasn't doing it already. -Unique file naming was broken because it was looking for _# and the default names ended with -#. -Disallow renaming of an ember in the library tree to an empty string. -Severe bug that prevented some variations from being read correctly from params generated outside this program. -Severe OpenCL randomization bug. The first x coordinates of the first points in the first kernel call of the first ember of a render since the OpenCL renderer object was created were not random and were mostly -1. -Severe bug when populating xform selection distributions that could sometimes cause a crash due to roundoff error. Fix by using double. -Limit the max number of variations in a random ember to MAX_CL_VARS, which is 8. This ensures they'll look the same on CPU and GPU. -Prevent user from saving stylesheet to default.qss, it's a special reserved filename. --Code changes -Generalize using the running sum output point inside of a variation for all cases: pre, reg and post. -Allow for array variables in variations where the address of each element is stored in m_Params. -Qualify all math functions with std:: -No longer use our own Clamp() in OpenCL, instead use the standard clamp(). -Redesign how functions are used in the variations OpenCL code. -Add tests to EmberTester to verify some of the new functionality. -Place more const and override qualifiers on functions where appropriate. -Add a global rand with a lock to be used very sparingly. -Use a map instead of a vector for bad param names in Xml parsing. -Prefix affine interpolation mode defines with "AFFINE_" to make their purpose more clear. -Allow for variations that change state during iteration by sending a separate copy of the ember to each rendering thread. -Implement this same functionality with a local struct in OpenCL. It's members are the total of all variables that need to change state within an ember. -Add Contains() function to Utils.h. -EmberRender: print names of kernels being printed with --dump_kernel option. -Clean up EmberTester to handle some of the recent changes. -Fix various casts. -Replace % 2 with & 1, even though the compiler was likely doing this already. -Add new file Variations06.h to accommodate new variations. -General cleanup.
2015-11-22 17:15:07 -05:00
static inline T GdoffsFosc(T p, T a) { return GdoffsFscl(-1 * std::cos(p * a * M_2PI)); }
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
static inline T GdoffsFlip(T a, T b, T c) { return (c * (b - a) + a); }
T m_DeltaX;//Params.
T m_DeltaY;
T m_AreaX;
T m_AreaY;
T m_CenterX;
T m_CenterY;
T m_Gamma;
T m_Square;
T m_Dx;//Precalc.
T m_Ax;
T m_Cx;
T m_Dy;
T m_Ay;
T m_Cy;
T m_B;
};
/// <summary>
/// octagon.
/// </summary>
template <typename T>
--Code changes -Split VariationList into .h/cpp files. --Make VariationList.cpp the only place where the variations files are included from. This speeds up the build, but variations can no longer be directly instantiated. --Variations are no longer exported.
2016-02-18 21:58:24 -05:00
class OctagonVariation : public ParametricVariation<T>
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
public:
--Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums.
2016-01-04 19:50:15 -05:00
OctagonVariation(T weight = 1.0) : ParametricVariation<T>("octagon", eVariationId::VAR_OCTAGON, weight)
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
Init();
}
PARVARCOPY(OctagonVariation)
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
T x2 = SQR(helper.In.x);
T y2 = SQR(helper.In.y);
T z2 = SQR(helper.In.z);
T r = m_Weight / Zeps(SQR(x2) + z2 + SQR(y2) + z2);
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
if (r < 2)
{
helper.Out.x = r * helper.In.x;
helper.Out.y = r * helper.In.y;
helper.Out.z = r * helper.In.z;
}
else
{
helper.Out.x = m_Weight * helper.In.x;
helper.Out.y = m_Weight * helper.In.y;
helper.Out.z = m_Weight * helper.In.z;
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
T t = m_Weight / Zeps(std::sqrt(SQR(helper.In.x)) + std::sqrt(helper.In.z) + std::sqrt(SQR(helper.In.y)) + std::sqrt(helper.In.z));
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
if (r >= 0)
{
helper.Out.x = t * helper.In.x;
helper.Out.y = t * helper.In.y;
helper.Out.z = t * helper.In.z;
}
else
{
helper.Out.x = m_Weight * helper.In.x;
helper.Out.y = m_Weight * helper.In.y;
helper.Out.z = m_Weight * helper.In.z;
}
if (helper.In.x >= 0)
helper.Out.x = m_Weight * (helper.In.x + m_X);
else
helper.Out.x = m_Weight * (helper.In.x - m_X);
if (helper.In.y >= 0)
helper.Out.y = m_Weight * (helper.In.y + m_Y);
else
helper.Out.y = m_Weight * (helper.In.y - m_Y);
if (helper.In.z >= 0)
helper.Out.z = m_Weight * (helper.In.z + m_Z);
else
helper.Out.z = m_Weight * (helper.In.z - m_Z);
}
}
--Code changes -Better use of const functions. -Ember.pro should not define BUILDING_EMBERCL, instead define BUILDING_EMBER.
2015-06-29 23:13:53 -04:00
virtual string OpenCLString() const override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
ostringstream ss, ss2;
0.4.1.3 Beta 10/14/2014 --User Changes Size is no longer fixed to the window size. Size scaling is done differently in the final render dialog. This fixes several bugs. Remove Xml saving size from settings and options dialog, it no longer applies. Final render can be broken into strips. Set default save path to the desktop if none is found in the settings file. Set default output size to 1920x1080 if none is found in the settings file. --Bug Fixes Better memory size reporting in final render dialog. --Code Changes Migrate to C++11, Qt 5.3.1, and Visual Studio 2013. Change most instances of unsigned int to size_t, and int to intmax_t. Add m_OrigPixPerUnit and m_ScaleType to Ember for scaling purposes. Replace some sprintf_s() calls in XmlToEmber with ostringstream. Move more non-templated members into RendererBase. Add CopyVec() overload that takes a per element function pointer. Add vector Memset(). Replace '&' with '+' instead of "&amp;" in XmlToEmber for much faster parsing. Break strips rendering out into EmberCommon and call from EmberRender and Fractorium. Make AddAndWriteBuffer() just call WriteBuffer(). Make AddAndWriteImage() delete the existing image first before replacing it. Add SetOutputTexture() to RendererCL to support making new textures in response to resize events. Remove multiple return statements in RendererCL, and replace with a bool that tracks results. Add ToDouble(), MakeEnd(), ToString() and Exists() wrappers in Fractorium. Add Size() wrapper in EmberFile. Make QString function arguments const QString&, and string with const string&. Make ShowCritical() wrapper for invoking a message box from another thread. Add combo box to TwoButtonWidget and rename.
2014-10-14 11:53:15 -04:00
intmax_t i = 0, varIndex = IndexInXform();
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
ss2 << "_" << XformIndexInEmber() << "]";
string index = ss2.str();
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
string weight = WeightDefineString();
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
string x = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string y = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string z = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
ss << "\t{\n"
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
<< "\t\treal_t x2 = SQR(vIn.x);\n"
<< "\t\treal_t y2 = SQR(vIn.y);\n"
<< "\t\treal_t z2 = SQR(vIn.z);\n"
<< "\t\treal_t r = " << weight << " / Zeps(fma(x2, x2, z2) + fma(y2, y2, z2));\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\n"
<< "\t\tif (r < 2)\n"
<< "\t\t{\n"
<< "\t\t vOut.x = r * vIn.x;\n"
<< "\t\t vOut.y = r * vIn.y;\n"
<< "\t\t vOut.z = r * vIn.z;\n"
<< "\t\t}\n"
<< "\t\telse\n"
<< "\t\t{\n"
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
<< "\t\t vOut.x = " << weight << " * vIn.x;\n"
<< "\t\t vOut.y = " << weight << " * vIn.y;\n"
<< "\t\t vOut.z = " << weight << " * vIn.z;\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\n"
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
<< "\t\t real_t t = " << weight << " / Zeps(sqrt(SQR(vIn.x)) + sqrt(vIn.z) + sqrt(SQR(vIn.y)) + sqrt(vIn.z));\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\n"
<< "\t\t if (r >= 0)\n"
<< "\t\t {\n"
<< "\t\t vOut.x = t * vIn.x;\n"
<< "\t\t vOut.y = t * vIn.y;\n"
<< "\t\t vOut.z = t * vIn.z;\n"
<< "\t\t }\n"
<< "\t\t else\n"
<< "\t\t {\n"
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
<< "\t\t vOut.x = " << weight << " * vIn.x;\n"
<< "\t\t vOut.y = " << weight << " * vIn.y;\n"
<< "\t\t vOut.z = " << weight << " * vIn.z;\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t\t }\n"
<< "\n"
<< "\t\t if (vIn.x >= 0)\n"
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
<< "\t\t vOut.x = " << weight << " * (vIn.x + " << x << ");\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t\t else\n"
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
<< "\t\t vOut.x = " << weight << " * (vIn.x - " << x << ");\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\n"
<< "\t\t if (vIn.y >= 0)\n"
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
<< "\t\t vOut.y = " << weight << " * (vIn.y + " << y << ");\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t\t else\n"
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
<< "\t\t vOut.y = " << weight << " * (vIn.y - " << y << ");\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\n"
<< "\t\t if (vIn.z >= 0)\n"
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
<< "\t\t vOut.z = " << weight << " * (vIn.z + " << z << ");\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t\t else\n"
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
<< "\t\t vOut.z = " << weight << " * (vIn.z - " << z << ");\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t\t}\n"
<< "\t}\n";
return ss.str();
}
--User changes -Add new variations: bubbleT3D, crob, hexaplay3D, hexcrop, hexes, hexnix3D, loonie2, loonie3, nBlur, octapol and synth. -Allow for pre/post versions of dc_bubble, dc_cylinder and dc_linear whereas before they were omitted. -When saving a file with multiple embers in it, detect if time values are all the same and if so, start them at zero and increment by 1 for each ember. -Allow for numerous quality increases to be coalesced into one. It will pick up at the end of the current render. -Show selection highlight on variations tree in response to mouse hover. This makes it easier to see for which variation or param the current mouse wheel action will apply. -Make default temporal samples be 100, whereas before it was 1000 which was overkill. -Require the shift key to be held with delete for deleting an ember to prevent it from triggering when the user enters delete in the edit box. -This wasn't otherwise fixable without writing a lot more code. --Bug fixes -EmberGenome was crashing when generating a sequence from a source file with more than 2 embers in it. -EmberGenome was improperly handling the first frame of a merge after the last frame of the loop. -These bugs were due to a previous commit. Revert parts of that commit. -Prevent a zoom value of less than 0 when reading from xml. -Slight optimization of the crescents, and mask variations, if the compiler wasn't doing it already. -Unique file naming was broken because it was looking for _# and the default names ended with -#. -Disallow renaming of an ember in the library tree to an empty string. -Severe bug that prevented some variations from being read correctly from params generated outside this program. -Severe OpenCL randomization bug. The first x coordinates of the first points in the first kernel call of the first ember of a render since the OpenCL renderer object was created were not random and were mostly -1. -Severe bug when populating xform selection distributions that could sometimes cause a crash due to roundoff error. Fix by using double. -Limit the max number of variations in a random ember to MAX_CL_VARS, which is 8. This ensures they'll look the same on CPU and GPU. -Prevent user from saving stylesheet to default.qss, it's a special reserved filename. --Code changes -Generalize using the running sum output point inside of a variation for all cases: pre, reg and post. -Allow for array variables in variations where the address of each element is stored in m_Params. -Qualify all math functions with std:: -No longer use our own Clamp() in OpenCL, instead use the standard clamp(). -Redesign how functions are used in the variations OpenCL code. -Add tests to EmberTester to verify some of the new functionality. -Place more const and override qualifiers on functions where appropriate. -Add a global rand with a lock to be used very sparingly. -Use a map instead of a vector for bad param names in Xml parsing. -Prefix affine interpolation mode defines with "AFFINE_" to make their purpose more clear. -Allow for variations that change state during iteration by sending a separate copy of the ember to each rendering thread. -Implement this same functionality with a local struct in OpenCL. It's members are the total of all variables that need to change state within an ember. -Add Contains() function to Utils.h. -EmberRender: print names of kernels being printed with --dump_kernel option. -Clean up EmberTester to handle some of the recent changes. -Fix various casts. -Replace % 2 with & 1, even though the compiler was likely doing this already. -Add new file Variations06.h to accommodate new variations. -General cleanup.
2015-11-22 17:15:07 -05:00
virtual vector<string> OpenCLGlobalFuncNames() const override
{
return vector<string> { "Zeps" };
}
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
protected:
void Init()
{
string prefix = Prefix();
m_Params.clear();
m_Params.push_back(ParamWithName<T>(&m_X, prefix + "octagon_x"));//Original used a prefix of octa_, which is incompatible with Ember's design.
m_Params.push_back(ParamWithName<T>(&m_Y, prefix + "octagon_y"));
m_Params.push_back(ParamWithName<T>(&m_Z, prefix + "octagon_z"));
}
private:
T m_X;
T m_Y;
T m_Z;
};
/// <summary>
/// trade.
/// </summary>
template <typename T>
--Code changes -Split VariationList into .h/cpp files. --Make VariationList.cpp the only place where the variations files are included from. This speeds up the build, but variations can no longer be directly instantiated. --Variations are no longer exported.
2016-02-18 21:58:24 -05:00
class TradeVariation : public ParametricVariation<T>
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
public:
--Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums.
2016-01-04 19:50:15 -05:00
TradeVariation(T weight = 1.0) : ParametricVariation<T>("trade", eVariationId::VAR_TRADE, weight)
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
Init();
}
PARVARCOPY(TradeVariation)
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
T r, temp, c1mx;
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
if (helper.In.x > 0)
{
c1mx = m_C1 - helper.In.x;
Explicitly namespace calls to sqrt and isnan osx/linux have alternative top-level implementations of these methods, so the calls are ambiguous given the - so these should be used explicitly
2015-04-13 07:32:58 -04:00
r = std::sqrt(SQR(c1mx) + SQR(helper.In.y));
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
if (r <= m_R1)
{
r *= m_R2 / m_R1;
06/09/2017 --User changes -dark.qss is now per-OS. -Properly set/reload palette when coming from the palette editor. The latter must be done if they've modified the current palette even if they've clicked cancel. --Bug fixes -Make the following variations safer by using Zeps(): conic, bipolar, edisc, whorl, tan, csc, cot, tanh, sech, csch, coth, auger, bwraps, hypertile3d, hypertile3d1, ortho, poincare, rational3, barycentroid, sschecks, cscq, cschq, scry_3D, splitbrdr, hexcrop, nblur, crob. -Fix bug enabling/disabling overwrite button in palette editor. -Small optimization for gdoffs, use precalcAtanYX. -Properly propagate z through circlesplit, cylinder2 and tile_log variations. -Some values in truchet_fill could've been NaN. --Code changes -Make most installation files read only. -Qualify many calls with std:: to ensure they're not colliding with glm:: -Use auto in more places.
2017-06-09 22:38:06 -04:00
temp = std::atan2(helper.In.y, c1mx);
--User changes -Add new variations: bubbleT3D, crob, hexaplay3D, hexcrop, hexes, hexnix3D, loonie2, loonie3, nBlur, octapol and synth. -Allow for pre/post versions of dc_bubble, dc_cylinder and dc_linear whereas before they were omitted. -When saving a file with multiple embers in it, detect if time values are all the same and if so, start them at zero and increment by 1 for each ember. -Allow for numerous quality increases to be coalesced into one. It will pick up at the end of the current render. -Show selection highlight on variations tree in response to mouse hover. This makes it easier to see for which variation or param the current mouse wheel action will apply. -Make default temporal samples be 100, whereas before it was 1000 which was overkill. -Require the shift key to be held with delete for deleting an ember to prevent it from triggering when the user enters delete in the edit box. -This wasn't otherwise fixable without writing a lot more code. --Bug fixes -EmberGenome was crashing when generating a sequence from a source file with more than 2 embers in it. -EmberGenome was improperly handling the first frame of a merge after the last frame of the loop. -These bugs were due to a previous commit. Revert parts of that commit. -Prevent a zoom value of less than 0 when reading from xml. -Slight optimization of the crescents, and mask variations, if the compiler wasn't doing it already. -Unique file naming was broken because it was looking for _# and the default names ended with -#. -Disallow renaming of an ember in the library tree to an empty string. -Severe bug that prevented some variations from being read correctly from params generated outside this program. -Severe OpenCL randomization bug. The first x coordinates of the first points in the first kernel call of the first ember of a render since the OpenCL renderer object was created were not random and were mostly -1. -Severe bug when populating xform selection distributions that could sometimes cause a crash due to roundoff error. Fix by using double. -Limit the max number of variations in a random ember to MAX_CL_VARS, which is 8. This ensures they'll look the same on CPU and GPU. -Prevent user from saving stylesheet to default.qss, it's a special reserved filename. --Code changes -Generalize using the running sum output point inside of a variation for all cases: pre, reg and post. -Allow for array variables in variations where the address of each element is stored in m_Params. -Qualify all math functions with std:: -No longer use our own Clamp() in OpenCL, instead use the standard clamp(). -Redesign how functions are used in the variations OpenCL code. -Add tests to EmberTester to verify some of the new functionality. -Place more const and override qualifiers on functions where appropriate. -Add a global rand with a lock to be used very sparingly. -Use a map instead of a vector for bad param names in Xml parsing. -Prefix affine interpolation mode defines with "AFFINE_" to make their purpose more clear. -Allow for variations that change state during iteration by sending a separate copy of the ember to each rendering thread. -Implement this same functionality with a local struct in OpenCL. It's members are the total of all variables that need to change state within an ember. -Add Contains() function to Utils.h. -EmberRender: print names of kernels being printed with --dump_kernel option. -Clean up EmberTester to handle some of the recent changes. -Fix various casts. -Replace % 2 with & 1, even though the compiler was likely doing this already. -Add new file Variations06.h to accommodate new variations. -General cleanup.
2015-11-22 17:15:07 -05:00
helper.Out.x = m_Weight * (r * std::cos(temp) - m_C2);
helper.Out.y = m_Weight * r * std::sin(temp);
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
}
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
else
{
helper.Out.x = m_Weight * helper.In.x;
helper.Out.y = m_Weight * helper.In.y;
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
}
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
}
else
{
c1mx = -m_C2 - helper.In.x;
Explicitly namespace calls to sqrt and isnan osx/linux have alternative top-level implementations of these methods, so the calls are ambiguous given the - so these should be used explicitly
2015-04-13 07:32:58 -04:00
r = std::sqrt(SQR(c1mx) + SQR(helper.In.y));
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
if (r <= m_R2)
{
r *= m_R1 / m_R2;
06/09/2017 --User changes -dark.qss is now per-OS. -Properly set/reload palette when coming from the palette editor. The latter must be done if they've modified the current palette even if they've clicked cancel. --Bug fixes -Make the following variations safer by using Zeps(): conic, bipolar, edisc, whorl, tan, csc, cot, tanh, sech, csch, coth, auger, bwraps, hypertile3d, hypertile3d1, ortho, poincare, rational3, barycentroid, sschecks, cscq, cschq, scry_3D, splitbrdr, hexcrop, nblur, crob. -Fix bug enabling/disabling overwrite button in palette editor. -Small optimization for gdoffs, use precalcAtanYX. -Properly propagate z through circlesplit, cylinder2 and tile_log variations. -Some values in truchet_fill could've been NaN. --Code changes -Make most installation files read only. -Qualify many calls with std:: to ensure they're not colliding with glm:: -Use auto in more places.
2017-06-09 22:38:06 -04:00
temp = std::atan2(helper.In.y, c1mx);
--User changes -Add new variations: bubbleT3D, crob, hexaplay3D, hexcrop, hexes, hexnix3D, loonie2, loonie3, nBlur, octapol and synth. -Allow for pre/post versions of dc_bubble, dc_cylinder and dc_linear whereas before they were omitted. -When saving a file with multiple embers in it, detect if time values are all the same and if so, start them at zero and increment by 1 for each ember. -Allow for numerous quality increases to be coalesced into one. It will pick up at the end of the current render. -Show selection highlight on variations tree in response to mouse hover. This makes it easier to see for which variation or param the current mouse wheel action will apply. -Make default temporal samples be 100, whereas before it was 1000 which was overkill. -Require the shift key to be held with delete for deleting an ember to prevent it from triggering when the user enters delete in the edit box. -This wasn't otherwise fixable without writing a lot more code. --Bug fixes -EmberGenome was crashing when generating a sequence from a source file with more than 2 embers in it. -EmberGenome was improperly handling the first frame of a merge after the last frame of the loop. -These bugs were due to a previous commit. Revert parts of that commit. -Prevent a zoom value of less than 0 when reading from xml. -Slight optimization of the crescents, and mask variations, if the compiler wasn't doing it already. -Unique file naming was broken because it was looking for _# and the default names ended with -#. -Disallow renaming of an ember in the library tree to an empty string. -Severe bug that prevented some variations from being read correctly from params generated outside this program. -Severe OpenCL randomization bug. The first x coordinates of the first points in the first kernel call of the first ember of a render since the OpenCL renderer object was created were not random and were mostly -1. -Severe bug when populating xform selection distributions that could sometimes cause a crash due to roundoff error. Fix by using double. -Limit the max number of variations in a random ember to MAX_CL_VARS, which is 8. This ensures they'll look the same on CPU and GPU. -Prevent user from saving stylesheet to default.qss, it's a special reserved filename. --Code changes -Generalize using the running sum output point inside of a variation for all cases: pre, reg and post. -Allow for array variables in variations where the address of each element is stored in m_Params. -Qualify all math functions with std:: -No longer use our own Clamp() in OpenCL, instead use the standard clamp(). -Redesign how functions are used in the variations OpenCL code. -Add tests to EmberTester to verify some of the new functionality. -Place more const and override qualifiers on functions where appropriate. -Add a global rand with a lock to be used very sparingly. -Use a map instead of a vector for bad param names in Xml parsing. -Prefix affine interpolation mode defines with "AFFINE_" to make their purpose more clear. -Allow for variations that change state during iteration by sending a separate copy of the ember to each rendering thread. -Implement this same functionality with a local struct in OpenCL. It's members are the total of all variables that need to change state within an ember. -Add Contains() function to Utils.h. -EmberRender: print names of kernels being printed with --dump_kernel option. -Clean up EmberTester to handle some of the recent changes. -Fix various casts. -Replace % 2 with & 1, even though the compiler was likely doing this already. -Add new file Variations06.h to accommodate new variations. -General cleanup.
2015-11-22 17:15:07 -05:00
helper.Out.x = m_Weight * (r * std::cos(temp) + m_C1);
helper.Out.y = m_Weight * r * std::sin(temp);
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
}
else
{
helper.Out.x = m_Weight * helper.In.x;
helper.Out.y = m_Weight * helper.In.y;
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
}
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
}
Numerous fixes 0.4.0.5 Beta 07/18/2014 --User Changes Allow for vibrancy values > 1. Add flatten and unflatten menu items. Automatically flatten like Apophysis does. Add plugin and new_linear tags to Xml to be compatible with Apophysis. --Bug Fixes Fix blur, blur3d, bubble, cropn, cross, curl, curl3d, epispiral, ho, julia3d, julia3dz, loonie, mirror_x, mirror_y, mirror_z, rotate_x, sinusoidal, spherical, spherical3d, stripes. Unique filename on final render was completely broken. Two severe OpenCL bugs. Random seeds were biased and fusing was being reset too often leading to results that differ from the CPU. Subtle, but sometimes severe bug in the setup of the xaos weights. Use properly defined epsilon by getting the value from std::numeric_limits, rather than hard coding 1e-6 or 1e-10. Omit incorrect usage of epsilon everywhere. It should not be automatically added to denominators. Rather, it should only be used if the denominator is zero. Force final render progress bars to 100 on completion. Sometimes they didn't seem to make it there. Make variation name and params comparisons be case insensitive. --Code Changes Make ForEach and FindIf wrappers around std::for_each and std::find_if.
2014-07-19 02:33:18 -04:00
--User changes -Allow for pausing the renderer in the main window. This makes is more efficient when entering many parameters, such as when following a tutorial. -Add support for new variations: erf, gamma, jac_cn, jac_dn, jac_sn, logDB, pressure_wave, pRose3D, splits3D, w, waves2b, x, xerf, y, z. -Inform user of the start and stop of file parsing in EmberAnimate because the files could potentially be very large. -Move the follwing fields to a new table called Animation: Interpolation, Affine Interpolation, Temporal Samples, Temporal Filter Width, Temporal Filter Type. -These currently have no effect on the interactive renderer and instead are used when running flames through EmberGenome to generate sequences, and then animating them in Fractorium or EmberAnimate. -Add new parameter overrides for EmberRender and EmberAnimate which directly assign values to all flames being rendered, rather than scale: --quality --demin --demax --Bug fixes -Left pad instead of right pad names of sequence outputs from EmberGenome. -Unique file naming was broken for files which already had an underscore in them. -Properly report that png is the default format of EmberRender and EmberAnimate output instead of erroneously claiming it was jpg. -Make command line programs search these folders in this order for the palette file: ./ ~/.fractorium ~/.config/fractorium /usr/share/fractorium /usr/local/share/fractorium -Fix possible bad values in hexes. -Better assignment of Z variables. -Fix boarders due to use of poorly implemented rint() function from flam3. Use std::rint() now. -wedge_sph was completely wrong due to having accidentally swapped the mapping of two parameters. -Make juliascope juliascope_power parameter be of type REAL_NONZERO since it's used as a denominator. -Make Z assignment compatible with the originals in: -arch, bcircle, bCollide, bent, bent2, bisplit, blob, blur_linear, blur_square, bMod, boarders, boarders2, bSwirl, bTransform, butterfly, cardioid, cell, circleblur, circlize, circlize2, circus, collideoscope, cos, cosine, cosh, coth, cpow, cpow2, crescents, cropn, csc, csch, curl, curve, dc_gridout, deltaa, diamond, disc2, eclipse, eCollide, edisc, eJulia, elliptic, eMod, eMotion, ennepers, epispiral, ePush, eRotate, eScale, eSwirl, ex, exp, expo, exponential, fan, fdisc, fibonacci, fibonacci2, fisheye, flipcircle, flipy, flower, flux, funnel, glynnia, GlynnSim1, GlynnSim2, GlynnSim3, gridout, handkerchief, heart, hole, idisc, julia, julian2, juliaNab, kaleidoscope, lazyTravis, Lissajous, mask, MobiusN, mobius_strip, modulus, murl, murl2, npolar, ortho, oscilloscope, parabola, perspective, petal, phoenix_julia, pie (was also inconsistent between cpu and gpu), poincare, popcorn, popcorn2, power, pow_block, rational3, rays, rblur, rings, rippled, roundspher, sec, secant2, sigmoid, sin, sineblur, sinh, sinusgrid, sphericaln, spiralwing, spirograph, split, squarize, squirrel, squish, sschecks, starblur, stripes, stwin, super_shape, tan, tancos, tangent, tanh, TwinTrian, twoface, unpolar, waves, wavesn, wedge_julia, whorl, xheart, zblur, zscale. --Code changes -Generalize Variation::PrecalcHelper() and rename to PrePostPrecalcHelper(). --Do the same for the OpenCL version and rename it PrePostPrecalcOpenCLString(). -Rename Variation::m_AssignType to m_PrePostAssignType since it's only relevant to pre/post variations.
2016-01-29 20:02:15 -05:00
helper.Out.z = DefaultZ(helper);
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
}
--Code changes -Better use of const functions. -Ember.pro should not define BUILDING_EMBERCL, instead define BUILDING_EMBER.
2015-06-29 23:13:53 -04:00
virtual string OpenCLString() const override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
ostringstream ss, ss2;
0.4.1.3 Beta 10/14/2014 --User Changes Size is no longer fixed to the window size. Size scaling is done differently in the final render dialog. This fixes several bugs. Remove Xml saving size from settings and options dialog, it no longer applies. Final render can be broken into strips. Set default save path to the desktop if none is found in the settings file. Set default output size to 1920x1080 if none is found in the settings file. --Bug Fixes Better memory size reporting in final render dialog. --Code Changes Migrate to C++11, Qt 5.3.1, and Visual Studio 2013. Change most instances of unsigned int to size_t, and int to intmax_t. Add m_OrigPixPerUnit and m_ScaleType to Ember for scaling purposes. Replace some sprintf_s() calls in XmlToEmber with ostringstream. Move more non-templated members into RendererBase. Add CopyVec() overload that takes a per element function pointer. Add vector Memset(). Replace '&' with '+' instead of "&amp;" in XmlToEmber for much faster parsing. Break strips rendering out into EmberCommon and call from EmberRender and Fractorium. Make AddAndWriteBuffer() just call WriteBuffer(). Make AddAndWriteImage() delete the existing image first before replacing it. Add SetOutputTexture() to RendererCL to support making new textures in response to resize events. Remove multiple return statements in RendererCL, and replace with a bool that tracks results. Add ToDouble(), MakeEnd(), ToString() and Exists() wrappers in Fractorium. Add Size() wrapper in EmberFile. Make QString function arguments const QString&, and string with const string&. Make ShowCritical() wrapper for invoking a message box from another thread. Add combo box to TwoButtonWidget and rename.
2014-10-14 11:53:15 -04:00
intmax_t i = 0, varIndex = IndexInXform();
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
ss2 << "_" << XformIndexInEmber() << "]";
string index = ss2.str();
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
string weight = WeightDefineString();
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
string r1 = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string d1 = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string r2 = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string d2 = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string c1 = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string c2 = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
ss << "\t{\n"
<< "\t\treal_t r, temp, c1mx;\n"
<< "\n"
<< "\t\tif (vIn.x > 0)\n"
<< "\t\t{\n"
<< "\t\t c1mx = " << c1 << " - vIn.x;\n"
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
<< "\t\t r = sqrt(fma(c1mx, c1mx, SQR(vIn.y)));\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\n"
<< "\t\t if (r <= " << r1 << ")\n"
<< "\t\t {\n"
<< "\t\t r *= " << r2 << " / " << r1 << ";\n"
<< "\t\t temp = atan2(vIn.y, c1mx);\n"
<< "\n"
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
<< "\t\t vOut.x = " << weight << " * fma(r, cos(temp), -" << c2 << ");\n"
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
<< "\t\t vOut.y = " << weight << " * r * sin(temp);\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t\t }\n"
<< "\t\t else\n"
<< "\t\t {\n"
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
<< "\t\t vOut.x = " << weight << " * vIn.x;\n"
<< "\t\t vOut.y = " << weight << " * vIn.y;\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t\t }\n"
<< "\t\t}\n"
<< "\t\telse\n"
<< "\t\t{\n"
<< "\t\t c1mx = -" << c2 << " - vIn.x;\n"
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
<< "\t\t r = sqrt(fma(c1mx, c1mx, SQR(vIn.y)));\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\n"
<< "\t\t if (r <= " << r2 << ")\n"
<< "\t\t {\n"
<< "\t\t r *= " << r1 << " / " << r2 << ";\n"
<< "\t\t temp = atan2(vIn.y, c1mx);\n"
<< "\n"
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
<< "\t\t vOut.x = " << weight << " * fma(r, cos(temp), " << c1 << ");\n"
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
<< "\t\t vOut.y = " << weight << " * r * sin(temp);\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t\t }\n"
<< "\t\t else\n"
<< "\t\t {\n"
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
<< "\t\t vOut.x = " << weight << " * vIn.x;\n"
<< "\t\t vOut.y = " << weight << " * vIn.y;\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t\t }\n"
<< "\t\t}\n"
Numerous fixes 0.4.0.5 Beta 07/18/2014 --User Changes Allow for vibrancy values > 1. Add flatten and unflatten menu items. Automatically flatten like Apophysis does. Add plugin and new_linear tags to Xml to be compatible with Apophysis. --Bug Fixes Fix blur, blur3d, bubble, cropn, cross, curl, curl3d, epispiral, ho, julia3d, julia3dz, loonie, mirror_x, mirror_y, mirror_z, rotate_x, sinusoidal, spherical, spherical3d, stripes. Unique filename on final render was completely broken. Two severe OpenCL bugs. Random seeds were biased and fusing was being reset too often leading to results that differ from the CPU. Subtle, but sometimes severe bug in the setup of the xaos weights. Use properly defined epsilon by getting the value from std::numeric_limits, rather than hard coding 1e-6 or 1e-10. Omit incorrect usage of epsilon everywhere. It should not be automatically added to denominators. Rather, it should only be used if the denominator is zero. Force final render progress bars to 100 on completion. Sometimes they didn't seem to make it there. Make variation name and params comparisons be case insensitive. --Code Changes Make ForEach and FindIf wrappers around std::for_each and std::find_if.
2014-07-19 02:33:18 -04:00
<< "\n"
--User changes -Allow for pausing the renderer in the main window. This makes is more efficient when entering many parameters, such as when following a tutorial. -Add support for new variations: erf, gamma, jac_cn, jac_dn, jac_sn, logDB, pressure_wave, pRose3D, splits3D, w, waves2b, x, xerf, y, z. -Inform user of the start and stop of file parsing in EmberAnimate because the files could potentially be very large. -Move the follwing fields to a new table called Animation: Interpolation, Affine Interpolation, Temporal Samples, Temporal Filter Width, Temporal Filter Type. -These currently have no effect on the interactive renderer and instead are used when running flames through EmberGenome to generate sequences, and then animating them in Fractorium or EmberAnimate. -Add new parameter overrides for EmberRender and EmberAnimate which directly assign values to all flames being rendered, rather than scale: --quality --demin --demax --Bug fixes -Left pad instead of right pad names of sequence outputs from EmberGenome. -Unique file naming was broken for files which already had an underscore in them. -Properly report that png is the default format of EmberRender and EmberAnimate output instead of erroneously claiming it was jpg. -Make command line programs search these folders in this order for the palette file: ./ ~/.fractorium ~/.config/fractorium /usr/share/fractorium /usr/local/share/fractorium -Fix possible bad values in hexes. -Better assignment of Z variables. -Fix boarders due to use of poorly implemented rint() function from flam3. Use std::rint() now. -wedge_sph was completely wrong due to having accidentally swapped the mapping of two parameters. -Make juliascope juliascope_power parameter be of type REAL_NONZERO since it's used as a denominator. -Make Z assignment compatible with the originals in: -arch, bcircle, bCollide, bent, bent2, bisplit, blob, blur_linear, blur_square, bMod, boarders, boarders2, bSwirl, bTransform, butterfly, cardioid, cell, circleblur, circlize, circlize2, circus, collideoscope, cos, cosine, cosh, coth, cpow, cpow2, crescents, cropn, csc, csch, curl, curve, dc_gridout, deltaa, diamond, disc2, eclipse, eCollide, edisc, eJulia, elliptic, eMod, eMotion, ennepers, epispiral, ePush, eRotate, eScale, eSwirl, ex, exp, expo, exponential, fan, fdisc, fibonacci, fibonacci2, fisheye, flipcircle, flipy, flower, flux, funnel, glynnia, GlynnSim1, GlynnSim2, GlynnSim3, gridout, handkerchief, heart, hole, idisc, julia, julian2, juliaNab, kaleidoscope, lazyTravis, Lissajous, mask, MobiusN, mobius_strip, modulus, murl, murl2, npolar, ortho, oscilloscope, parabola, perspective, petal, phoenix_julia, pie (was also inconsistent between cpu and gpu), poincare, popcorn, popcorn2, power, pow_block, rational3, rays, rblur, rings, rippled, roundspher, sec, secant2, sigmoid, sin, sineblur, sinh, sinusgrid, sphericaln, spiralwing, spirograph, split, squarize, squirrel, squish, sschecks, starblur, stripes, stwin, super_shape, tan, tancos, tangent, tanh, TwinTrian, twoface, unpolar, waves, wavesn, wedge_julia, whorl, xheart, zblur, zscale. --Code changes -Generalize Variation::PrecalcHelper() and rename to PrePostPrecalcHelper(). --Do the same for the OpenCL version and rename it PrePostPrecalcOpenCLString(). -Rename Variation::m_AssignType to m_PrePostAssignType since it's only relevant to pre/post variations.
2016-01-29 20:02:15 -05:00
<< "\t\tvOut.z = " << DefaultZCl()
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t}\n";
return ss.str();
}
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
virtual void Precalc() override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
m_C1 = m_R1 + m_D1;
m_C2 = m_R2 + m_D2;
}
protected:
void Init()
{
string prefix = Prefix();
m_Params.clear();
--Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums.
2016-01-04 19:50:15 -05:00
m_Params.push_back(ParamWithName<T>(&m_R1, prefix + "trade_r1", 1, eParamType::REAL, EPS, TMAX));
m_Params.push_back(ParamWithName<T>(&m_D1, prefix + "trade_d1", 1, eParamType::REAL, 0, TMAX));
m_Params.push_back(ParamWithName<T>(&m_R2, prefix + "trade_r2", 1, eParamType::REAL, EPS, TMAX));
m_Params.push_back(ParamWithName<T>(&m_D2, prefix + "trade_d2", 1, eParamType::REAL, 0, TMAX));
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
m_Params.push_back(ParamWithName<T>(true, &m_C1, prefix + "trade_c1"));
m_Params.push_back(ParamWithName<T>(true, &m_C2, prefix + "trade_c2"));
}
private:
T m_R1;
T m_D1;
T m_R2;
T m_D2;
T m_C1;//Precalc.
T m_C2;
};
/// <summary>
/// Juliac.
/// </summary>
template <typename T>
--Code changes -Split VariationList into .h/cpp files. --Make VariationList.cpp the only place where the variations files are included from. This speeds up the build, but variations can no longer be directly instantiated. --Variations are no longer exported.
2016-02-18 21:58:24 -05:00
class JuliacVariation : public ParametricVariation<T>
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
public:
--Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums.
2016-01-04 19:50:15 -05:00
JuliacVariation(T weight = 1.0) : ParametricVariation<T>("Juliac", eVariationId::VAR_JULIAC, weight, true, false, false, false, true)
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
Init();
}
PARVARCOPY(JuliacVariation)
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
T arg = helper.m_PrecalcAtanyx + fmod(T(rand.Rand()), T(1 / m_ReInv)) * M_2PI;
--User changes -Add new variations: bubbleT3D, crob, hexaplay3D, hexcrop, hexes, hexnix3D, loonie2, loonie3, nBlur, octapol and synth. -Allow for pre/post versions of dc_bubble, dc_cylinder and dc_linear whereas before they were omitted. -When saving a file with multiple embers in it, detect if time values are all the same and if so, start them at zero and increment by 1 for each ember. -Allow for numerous quality increases to be coalesced into one. It will pick up at the end of the current render. -Show selection highlight on variations tree in response to mouse hover. This makes it easier to see for which variation or param the current mouse wheel action will apply. -Make default temporal samples be 100, whereas before it was 1000 which was overkill. -Require the shift key to be held with delete for deleting an ember to prevent it from triggering when the user enters delete in the edit box. -This wasn't otherwise fixable without writing a lot more code. --Bug fixes -EmberGenome was crashing when generating a sequence from a source file with more than 2 embers in it. -EmberGenome was improperly handling the first frame of a merge after the last frame of the loop. -These bugs were due to a previous commit. Revert parts of that commit. -Prevent a zoom value of less than 0 when reading from xml. -Slight optimization of the crescents, and mask variations, if the compiler wasn't doing it already. -Unique file naming was broken because it was looking for _# and the default names ended with -#. -Disallow renaming of an ember in the library tree to an empty string. -Severe bug that prevented some variations from being read correctly from params generated outside this program. -Severe OpenCL randomization bug. The first x coordinates of the first points in the first kernel call of the first ember of a render since the OpenCL renderer object was created were not random and were mostly -1. -Severe bug when populating xform selection distributions that could sometimes cause a crash due to roundoff error. Fix by using double. -Limit the max number of variations in a random ember to MAX_CL_VARS, which is 8. This ensures they'll look the same on CPU and GPU. -Prevent user from saving stylesheet to default.qss, it's a special reserved filename. --Code changes -Generalize using the running sum output point inside of a variation for all cases: pre, reg and post. -Allow for array variables in variations where the address of each element is stored in m_Params. -Qualify all math functions with std:: -No longer use our own Clamp() in OpenCL, instead use the standard clamp(). -Redesign how functions are used in the variations OpenCL code. -Add tests to EmberTester to verify some of the new functionality. -Place more const and override qualifiers on functions where appropriate. -Add a global rand with a lock to be used very sparingly. -Use a map instead of a vector for bad param names in Xml parsing. -Prefix affine interpolation mode defines with "AFFINE_" to make their purpose more clear. -Allow for variations that change state during iteration by sending a separate copy of the ember to each rendering thread. -Implement this same functionality with a local struct in OpenCL. It's members are the total of all variables that need to change state within an ember. -Add Contains() function to Utils.h. -EmberRender: print names of kernels being printed with --dump_kernel option. -Clean up EmberTester to handle some of the recent changes. -Fix various casts. -Replace % 2 with & 1, even though the compiler was likely doing this already. -Add new file Variations06.h to accommodate new variations. -General cleanup.
2015-11-22 17:15:07 -05:00
T lnmod = m_Dist * T(0.5) * std::log(helper.m_PrecalcSumSquares);
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
T temp = arg * m_ReInv + lnmod * m_Im100;
--User changes -Add new variations: bubbleT3D, crob, hexaplay3D, hexcrop, hexes, hexnix3D, loonie2, loonie3, nBlur, octapol and synth. -Allow for pre/post versions of dc_bubble, dc_cylinder and dc_linear whereas before they were omitted. -When saving a file with multiple embers in it, detect if time values are all the same and if so, start them at zero and increment by 1 for each ember. -Allow for numerous quality increases to be coalesced into one. It will pick up at the end of the current render. -Show selection highlight on variations tree in response to mouse hover. This makes it easier to see for which variation or param the current mouse wheel action will apply. -Make default temporal samples be 100, whereas before it was 1000 which was overkill. -Require the shift key to be held with delete for deleting an ember to prevent it from triggering when the user enters delete in the edit box. -This wasn't otherwise fixable without writing a lot more code. --Bug fixes -EmberGenome was crashing when generating a sequence from a source file with more than 2 embers in it. -EmberGenome was improperly handling the first frame of a merge after the last frame of the loop. -These bugs were due to a previous commit. Revert parts of that commit. -Prevent a zoom value of less than 0 when reading from xml. -Slight optimization of the crescents, and mask variations, if the compiler wasn't doing it already. -Unique file naming was broken because it was looking for _# and the default names ended with -#. -Disallow renaming of an ember in the library tree to an empty string. -Severe bug that prevented some variations from being read correctly from params generated outside this program. -Severe OpenCL randomization bug. The first x coordinates of the first points in the first kernel call of the first ember of a render since the OpenCL renderer object was created were not random and were mostly -1. -Severe bug when populating xform selection distributions that could sometimes cause a crash due to roundoff error. Fix by using double. -Limit the max number of variations in a random ember to MAX_CL_VARS, which is 8. This ensures they'll look the same on CPU and GPU. -Prevent user from saving stylesheet to default.qss, it's a special reserved filename. --Code changes -Generalize using the running sum output point inside of a variation for all cases: pre, reg and post. -Allow for array variables in variations where the address of each element is stored in m_Params. -Qualify all math functions with std:: -No longer use our own Clamp() in OpenCL, instead use the standard clamp(). -Redesign how functions are used in the variations OpenCL code. -Add tests to EmberTester to verify some of the new functionality. -Place more const and override qualifiers on functions where appropriate. -Add a global rand with a lock to be used very sparingly. -Use a map instead of a vector for bad param names in Xml parsing. -Prefix affine interpolation mode defines with "AFFINE_" to make their purpose more clear. -Allow for variations that change state during iteration by sending a separate copy of the ember to each rendering thread. -Implement this same functionality with a local struct in OpenCL. It's members are the total of all variables that need to change state within an ember. -Add Contains() function to Utils.h. -EmberRender: print names of kernels being printed with --dump_kernel option. -Clean up EmberTester to handle some of the recent changes. -Fix various casts. -Replace % 2 with & 1, even though the compiler was likely doing this already. -Add new file Variations06.h to accommodate new variations. -General cleanup.
2015-11-22 17:15:07 -05:00
T mod2 = std::exp(lnmod * m_ReInv - arg * m_Im100);
helper.Out.x = m_Weight * mod2 * std::cos(temp);
helper.Out.y = m_Weight * mod2 * std::sin(temp);
--User changes -Allow for pausing the renderer in the main window. This makes is more efficient when entering many parameters, such as when following a tutorial. -Add support for new variations: erf, gamma, jac_cn, jac_dn, jac_sn, logDB, pressure_wave, pRose3D, splits3D, w, waves2b, x, xerf, y, z. -Inform user of the start and stop of file parsing in EmberAnimate because the files could potentially be very large. -Move the follwing fields to a new table called Animation: Interpolation, Affine Interpolation, Temporal Samples, Temporal Filter Width, Temporal Filter Type. -These currently have no effect on the interactive renderer and instead are used when running flames through EmberGenome to generate sequences, and then animating them in Fractorium or EmberAnimate. -Add new parameter overrides for EmberRender and EmberAnimate which directly assign values to all flames being rendered, rather than scale: --quality --demin --demax --Bug fixes -Left pad instead of right pad names of sequence outputs from EmberGenome. -Unique file naming was broken for files which already had an underscore in them. -Properly report that png is the default format of EmberRender and EmberAnimate output instead of erroneously claiming it was jpg. -Make command line programs search these folders in this order for the palette file: ./ ~/.fractorium ~/.config/fractorium /usr/share/fractorium /usr/local/share/fractorium -Fix possible bad values in hexes. -Better assignment of Z variables. -Fix boarders due to use of poorly implemented rint() function from flam3. Use std::rint() now. -wedge_sph was completely wrong due to having accidentally swapped the mapping of two parameters. -Make juliascope juliascope_power parameter be of type REAL_NONZERO since it's used as a denominator. -Make Z assignment compatible with the originals in: -arch, bcircle, bCollide, bent, bent2, bisplit, blob, blur_linear, blur_square, bMod, boarders, boarders2, bSwirl, bTransform, butterfly, cardioid, cell, circleblur, circlize, circlize2, circus, collideoscope, cos, cosine, cosh, coth, cpow, cpow2, crescents, cropn, csc, csch, curl, curve, dc_gridout, deltaa, diamond, disc2, eclipse, eCollide, edisc, eJulia, elliptic, eMod, eMotion, ennepers, epispiral, ePush, eRotate, eScale, eSwirl, ex, exp, expo, exponential, fan, fdisc, fibonacci, fibonacci2, fisheye, flipcircle, flipy, flower, flux, funnel, glynnia, GlynnSim1, GlynnSim2, GlynnSim3, gridout, handkerchief, heart, hole, idisc, julia, julian2, juliaNab, kaleidoscope, lazyTravis, Lissajous, mask, MobiusN, mobius_strip, modulus, murl, murl2, npolar, ortho, oscilloscope, parabola, perspective, petal, phoenix_julia, pie (was also inconsistent between cpu and gpu), poincare, popcorn, popcorn2, power, pow_block, rational3, rays, rblur, rings, rippled, roundspher, sec, secant2, sigmoid, sin, sineblur, sinh, sinusgrid, sphericaln, spiralwing, spirograph, split, squarize, squirrel, squish, sschecks, starblur, stripes, stwin, super_shape, tan, tancos, tangent, tanh, TwinTrian, twoface, unpolar, waves, wavesn, wedge_julia, whorl, xheart, zblur, zscale. --Code changes -Generalize Variation::PrecalcHelper() and rename to PrePostPrecalcHelper(). --Do the same for the OpenCL version and rename it PrePostPrecalcOpenCLString(). -Rename Variation::m_AssignType to m_PrePostAssignType since it's only relevant to pre/post variations.
2016-01-29 20:02:15 -05:00
helper.Out.z = DefaultZ(helper);
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
}
--Code changes -Better use of const functions. -Ember.pro should not define BUILDING_EMBERCL, instead define BUILDING_EMBER.
2015-06-29 23:13:53 -04:00
virtual string OpenCLString() const override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
ostringstream ss, ss2;
0.4.1.3 Beta 10/14/2014 --User Changes Size is no longer fixed to the window size. Size scaling is done differently in the final render dialog. This fixes several bugs. Remove Xml saving size from settings and options dialog, it no longer applies. Final render can be broken into strips. Set default save path to the desktop if none is found in the settings file. Set default output size to 1920x1080 if none is found in the settings file. --Bug Fixes Better memory size reporting in final render dialog. --Code Changes Migrate to C++11, Qt 5.3.1, and Visual Studio 2013. Change most instances of unsigned int to size_t, and int to intmax_t. Add m_OrigPixPerUnit and m_ScaleType to Ember for scaling purposes. Replace some sprintf_s() calls in XmlToEmber with ostringstream. Move more non-templated members into RendererBase. Add CopyVec() overload that takes a per element function pointer. Add vector Memset(). Replace '&' with '+' instead of "&amp;" in XmlToEmber for much faster parsing. Break strips rendering out into EmberCommon and call from EmberRender and Fractorium. Make AddAndWriteBuffer() just call WriteBuffer(). Make AddAndWriteImage() delete the existing image first before replacing it. Add SetOutputTexture() to RendererCL to support making new textures in response to resize events. Remove multiple return statements in RendererCL, and replace with a bool that tracks results. Add ToDouble(), MakeEnd(), ToString() and Exists() wrappers in Fractorium. Add Size() wrapper in EmberFile. Make QString function arguments const QString&, and string with const string&. Make ShowCritical() wrapper for invoking a message box from another thread. Add combo box to TwoButtonWidget and rename.
2014-10-14 11:53:15 -04:00
intmax_t i = 0, varIndex = IndexInXform();
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
ss2 << "_" << XformIndexInEmber() << "]";
string index = ss2.str();
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
string weight = WeightDefineString();
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
string re = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string im = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string dist = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
string reInv = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string im100 = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
ss << "\t{\n"
--Bug fixes -Ensure all OpenCL functions that take multiple arguments cast literal values to real_t.
2019-05-23 22:44:27 -04:00
<< "\t\treal_t arg = fma(fmod((real_t)MwcNext(mwc), (real_t)((real_t)(1.0) / " << reInv << ")), M_2PI, precalcAtanyx);\n"
Ditch T(f) define in opencl programs in preference to already-defined real_t. Also removed some erroneously-added double-'T(' statements
2015-02-26 16:18:55 -05:00
<< "\t\treal_t lnmod = " << dist << " * (real_t)(0.5) * log(precalcSumSquares);\n"
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
<< "\t\treal_t temp = fma(arg, " << reInv << ", lnmod * " << im100 << ");\n"
<< "\t\treal_t mod2 = exp(fma(lnmod, " << reInv << ", -(arg * " << im100 << ")));\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\n"
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
<< "\t\tvOut.x = " << weight << " * mod2 * cos(temp);\n"
<< "\t\tvOut.y = " << weight << " * mod2 * sin(temp);\n"
--User changes -Allow for pausing the renderer in the main window. This makes is more efficient when entering many parameters, such as when following a tutorial. -Add support for new variations: erf, gamma, jac_cn, jac_dn, jac_sn, logDB, pressure_wave, pRose3D, splits3D, w, waves2b, x, xerf, y, z. -Inform user of the start and stop of file parsing in EmberAnimate because the files could potentially be very large. -Move the follwing fields to a new table called Animation: Interpolation, Affine Interpolation, Temporal Samples, Temporal Filter Width, Temporal Filter Type. -These currently have no effect on the interactive renderer and instead are used when running flames through EmberGenome to generate sequences, and then animating them in Fractorium or EmberAnimate. -Add new parameter overrides for EmberRender and EmberAnimate which directly assign values to all flames being rendered, rather than scale: --quality --demin --demax --Bug fixes -Left pad instead of right pad names of sequence outputs from EmberGenome. -Unique file naming was broken for files which already had an underscore in them. -Properly report that png is the default format of EmberRender and EmberAnimate output instead of erroneously claiming it was jpg. -Make command line programs search these folders in this order for the palette file: ./ ~/.fractorium ~/.config/fractorium /usr/share/fractorium /usr/local/share/fractorium -Fix possible bad values in hexes. -Better assignment of Z variables. -Fix boarders due to use of poorly implemented rint() function from flam3. Use std::rint() now. -wedge_sph was completely wrong due to having accidentally swapped the mapping of two parameters. -Make juliascope juliascope_power parameter be of type REAL_NONZERO since it's used as a denominator. -Make Z assignment compatible with the originals in: -arch, bcircle, bCollide, bent, bent2, bisplit, blob, blur_linear, blur_square, bMod, boarders, boarders2, bSwirl, bTransform, butterfly, cardioid, cell, circleblur, circlize, circlize2, circus, collideoscope, cos, cosine, cosh, coth, cpow, cpow2, crescents, cropn, csc, csch, curl, curve, dc_gridout, deltaa, diamond, disc2, eclipse, eCollide, edisc, eJulia, elliptic, eMod, eMotion, ennepers, epispiral, ePush, eRotate, eScale, eSwirl, ex, exp, expo, exponential, fan, fdisc, fibonacci, fibonacci2, fisheye, flipcircle, flipy, flower, flux, funnel, glynnia, GlynnSim1, GlynnSim2, GlynnSim3, gridout, handkerchief, heart, hole, idisc, julia, julian2, juliaNab, kaleidoscope, lazyTravis, Lissajous, mask, MobiusN, mobius_strip, modulus, murl, murl2, npolar, ortho, oscilloscope, parabola, perspective, petal, phoenix_julia, pie (was also inconsistent between cpu and gpu), poincare, popcorn, popcorn2, power, pow_block, rational3, rays, rblur, rings, rippled, roundspher, sec, secant2, sigmoid, sin, sineblur, sinh, sinusgrid, sphericaln, spiralwing, spirograph, split, squarize, squirrel, squish, sschecks, starblur, stripes, stwin, super_shape, tan, tancos, tangent, tanh, TwinTrian, twoface, unpolar, waves, wavesn, wedge_julia, whorl, xheart, zblur, zscale. --Code changes -Generalize Variation::PrecalcHelper() and rename to PrePostPrecalcHelper(). --Do the same for the OpenCL version and rename it PrePostPrecalcOpenCLString(). -Rename Variation::m_AssignType to m_PrePostAssignType since it's only relevant to pre/post variations.
2016-01-29 20:02:15 -05:00
<< "\t\tvOut.z = " << DefaultZCl()
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t}\n";
return ss.str();
}
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
virtual void Precalc() override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
Numerous fixes 0.4.0.5 Beta 07/18/2014 --User Changes Allow for vibrancy values > 1. Add flatten and unflatten menu items. Automatically flatten like Apophysis does. Add plugin and new_linear tags to Xml to be compatible with Apophysis. --Bug Fixes Fix blur, blur3d, bubble, cropn, cross, curl, curl3d, epispiral, ho, julia3d, julia3dz, loonie, mirror_x, mirror_y, mirror_z, rotate_x, sinusoidal, spherical, spherical3d, stripes. Unique filename on final render was completely broken. Two severe OpenCL bugs. Random seeds were biased and fusing was being reset too often leading to results that differ from the CPU. Subtle, but sometimes severe bug in the setup of the xaos weights. Use properly defined epsilon by getting the value from std::numeric_limits, rather than hard coding 1e-6 or 1e-10. Omit incorrect usage of epsilon everywhere. It should not be automatically added to denominators. Rather, it should only be used if the denominator is zero. Force final render progress bars to 100 on completion. Sometimes they didn't seem to make it there. Make variation name and params comparisons be case insensitive. --Code Changes Make ForEach and FindIf wrappers around std::for_each and std::find_if.
2014-07-19 02:33:18 -04:00
m_ReInv = 1 / Zeps(m_Re);
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
m_Im100 = m_Im * T(0.01);
}
protected:
void Init()
{
string prefix = Prefix();
m_Params.clear();
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
m_Params.push_back(ParamWithName<T>(&m_Re, prefix + "Juliac_re", 2));
m_Params.push_back(ParamWithName<T>(&m_Im, prefix + "Juliac_im", 1));
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
m_Params.push_back(ParamWithName<T>(&m_Dist, prefix + "Juliac_dist", 1));
m_Params.push_back(ParamWithName<T>(true, &m_ReInv, prefix + "Juliac_re_inv"));
m_Params.push_back(ParamWithName<T>(true, &m_Im100, prefix + "Juliac_im100"));
}
private:
T m_Re;
T m_Im;
T m_Dist;
T m_ReInv;
T m_Im100;
};
/// <summary>
/// blade3D.
/// </summary>
template <typename T>
--Code changes -Split VariationList into .h/cpp files. --Make VariationList.cpp the only place where the variations files are included from. This speeds up the build, but variations can no longer be directly instantiated. --Variations are no longer exported.
2016-02-18 21:58:24 -05:00
class Blade3DVariation : public Variation<T>
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
public:
--Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums.
2016-01-04 19:50:15 -05:00
Blade3DVariation(T weight = 1.0) : Variation<T>("blade3D", eVariationId::VAR_BLADE3D, weight, true, true) { }
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
VARCOPY(Blade3DVariation)
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
T r = rand.Frand01<T>() * m_Weight * helper.m_PrecalcSqrtSumSquares;
T sinr, cosr;
sincos(r, &sinr, &cosr);
helper.Out.x = m_Weight * helper.In.x * (cosr + sinr);
helper.Out.y = m_Weight * helper.In.x * (cosr - sinr);
helper.Out.z = m_Weight * helper.In.z * (sinr - cosr);
}
--Code changes -Better use of const functions. -Ember.pro should not define BUILDING_EMBERCL, instead define BUILDING_EMBER.
2015-06-29 23:13:53 -04:00
virtual string OpenCLString() const override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
ostringstream ss;
0.4.1.3 Beta 10/14/2014 --User Changes Size is no longer fixed to the window size. Size scaling is done differently in the final render dialog. This fixes several bugs. Remove Xml saving size from settings and options dialog, it no longer applies. Final render can be broken into strips. Set default save path to the desktop if none is found in the settings file. Set default output size to 1920x1080 if none is found in the settings file. --Bug Fixes Better memory size reporting in final render dialog. --Code Changes Migrate to C++11, Qt 5.3.1, and Visual Studio 2013. Change most instances of unsigned int to size_t, and int to intmax_t. Add m_OrigPixPerUnit and m_ScaleType to Ember for scaling purposes. Replace some sprintf_s() calls in XmlToEmber with ostringstream. Move more non-templated members into RendererBase. Add CopyVec() overload that takes a per element function pointer. Add vector Memset(). Replace '&' with '+' instead of "&amp;" in XmlToEmber for much faster parsing. Break strips rendering out into EmberCommon and call from EmberRender and Fractorium. Make AddAndWriteBuffer() just call WriteBuffer(). Make AddAndWriteImage() delete the existing image first before replacing it. Add SetOutputTexture() to RendererCL to support making new textures in response to resize events. Remove multiple return statements in RendererCL, and replace with a bool that tracks results. Add ToDouble(), MakeEnd(), ToString() and Exists() wrappers in Fractorium. Add Size() wrapper in EmberFile. Make QString function arguments const QString&, and string with const string&. Make ShowCritical() wrapper for invoking a message box from another thread. Add combo box to TwoButtonWidget and rename.
2014-10-14 11:53:15 -04:00
intmax_t varIndex = IndexInXform();
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
string weight = WeightDefineString();
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
ss << "\t{\n"
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
<< "\t\treal_t r = MwcNext01(mwc) * " << weight << " * precalcSqrtSumSquares;\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t\treal_t sinr = sin(r);\n"
<< "\t\treal_t cosr = cos(r);\n"
<< "\n"
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
<< "\t\tvOut.x = " << weight << " * vIn.x * (cosr + sinr);\n"
<< "\t\tvOut.y = " << weight << " * vIn.x * (cosr - sinr);\n"
<< "\t\tvOut.z = " << weight << " * vIn.z * (sinr - cosr);\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t}\n";
return ss.str();
}
};
/// <summary>
/// Blob3D.
/// </summary>
template <typename T>
--Code changes -Split VariationList into .h/cpp files. --Make VariationList.cpp the only place where the variations files are included from. This speeds up the build, but variations can no longer be directly instantiated. --Variations are no longer exported.
2016-02-18 21:58:24 -05:00
class Blob3DVariation : public ParametricVariation<T>
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
public:
--Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums.
2016-01-04 19:50:15 -05:00
Blob3DVariation(T weight = 1.0) : ParametricVariation<T>("blob3D", eVariationId::VAR_BLOB3D, weight, true, true, true, true)
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
Init();
}
PARVARCOPY(Blob3DVariation)
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
--User changes -Add new variations: bubbleT3D, crob, hexaplay3D, hexcrop, hexes, hexnix3D, loonie2, loonie3, nBlur, octapol and synth. -Allow for pre/post versions of dc_bubble, dc_cylinder and dc_linear whereas before they were omitted. -When saving a file with multiple embers in it, detect if time values are all the same and if so, start them at zero and increment by 1 for each ember. -Allow for numerous quality increases to be coalesced into one. It will pick up at the end of the current render. -Show selection highlight on variations tree in response to mouse hover. This makes it easier to see for which variation or param the current mouse wheel action will apply. -Make default temporal samples be 100, whereas before it was 1000 which was overkill. -Require the shift key to be held with delete for deleting an ember to prevent it from triggering when the user enters delete in the edit box. -This wasn't otherwise fixable without writing a lot more code. --Bug fixes -EmberGenome was crashing when generating a sequence from a source file with more than 2 embers in it. -EmberGenome was improperly handling the first frame of a merge after the last frame of the loop. -These bugs were due to a previous commit. Revert parts of that commit. -Prevent a zoom value of less than 0 when reading from xml. -Slight optimization of the crescents, and mask variations, if the compiler wasn't doing it already. -Unique file naming was broken because it was looking for _# and the default names ended with -#. -Disallow renaming of an ember in the library tree to an empty string. -Severe bug that prevented some variations from being read correctly from params generated outside this program. -Severe OpenCL randomization bug. The first x coordinates of the first points in the first kernel call of the first ember of a render since the OpenCL renderer object was created were not random and were mostly -1. -Severe bug when populating xform selection distributions that could sometimes cause a crash due to roundoff error. Fix by using double. -Limit the max number of variations in a random ember to MAX_CL_VARS, which is 8. This ensures they'll look the same on CPU and GPU. -Prevent user from saving stylesheet to default.qss, it's a special reserved filename. --Code changes -Generalize using the running sum output point inside of a variation for all cases: pre, reg and post. -Allow for array variables in variations where the address of each element is stored in m_Params. -Qualify all math functions with std:: -No longer use our own Clamp() in OpenCL, instead use the standard clamp(). -Redesign how functions are used in the variations OpenCL code. -Add tests to EmberTester to verify some of the new functionality. -Place more const and override qualifiers on functions where appropriate. -Add a global rand with a lock to be used very sparingly. -Use a map instead of a vector for bad param names in Xml parsing. -Prefix affine interpolation mode defines with "AFFINE_" to make their purpose more clear. -Allow for variations that change state during iteration by sending a separate copy of the ember to each rendering thread. -Implement this same functionality with a local struct in OpenCL. It's members are the total of all variables that need to change state within an ember. -Add Contains() function to Utils.h. -EmberRender: print names of kernels being printed with --dump_kernel option. -Clean up EmberTester to handle some of the recent changes. -Fix various casts. -Replace % 2 with & 1, even though the compiler was likely doing this already. -Add new file Variations06.h to accommodate new variations. -General cleanup.
2015-11-22 17:15:07 -05:00
T r = helper.m_PrecalcSqrtSumSquares * (m_BlobLow + m_BlobDiff * (T(0.5) + T(0.5) * std::sin(m_BlobWaves * helper.m_PrecalcAtanxy)));
--Bug fixes -Fix variations: blob2 (broken on CPU), epispiral, hole. -Fix reading variations from Apophysis: supershape. -Bump render was broken. --Code changes -Flip precalc sina/cosa and apply everywhere by flipping the usage. Flam3 had these reversed and it made the code confusing to read.
2019-12-28 00:04:41 -05:00
helper.Out.x = m_Weight * helper.m_PrecalcCosa * r;//Flipped from original JWildfire plugin which did atan2(x, y) then sin, cos.
helper.Out.y = m_Weight * helper.m_PrecalcSina * r;//Here we do atan(y, x) then cos, sin.
--User changes -Add new variations: bubbleT3D, crob, hexaplay3D, hexcrop, hexes, hexnix3D, loonie2, loonie3, nBlur, octapol and synth. -Allow for pre/post versions of dc_bubble, dc_cylinder and dc_linear whereas before they were omitted. -When saving a file with multiple embers in it, detect if time values are all the same and if so, start them at zero and increment by 1 for each ember. -Allow for numerous quality increases to be coalesced into one. It will pick up at the end of the current render. -Show selection highlight on variations tree in response to mouse hover. This makes it easier to see for which variation or param the current mouse wheel action will apply. -Make default temporal samples be 100, whereas before it was 1000 which was overkill. -Require the shift key to be held with delete for deleting an ember to prevent it from triggering when the user enters delete in the edit box. -This wasn't otherwise fixable without writing a lot more code. --Bug fixes -EmberGenome was crashing when generating a sequence from a source file with more than 2 embers in it. -EmberGenome was improperly handling the first frame of a merge after the last frame of the loop. -These bugs were due to a previous commit. Revert parts of that commit. -Prevent a zoom value of less than 0 when reading from xml. -Slight optimization of the crescents, and mask variations, if the compiler wasn't doing it already. -Unique file naming was broken because it was looking for _# and the default names ended with -#. -Disallow renaming of an ember in the library tree to an empty string. -Severe bug that prevented some variations from being read correctly from params generated outside this program. -Severe OpenCL randomization bug. The first x coordinates of the first points in the first kernel call of the first ember of a render since the OpenCL renderer object was created were not random and were mostly -1. -Severe bug when populating xform selection distributions that could sometimes cause a crash due to roundoff error. Fix by using double. -Limit the max number of variations in a random ember to MAX_CL_VARS, which is 8. This ensures they'll look the same on CPU and GPU. -Prevent user from saving stylesheet to default.qss, it's a special reserved filename. --Code changes -Generalize using the running sum output point inside of a variation for all cases: pre, reg and post. -Allow for array variables in variations where the address of each element is stored in m_Params. -Qualify all math functions with std:: -No longer use our own Clamp() in OpenCL, instead use the standard clamp(). -Redesign how functions are used in the variations OpenCL code. -Add tests to EmberTester to verify some of the new functionality. -Place more const and override qualifiers on functions where appropriate. -Add a global rand with a lock to be used very sparingly. -Use a map instead of a vector for bad param names in Xml parsing. -Prefix affine interpolation mode defines with "AFFINE_" to make their purpose more clear. -Allow for variations that change state during iteration by sending a separate copy of the ember to each rendering thread. -Implement this same functionality with a local struct in OpenCL. It's members are the total of all variables that need to change state within an ember. -Add Contains() function to Utils.h. -EmberRender: print names of kernels being printed with --dump_kernel option. -Clean up EmberTester to handle some of the recent changes. -Fix various casts. -Replace % 2 with & 1, even though the compiler was likely doing this already. -Add new file Variations06.h to accommodate new variations. -General cleanup.
2015-11-22 17:15:07 -05:00
helper.Out.z = m_Weight * std::sin(m_BlobWaves * helper.m_PrecalcAtanxy) * r;
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
}
--Code changes -Better use of const functions. -Ember.pro should not define BUILDING_EMBERCL, instead define BUILDING_EMBER.
2015-06-29 23:13:53 -04:00
virtual string OpenCLString() const override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
ostringstream ss, ss2;
0.4.1.3 Beta 10/14/2014 --User Changes Size is no longer fixed to the window size. Size scaling is done differently in the final render dialog. This fixes several bugs. Remove Xml saving size from settings and options dialog, it no longer applies. Final render can be broken into strips. Set default save path to the desktop if none is found in the settings file. Set default output size to 1920x1080 if none is found in the settings file. --Bug Fixes Better memory size reporting in final render dialog. --Code Changes Migrate to C++11, Qt 5.3.1, and Visual Studio 2013. Change most instances of unsigned int to size_t, and int to intmax_t. Add m_OrigPixPerUnit and m_ScaleType to Ember for scaling purposes. Replace some sprintf_s() calls in XmlToEmber with ostringstream. Move more non-templated members into RendererBase. Add CopyVec() overload that takes a per element function pointer. Add vector Memset(). Replace '&' with '+' instead of "&amp;" in XmlToEmber for much faster parsing. Break strips rendering out into EmberCommon and call from EmberRender and Fractorium. Make AddAndWriteBuffer() just call WriteBuffer(). Make AddAndWriteImage() delete the existing image first before replacing it. Add SetOutputTexture() to RendererCL to support making new textures in response to resize events. Remove multiple return statements in RendererCL, and replace with a bool that tracks results. Add ToDouble(), MakeEnd(), ToString() and Exists() wrappers in Fractorium. Add Size() wrapper in EmberFile. Make QString function arguments const QString&, and string with const string&. Make ShowCritical() wrapper for invoking a message box from another thread. Add combo box to TwoButtonWidget and rename.
2014-10-14 11:53:15 -04:00
intmax_t i = 0, varIndex = IndexInXform();
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
ss2 << "_" << XformIndexInEmber() << "]";
string index = ss2.str();
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
string weight = WeightDefineString();
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
string blobLow = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string blobHigh = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
string blobWaves = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
string blobDiff = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
ss << "\t{\n"
--User changes -Add buttons to copy and paste affine transforms. -Show xform names on the column headers of the xaos table. -Add a color-coded third column to the variations tree which shows any properties of each variation which are non-standard. -Draw a transparent circle over hovered xforms. -Change how xforms respond to dragging. Rotate only is now the default, and scale will only happen with shift. --Optionally do scale and rotate when holding shift, via a setting in the options dialog. --Bug fixes -Snapping when dragging was wrong sometimes. -The program would very rarely crash on startup due to some values being in an uninitialized state. --Code changes -Change almost every variation to use fma() in OpenCL when doing computations of the form a * b + c. This provides a slight speedup, mostly in double precision mode. -Also apply fma() to affine calcs. -Cleanup of OpenGL affine drawing code. -Separate the concept of hovering and selecting xforms.
2018-09-15 06:11:12 -04:00
<< "\t\treal_t r = precalcSqrtSumSquares * fma(" << blobDiff << ", fma((real_t)(0.5), sin(" << blobWaves << " * precalcAtanxy), (real_t)(0.5)), " << blobLow << ");\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\n"
--Bug fixes -Fix variations: blob2 (broken on CPU), epispiral, hole. -Fix reading variations from Apophysis: supershape. -Bump render was broken. --Code changes -Flip precalc sina/cosa and apply everywhere by flipping the usage. Flam3 had these reversed and it made the code confusing to read.
2019-12-28 00:04:41 -05:00
<< "\t\tvOut.x = " << weight << " * (precalcCosa * r);\n"
<< "\t\tvOut.y = " << weight << " * (precalcSina * r);\n"
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
<< "\t\tvOut.z = " << weight << " * (sin(" << blobWaves << " * precalcAtanxy) * r);\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t}\n";
return ss.str();
}
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
virtual void Precalc() override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
m_BlobDiff = m_BlobHigh - m_BlobLow;
}
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
virtual void Random(QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
m_BlobLow = T(0.2) + T(0.5) * rand.Frand01<T>();
m_BlobHigh = T(0.8) + T(0.4) * rand.Frand01<T>();
More Linux work. Convert all casts to new style, away from legacy.
2014-12-07 02:51:44 -05:00
m_BlobWaves = T(int(2 + 5 * rand.Frand01<T>()));
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
}
protected:
void Init()
{
string prefix = Prefix();
m_Params.clear();
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
m_Params.push_back(ParamWithName<T>(&m_BlobLow, prefix + "blob3D_low"));
m_Params.push_back(ParamWithName<T>(&m_BlobHigh, prefix + "blob3D_high", 1));
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
m_Params.push_back(ParamWithName<T>(&m_BlobWaves, prefix + "blob3D_waves", 1));
m_Params.push_back(ParamWithName<T>(true, &m_BlobDiff, prefix + "blob3D_diff"));//Precalc.
}
private:
T m_BlobLow;
T m_BlobHigh;
T m_BlobWaves;
T m_BlobDiff;//Precalc.
};
/// <summary>
/// blocky.
/// </summary>
template <typename T>
--Code changes -Split VariationList into .h/cpp files. --Make VariationList.cpp the only place where the variations files are included from. This speeds up the build, but variations can no longer be directly instantiated. --Variations are no longer exported.
2016-02-18 21:58:24 -05:00
class BlockyVariation : public ParametricVariation<T>
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
public:
--Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums.
2016-01-04 19:50:15 -05:00
BlockyVariation(T weight = 1.0) : ParametricVariation<T>("blocky", eVariationId::VAR_BLOCKY, weight, true)
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
Init();
}
PARVARCOPY(BlockyVariation)
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
--User changes -Add new variations: bubbleT3D, crob, hexaplay3D, hexcrop, hexes, hexnix3D, loonie2, loonie3, nBlur, octapol and synth. -Allow for pre/post versions of dc_bubble, dc_cylinder and dc_linear whereas before they were omitted. -When saving a file with multiple embers in it, detect if time values are all the same and if so, start them at zero and increment by 1 for each ember. -Allow for numerous quality increases to be coalesced into one. It will pick up at the end of the current render. -Show selection highlight on variations tree in response to mouse hover. This makes it easier to see for which variation or param the current mouse wheel action will apply. -Make default temporal samples be 100, whereas before it was 1000 which was overkill. -Require the shift key to be held with delete for deleting an ember to prevent it from triggering when the user enters delete in the edit box. -This wasn't otherwise fixable without writing a lot more code. --Bug fixes -EmberGenome was crashing when generating a sequence from a source file with more than 2 embers in it. -EmberGenome was improperly handling the first frame of a merge after the last frame of the loop. -These bugs were due to a previous commit. Revert parts of that commit. -Prevent a zoom value of less than 0 when reading from xml. -Slight optimization of the crescents, and mask variations, if the compiler wasn't doing it already. -Unique file naming was broken because it was looking for _# and the default names ended with -#. -Disallow renaming of an ember in the library tree to an empty string. -Severe bug that prevented some variations from being read correctly from params generated outside this program. -Severe OpenCL randomization bug. The first x coordinates of the first points in the first kernel call of the first ember of a render since the OpenCL renderer object was created were not random and were mostly -1. -Severe bug when populating xform selection distributions that could sometimes cause a crash due to roundoff error. Fix by using double. -Limit the max number of variations in a random ember to MAX_CL_VARS, which is 8. This ensures they'll look the same on CPU and GPU. -Prevent user from saving stylesheet to default.qss, it's a special reserved filename. --Code changes -Generalize using the running sum output point inside of a variation for all cases: pre, reg and post. -Allow for array variables in variations where the address of each element is stored in m_Params. -Qualify all math functions with std:: -No longer use our own Clamp() in OpenCL, instead use the standard clamp(). -Redesign how functions are used in the variations OpenCL code. -Add tests to EmberTester to verify some of the new functionality. -Place more const and override qualifiers on functions where appropriate. -Add a global rand with a lock to be used very sparingly. -Use a map instead of a vector for bad param names in Xml parsing. -Prefix affine interpolation mode defines with "AFFINE_" to make their purpose more clear. -Allow for variations that change state during iteration by sending a separate copy of the ember to each rendering thread. -Implement this same functionality with a local struct in OpenCL. It's members are the total of all variables that need to change state within an ember. -Add Contains() function to Utils.h. -EmberRender: print names of kernels being printed with --dump_kernel option. -Clean up EmberTester to handle some of the recent changes. -Fix various casts. -Replace % 2 with & 1, even though the compiler was likely doing this already. -Add new file Variations06.h to accommodate new variations. -General cleanup.
2015-11-22 17:15:07 -05:00
T t = Zeps((std::cos(helper.In.x) + std::cos(helper.In.y)) / m_Mp + 1);
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
T r = m_Weight / t;
T tmp = helper.m_PrecalcSumSquares + 1;
T x2 = 2 * helper.In.x;
T y2 = 2 * helper.In.y;
Explicitly namespace calls to sqrt and isnan osx/linux have alternative top-level implementations of these methods, so the calls are ambiguous given the - so these should be used explicitly
2015-04-13 07:32:58 -04:00
T xmax = T(0.5) * (std::sqrt(tmp + x2) + std::sqrt(tmp - x2));
T ymax = T(0.5) * (std::sqrt(tmp + y2) + std::sqrt(tmp - y2));
Numerous fixes 0.4.0.5 Beta 07/18/2014 --User Changes Allow for vibrancy values > 1. Add flatten and unflatten menu items. Automatically flatten like Apophysis does. Add plugin and new_linear tags to Xml to be compatible with Apophysis. --Bug Fixes Fix blur, blur3d, bubble, cropn, cross, curl, curl3d, epispiral, ho, julia3d, julia3dz, loonie, mirror_x, mirror_y, mirror_z, rotate_x, sinusoidal, spherical, spherical3d, stripes. Unique filename on final render was completely broken. Two severe OpenCL bugs. Random seeds were biased and fusing was being reset too often leading to results that differ from the CPU. Subtle, but sometimes severe bug in the setup of the xaos weights. Use properly defined epsilon by getting the value from std::numeric_limits, rather than hard coding 1e-6 or 1e-10. Omit incorrect usage of epsilon everywhere. It should not be automatically added to denominators. Rather, it should only be used if the denominator is zero. Force final render progress bars to 100 on completion. Sometimes they didn't seem to make it there. Make variation name and params comparisons be case insensitive. --Code Changes Make ForEach and FindIf wrappers around std::for_each and std::find_if.
2014-07-19 02:33:18 -04:00
T a = helper.In.x / Zeps(xmax);
--Code changes -Make usage of VarFuncs more consistent by moving some global functions there. -Make CircleTrans1 a little safer by calling Zeps() on m_Sc during precalc. -Begin prepping for 0.9.9.5 release.
2016-03-04 21:54:06 -05:00
T b = VarFuncs<T>::SafeSqrt(1 - SQR(a));
06/09/2017 --User changes -dark.qss is now per-OS. -Properly set/reload palette when coming from the palette editor. The latter must be done if they've modified the current palette even if they've clicked cancel. --Bug fixes -Make the following variations safer by using Zeps(): conic, bipolar, edisc, whorl, tan, csc, cot, tanh, sech, csch, coth, auger, bwraps, hypertile3d, hypertile3d1, ortho, poincare, rational3, barycentroid, sschecks, cscq, cschq, scry_3D, splitbrdr, hexcrop, nblur, crob. -Fix bug enabling/disabling overwrite button in palette editor. -Small optimization for gdoffs, use precalcAtanYX. -Properly propagate z through circlesplit, cylinder2 and tile_log variations. -Some values in truchet_fill could've been NaN. --Code changes -Make most installation files read only. -Qualify many calls with std:: to ensure they're not colliding with glm:: -Use auto in more places.
2017-06-09 22:38:06 -04:00
helper.Out.x = m_Vx * std::atan2(a, b) * r;
Numerous fixes 0.4.0.5 Beta 07/18/2014 --User Changes Allow for vibrancy values > 1. Add flatten and unflatten menu items. Automatically flatten like Apophysis does. Add plugin and new_linear tags to Xml to be compatible with Apophysis. --Bug Fixes Fix blur, blur3d, bubble, cropn, cross, curl, curl3d, epispiral, ho, julia3d, julia3dz, loonie, mirror_x, mirror_y, mirror_z, rotate_x, sinusoidal, spherical, spherical3d, stripes. Unique filename on final render was completely broken. Two severe OpenCL bugs. Random seeds were biased and fusing was being reset too often leading to results that differ from the CPU. Subtle, but sometimes severe bug in the setup of the xaos weights. Use properly defined epsilon by getting the value from std::numeric_limits, rather than hard coding 1e-6 or 1e-10. Omit incorrect usage of epsilon everywhere. It should not be automatically added to denominators. Rather, it should only be used if the denominator is zero. Force final render progress bars to 100 on completion. Sometimes they didn't seem to make it there. Make variation name and params comparisons be case insensitive. --Code Changes Make ForEach and FindIf wrappers around std::for_each and std::find_if.
2014-07-19 02:33:18 -04:00
a = helper.In.y / Zeps(ymax);
--Code changes -Make usage of VarFuncs more consistent by moving some global functions there. -Make CircleTrans1 a little safer by calling Zeps() on m_Sc during precalc. -Begin prepping for 0.9.9.5 release.
2016-03-04 21:54:06 -05:00
b = VarFuncs<T>::SafeSqrt(1 - SQR(a));
06/09/2017 --User changes -dark.qss is now per-OS. -Properly set/reload palette when coming from the palette editor. The latter must be done if they've modified the current palette even if they've clicked cancel. --Bug fixes -Make the following variations safer by using Zeps(): conic, bipolar, edisc, whorl, tan, csc, cot, tanh, sech, csch, coth, auger, bwraps, hypertile3d, hypertile3d1, ortho, poincare, rational3, barycentroid, sschecks, cscq, cschq, scry_3D, splitbrdr, hexcrop, nblur, crob. -Fix bug enabling/disabling overwrite button in palette editor. -Small optimization for gdoffs, use precalcAtanYX. -Properly propagate z through circlesplit, cylinder2 and tile_log variations. -Some values in truchet_fill could've been NaN. --Code changes -Make most installation files read only. -Qualify many calls with std:: to ensure they're not colliding with glm:: -Use auto in more places.
2017-06-09 22:38:06 -04:00
helper.Out.y = m_Vy * std::atan2(a, b) * r;
--User changes -Allow for pausing the renderer in the main window. This makes is more efficient when entering many parameters, such as when following a tutorial. -Add support for new variations: erf, gamma, jac_cn, jac_dn, jac_sn, logDB, pressure_wave, pRose3D, splits3D, w, waves2b, x, xerf, y, z. -Inform user of the start and stop of file parsing in EmberAnimate because the files could potentially be very large. -Move the follwing fields to a new table called Animation: Interpolation, Affine Interpolation, Temporal Samples, Temporal Filter Width, Temporal Filter Type. -These currently have no effect on the interactive renderer and instead are used when running flames through EmberGenome to generate sequences, and then animating them in Fractorium or EmberAnimate. -Add new parameter overrides for EmberRender and EmberAnimate which directly assign values to all flames being rendered, rather than scale: --quality --demin --demax --Bug fixes -Left pad instead of right pad names of sequence outputs from EmberGenome. -Unique file naming was broken for files which already had an underscore in them. -Properly report that png is the default format of EmberRender and EmberAnimate output instead of erroneously claiming it was jpg. -Make command line programs search these folders in this order for the palette file: ./ ~/.fractorium ~/.config/fractorium /usr/share/fractorium /usr/local/share/fractorium -Fix possible bad values in hexes. -Better assignment of Z variables. -Fix boarders due to use of poorly implemented rint() function from flam3. Use std::rint() now. -wedge_sph was completely wrong due to having accidentally swapped the mapping of two parameters. -Make juliascope juliascope_power parameter be of type REAL_NONZERO since it's used as a denominator. -Make Z assignment compatible with the originals in: -arch, bcircle, bCollide, bent, bent2, bisplit, blob, blur_linear, blur_square, bMod, boarders, boarders2, bSwirl, bTransform, butterfly, cardioid, cell, circleblur, circlize, circlize2, circus, collideoscope, cos, cosine, cosh, coth, cpow, cpow2, crescents, cropn, csc, csch, curl, curve, dc_gridout, deltaa, diamond, disc2, eclipse, eCollide, edisc, eJulia, elliptic, eMod, eMotion, ennepers, epispiral, ePush, eRotate, eScale, eSwirl, ex, exp, expo, exponential, fan, fdisc, fibonacci, fibonacci2, fisheye, flipcircle, flipy, flower, flux, funnel, glynnia, GlynnSim1, GlynnSim2, GlynnSim3, gridout, handkerchief, heart, hole, idisc, julia, julian2, juliaNab, kaleidoscope, lazyTravis, Lissajous, mask, MobiusN, mobius_strip, modulus, murl, murl2, npolar, ortho, oscilloscope, parabola, perspective, petal, phoenix_julia, pie (was also inconsistent between cpu and gpu), poincare, popcorn, popcorn2, power, pow_block, rational3, rays, rblur, rings, rippled, roundspher, sec, secant2, sigmoid, sin, sineblur, sinh, sinusgrid, sphericaln, spiralwing, spirograph, split, squarize, squirrel, squish, sschecks, starblur, stripes, stwin, super_shape, tan, tancos, tangent, tanh, TwinTrian, twoface, unpolar, waves, wavesn, wedge_julia, whorl, xheart, zblur, zscale. --Code changes -Generalize Variation::PrecalcHelper() and rename to PrePostPrecalcHelper(). --Do the same for the OpenCL version and rename it PrePostPrecalcOpenCLString(). -Rename Variation::m_AssignType to m_PrePostAssignType since it's only relevant to pre/post variations.
2016-01-29 20:02:15 -05:00
helper.Out.z = DefaultZ(helper);
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
}
--Code changes -Better use of const functions. -Ember.pro should not define BUILDING_EMBERCL, instead define BUILDING_EMBER.
2015-06-29 23:13:53 -04:00
virtual string OpenCLString() const override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
ostringstream ss, ss2;
0.4.1.3 Beta 10/14/2014 --User Changes Size is no longer fixed to the window size. Size scaling is done differently in the final render dialog. This fixes several bugs. Remove Xml saving size from settings and options dialog, it no longer applies. Final render can be broken into strips. Set default save path to the desktop if none is found in the settings file. Set default output size to 1920x1080 if none is found in the settings file. --Bug Fixes Better memory size reporting in final render dialog. --Code Changes Migrate to C++11, Qt 5.3.1, and Visual Studio 2013. Change most instances of unsigned int to size_t, and int to intmax_t. Add m_OrigPixPerUnit and m_ScaleType to Ember for scaling purposes. Replace some sprintf_s() calls in XmlToEmber with ostringstream. Move more non-templated members into RendererBase. Add CopyVec() overload that takes a per element function pointer. Add vector Memset(). Replace '&' with '+' instead of "&amp;" in XmlToEmber for much faster parsing. Break strips rendering out into EmberCommon and call from EmberRender and Fractorium. Make AddAndWriteBuffer() just call WriteBuffer(). Make AddAndWriteImage() delete the existing image first before replacing it. Add SetOutputTexture() to RendererCL to support making new textures in response to resize events. Remove multiple return statements in RendererCL, and replace with a bool that tracks results. Add ToDouble(), MakeEnd(), ToString() and Exists() wrappers in Fractorium. Add Size() wrapper in EmberFile. Make QString function arguments const QString&, and string with const string&. Make ShowCritical() wrapper for invoking a message box from another thread. Add combo box to TwoButtonWidget and rename.
2014-10-14 11:53:15 -04:00
intmax_t i = 0, varIndex = IndexInXform();
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
ss2 << "_" << XformIndexInEmber() << "]";
string index = ss2.str();
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
string weight = WeightDefineString();
string x = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string y = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
string mp = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
string v = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
string vx = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
string vy = "parVars[" + ToUpper(m_Params[i++].Name()) + index;
ss << "\t{\n"
Numerous fixes 0.4.0.5 Beta 07/18/2014 --User Changes Allow for vibrancy values > 1. Add flatten and unflatten menu items. Automatically flatten like Apophysis does. Add plugin and new_linear tags to Xml to be compatible with Apophysis. --Bug Fixes Fix blur, blur3d, bubble, cropn, cross, curl, curl3d, epispiral, ho, julia3d, julia3dz, loonie, mirror_x, mirror_y, mirror_z, rotate_x, sinusoidal, spherical, spherical3d, stripes. Unique filename on final render was completely broken. Two severe OpenCL bugs. Random seeds were biased and fusing was being reset too often leading to results that differ from the CPU. Subtle, but sometimes severe bug in the setup of the xaos weights. Use properly defined epsilon by getting the value from std::numeric_limits, rather than hard coding 1e-6 or 1e-10. Omit incorrect usage of epsilon everywhere. It should not be automatically added to denominators. Rather, it should only be used if the denominator is zero. Force final render progress bars to 100 on completion. Sometimes they didn't seem to make it there. Make variation name and params comparisons be case insensitive. --Code Changes Make ForEach and FindIf wrappers around std::for_each and std::find_if.
2014-07-19 02:33:18 -04:00
<< "\t\treal_t t = Zeps((cos(vIn.x) + cos(vIn.y)) / " << mp << " + 1);\n"
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
<< "\t\treal_t r = " << weight << " / t;\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t\treal_t tmp = precalcSumSquares + 1;\n"
<< "\t\treal_t x2 = 2 * vIn.x;\n"
<< "\t\treal_t y2 = 2 * vIn.y;\n"
Ditch T(f) define in opencl programs in preference to already-defined real_t. Also removed some erroneously-added double-'T(' statements
2015-02-26 16:18:55 -05:00
<< "\t\treal_t xmax = (real_t)(0.5) * (sqrt(tmp + x2) + sqrt(tmp - x2));\n"
<< "\t\treal_t ymax = (real_t)(0.5) * (sqrt(tmp + y2) + sqrt(tmp - y2));\n"
Numerous fixes 0.4.0.5 Beta 07/18/2014 --User Changes Allow for vibrancy values > 1. Add flatten and unflatten menu items. Automatically flatten like Apophysis does. Add plugin and new_linear tags to Xml to be compatible with Apophysis. --Bug Fixes Fix blur, blur3d, bubble, cropn, cross, curl, curl3d, epispiral, ho, julia3d, julia3dz, loonie, mirror_x, mirror_y, mirror_z, rotate_x, sinusoidal, spherical, spherical3d, stripes. Unique filename on final render was completely broken. Two severe OpenCL bugs. Random seeds were biased and fusing was being reset too often leading to results that differ from the CPU. Subtle, but sometimes severe bug in the setup of the xaos weights. Use properly defined epsilon by getting the value from std::numeric_limits, rather than hard coding 1e-6 or 1e-10. Omit incorrect usage of epsilon everywhere. It should not be automatically added to denominators. Rather, it should only be used if the denominator is zero. Force final render progress bars to 100 on completion. Sometimes they didn't seem to make it there. Make variation name and params comparisons be case insensitive. --Code Changes Make ForEach and FindIf wrappers around std::for_each and std::find_if.
2014-07-19 02:33:18 -04:00
<< "\t\treal_t a = vIn.x / Zeps(xmax);\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t\treal_t b = SafeSqrt(1 - SQR(a));\n"
<< "\n"
<< "\t\tvOut.x = " << vx << " * atan2(a, b) * r;\n"
<< "\n"
Numerous fixes 0.4.0.5 Beta 07/18/2014 --User Changes Allow for vibrancy values > 1. Add flatten and unflatten menu items. Automatically flatten like Apophysis does. Add plugin and new_linear tags to Xml to be compatible with Apophysis. --Bug Fixes Fix blur, blur3d, bubble, cropn, cross, curl, curl3d, epispiral, ho, julia3d, julia3dz, loonie, mirror_x, mirror_y, mirror_z, rotate_x, sinusoidal, spherical, spherical3d, stripes. Unique filename on final render was completely broken. Two severe OpenCL bugs. Random seeds were biased and fusing was being reset too often leading to results that differ from the CPU. Subtle, but sometimes severe bug in the setup of the xaos weights. Use properly defined epsilon by getting the value from std::numeric_limits, rather than hard coding 1e-6 or 1e-10. Omit incorrect usage of epsilon everywhere. It should not be automatically added to denominators. Rather, it should only be used if the denominator is zero. Force final render progress bars to 100 on completion. Sometimes they didn't seem to make it there. Make variation name and params comparisons be case insensitive. --Code Changes Make ForEach and FindIf wrappers around std::for_each and std::find_if.
2014-07-19 02:33:18 -04:00
<< "\t\ta = vIn.y / Zeps(ymax);\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t\tb = SafeSqrt(1 - SQR(a));\n"
<< "\n"
<< "\t\tvOut.y = " << vy << " * atan2(a, b) * r;\n"
--User changes -Allow for pausing the renderer in the main window. This makes is more efficient when entering many parameters, such as when following a tutorial. -Add support for new variations: erf, gamma, jac_cn, jac_dn, jac_sn, logDB, pressure_wave, pRose3D, splits3D, w, waves2b, x, xerf, y, z. -Inform user of the start and stop of file parsing in EmberAnimate because the files could potentially be very large. -Move the follwing fields to a new table called Animation: Interpolation, Affine Interpolation, Temporal Samples, Temporal Filter Width, Temporal Filter Type. -These currently have no effect on the interactive renderer and instead are used when running flames through EmberGenome to generate sequences, and then animating them in Fractorium or EmberAnimate. -Add new parameter overrides for EmberRender and EmberAnimate which directly assign values to all flames being rendered, rather than scale: --quality --demin --demax --Bug fixes -Left pad instead of right pad names of sequence outputs from EmberGenome. -Unique file naming was broken for files which already had an underscore in them. -Properly report that png is the default format of EmberRender and EmberAnimate output instead of erroneously claiming it was jpg. -Make command line programs search these folders in this order for the palette file: ./ ~/.fractorium ~/.config/fractorium /usr/share/fractorium /usr/local/share/fractorium -Fix possible bad values in hexes. -Better assignment of Z variables. -Fix boarders due to use of poorly implemented rint() function from flam3. Use std::rint() now. -wedge_sph was completely wrong due to having accidentally swapped the mapping of two parameters. -Make juliascope juliascope_power parameter be of type REAL_NONZERO since it's used as a denominator. -Make Z assignment compatible with the originals in: -arch, bcircle, bCollide, bent, bent2, bisplit, blob, blur_linear, blur_square, bMod, boarders, boarders2, bSwirl, bTransform, butterfly, cardioid, cell, circleblur, circlize, circlize2, circus, collideoscope, cos, cosine, cosh, coth, cpow, cpow2, crescents, cropn, csc, csch, curl, curve, dc_gridout, deltaa, diamond, disc2, eclipse, eCollide, edisc, eJulia, elliptic, eMod, eMotion, ennepers, epispiral, ePush, eRotate, eScale, eSwirl, ex, exp, expo, exponential, fan, fdisc, fibonacci, fibonacci2, fisheye, flipcircle, flipy, flower, flux, funnel, glynnia, GlynnSim1, GlynnSim2, GlynnSim3, gridout, handkerchief, heart, hole, idisc, julia, julian2, juliaNab, kaleidoscope, lazyTravis, Lissajous, mask, MobiusN, mobius_strip, modulus, murl, murl2, npolar, ortho, oscilloscope, parabola, perspective, petal, phoenix_julia, pie (was also inconsistent between cpu and gpu), poincare, popcorn, popcorn2, power, pow_block, rational3, rays, rblur, rings, rippled, roundspher, sec, secant2, sigmoid, sin, sineblur, sinh, sinusgrid, sphericaln, spiralwing, spirograph, split, squarize, squirrel, squish, sschecks, starblur, stripes, stwin, super_shape, tan, tancos, tangent, tanh, TwinTrian, twoface, unpolar, waves, wavesn, wedge_julia, whorl, xheart, zblur, zscale. --Code changes -Generalize Variation::PrecalcHelper() and rename to PrePostPrecalcHelper(). --Do the same for the OpenCL version and rename it PrePostPrecalcOpenCLString(). -Rename Variation::m_AssignType to m_PrePostAssignType since it's only relevant to pre/post variations.
2016-01-29 20:02:15 -05:00
<< "\t\tvOut.z = " << DefaultZCl()
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
<< "\t}\n";
return ss.str();
}
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
--User changes -Add new variations: bubbleT3D, crob, hexaplay3D, hexcrop, hexes, hexnix3D, loonie2, loonie3, nBlur, octapol and synth. -Allow for pre/post versions of dc_bubble, dc_cylinder and dc_linear whereas before they were omitted. -When saving a file with multiple embers in it, detect if time values are all the same and if so, start them at zero and increment by 1 for each ember. -Allow for numerous quality increases to be coalesced into one. It will pick up at the end of the current render. -Show selection highlight on variations tree in response to mouse hover. This makes it easier to see for which variation or param the current mouse wheel action will apply. -Make default temporal samples be 100, whereas before it was 1000 which was overkill. -Require the shift key to be held with delete for deleting an ember to prevent it from triggering when the user enters delete in the edit box. -This wasn't otherwise fixable without writing a lot more code. --Bug fixes -EmberGenome was crashing when generating a sequence from a source file with more than 2 embers in it. -EmberGenome was improperly handling the first frame of a merge after the last frame of the loop. -These bugs were due to a previous commit. Revert parts of that commit. -Prevent a zoom value of less than 0 when reading from xml. -Slight optimization of the crescents, and mask variations, if the compiler wasn't doing it already. -Unique file naming was broken because it was looking for _# and the default names ended with -#. -Disallow renaming of an ember in the library tree to an empty string. -Severe bug that prevented some variations from being read correctly from params generated outside this program. -Severe OpenCL randomization bug. The first x coordinates of the first points in the first kernel call of the first ember of a render since the OpenCL renderer object was created were not random and were mostly -1. -Severe bug when populating xform selection distributions that could sometimes cause a crash due to roundoff error. Fix by using double. -Limit the max number of variations in a random ember to MAX_CL_VARS, which is 8. This ensures they'll look the same on CPU and GPU. -Prevent user from saving stylesheet to default.qss, it's a special reserved filename. --Code changes -Generalize using the running sum output point inside of a variation for all cases: pre, reg and post. -Allow for array variables in variations where the address of each element is stored in m_Params. -Qualify all math functions with std:: -No longer use our own Clamp() in OpenCL, instead use the standard clamp(). -Redesign how functions are used in the variations OpenCL code. -Add tests to EmberTester to verify some of the new functionality. -Place more const and override qualifiers on functions where appropriate. -Add a global rand with a lock to be used very sparingly. -Use a map instead of a vector for bad param names in Xml parsing. -Prefix affine interpolation mode defines with "AFFINE_" to make their purpose more clear. -Allow for variations that change state during iteration by sending a separate copy of the ember to each rendering thread. -Implement this same functionality with a local struct in OpenCL. It's members are the total of all variables that need to change state within an ember. -Add Contains() function to Utils.h. -EmberRender: print names of kernels being printed with --dump_kernel option. -Clean up EmberTester to handle some of the recent changes. -Fix various casts. -Replace % 2 with & 1, even though the compiler was likely doing this already. -Add new file Variations06.h to accommodate new variations. -General cleanup.
2015-11-22 17:15:07 -05:00
virtual vector<string> OpenCLGlobalFuncNames() const override
{
return vector<string> { "SafeSqrt", "Zeps" };
}
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
virtual void Precalc() override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
{
m_V = m_Weight / T(M_PI_2);
m_Vx = m_V * m_X;
m_Vy = m_V * m_Y;
}
protected:
void Init()
{
string prefix = Prefix();
m_Params.clear();
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
m_Params.push_back(ParamWithName<T>(&m_X, prefix + "blocky_x", 1));
m_Params.push_back(ParamWithName<T>(&m_Y, prefix + "blocky_y", 1));
--Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums.
2016-01-04 19:50:15 -05:00
m_Params.push_back(ParamWithName<T>(&m_Mp, prefix + "blocky_mp", 4, eParamType::REAL_NONZERO));
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
m_Params.push_back(ParamWithName<T>(true, &m_V, prefix + "blocky_v"));//Precalc.
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
m_Params.push_back(ParamWithName<T>(true, &m_Vx, prefix + "blocky_vx"));
m_Params.push_back(ParamWithName<T>(true, &m_Vy, prefix + "blocky_vy"));
}
private:
T m_X;
T m_Y;
T m_Mp;
T m_V;//Precalc.
T m_Vx;
T m_Vy;
};
--User changes: -Show common folder locations such as documents, downloads, pictures in the sidebar in all file dialogs. -Warning message about exceeding memory in final render dialog now suggests strips as the solution to the problem. -Strips now has a tooltip explaining what it does. -Allow more digits in the spinners on the color section the flame tab. -Add manually adjustable size spinners in the final render dialog. Percentage scale and absolute size are fully synced. -Default prefix in final render is now the filename when doing animations (coming from sequence section of the library tab). -Changed the elliptic variation back to using a less precise version for float, and a more precise version for double. The last release had it always using double. -New applied xaos table that shows a read-only view of actual weights by taking the base xform weights and multiplying them by the xaos values. -New table in the xaos tab that gives a graphical representation of the probability that each xform is chosen, with and without xaos. -Add button to transpose the xaos rows and columns. -Add support for importing .chaos files from Chaotica. --Pasting back to Chaotica will work for most, but not all, variations due to incompatible parameter names in some. -Curves are now splines instead of Bezier. This adds compatibility with Chaotica, but breaks it for Apophysis. Xmls are still pastable, but the color curves will look different. --The curve editor on the palette tab can now add points by clicking on the lines and remove points by clicking on the points themselves, just like Chaotica. --Splines are saved in four new xml fields: overall_curve, red_curve, green_curve and blue_curve. -Allow for specifying the percentage of a sub batch each thread should iterate through per kernel call when running with OpenCL. This gives a roughly 1% performance increase due to having to make less kernel calls while iterating. --This field is present for interactive editing (where it's not very useful) and in the final render dialog. --On the command line, this is specified as --sbpctth for EmberRender and EmberAnimate. -Allow double clicking to toggle the supersample field in the flame tab between 1 and 2 for easily checking the effect of the field. -When showing affine values as polar coordinates, show angles normalized to 360 to match Chaotica. -Fuse Count spinner now toggles between 15 and 100 when double clicking for easily checking the effect of the field. -Added field for limiting the range in the x and y direction that the initial points are chosen from. -Added a field called K2 which is an alternative way to set brightness, ignored when zero. --This has no effect for many variations, but hs a noticeable effect for some. -Added new variations: arcsech arcsech2 arcsinh arctanh asteria block bwraps_rand circlecrop2 coth_spiral crackle2 depth_blur depth_blur2 depth_gaussian depth_gaussian2 depth_ngon depth_ngon2 depth_sine depth_sine2 dragonfire dspherical dust excinis exp2 flipx flowerdb foci_p gaussian glynnia2 glynnsim4 glynnsim5 henon henon hex_rand hex_truchet hypershift lazyjess lens lozi lozi modulusx modulusy oscilloscope2 point_symmetry pointsymmetry projective pulse rotate scry2 shift smartshape spher squares starblur2 swirl3 swirl3r tanh_spiral target0 target2 tile_hlp truchet_glyph truchet_inv truchet_knot unicorngaloshen vibration vibration2 --hex_truchet, hex_rand should always use double. They are extremely sensitive. --Bug fixes: -Bounds sign was flipped for x coordinate of world space when center was not zero. -Right clicking and dragging spinner showed menu on mouse up, even if it was very far away. -Text boxes for size in final render dialog were hard to type in. Same bug as xform weight used to be so fix the same way. -Fix spelling to be plural in toggle color speed box. -Stop using the blank user palette to generate flames. Either put colored palettes in it, or exclude it from randoms. -Clicking the random palette button for a palette file with only one palette in it would freeze the program. -Clicking none scale in final render did not re-render the preview. -Use less precision on random xaos. No need for 12 decimal places. -The term sub batch is overloaded in the options dialog. Change the naming and tooltip of those settings for cpu and opencl. --Also made clear in the tooltip for the default opencl quality setting that the value is per device. -The arrows spinner in palette editor appears like a read-only label. Made it look like a spinner. -Fix border colors for various spin boxes and table headers in the style sheet. Requires reload. -Fix a bug in the bwraps variation which would produce different results than Chaotica and Apophysis. -Synth was allowed to be selected for random flame generation when using an Nvidia card but it shouldn't have been because Nvidia has a hard time compiling synth. -A casting bug in the OpenCL kernels for log scaling and density filtering was preventing successful compilations on Intel iGPUs. Fixed even though we don't support anything other than AMD and Nvidia. -Palette rotation (click and drag) position was not being reset when loading a new flame. -When the xform circles were hidden, opening and closing the options dialog would improperly reshow them. -Double click toggle was broken on integer spin boxes. -Fixed tab order of some controls. -Creating a palette from a jpg in the palette editor only produced a single color. --Needed to package imageformats/qjpeg.dll with the Windows installer. -The basic memory benchmark test flame was not really testing memory. Make it more spread out. -Remove the temporal samples field from the flame tab, it was never used because it's only an animation parameter which is specified in the final render dialog or on the command line with EmberAnimate. --Code changes: -Add IsEmpty() to Palette to determine if a palette is all black. -Attempt to avoid selecting a blank palette in PaletteList::GetRandomPalette(). -Add function ScanForChaosNodes() and some associated helper functions in XmlToEmber. -Make variation param name correction be case insensitive in XmlToEmber. -Report error when assigning a variation param value in XmlToEmber. -Add SubBatchPercentPerThread() method to RendererCL. -Override enterEvent() and leaveEvent() in DoubleSpinBox and SpinBox to prevent the context menu from showing up on right mouse up after already leaving the spinner. -Filtering the mouse wheel event in TableWidget no longer appears to be needed. It was probably an old Qt bug that has been fixed. -Gui/ember syncing code in the final render dialog needed to be reworked to accommodate absolute sizes.
2019-04-13 22:00:46 -04:00
/// <summary>
/// block.
/// By TyrantWave.
/// </summary>
template <typename T>
class BlockVariation : public ParametricVariation<T>
{
public:
BlockVariation(T weight = 1.0) : ParametricVariation<T>("block", eVariationId::VAR_BLOCK, weight, true)
{
Init();
}
PARVARCOPY(BlockVariation)
virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
{
T tmp = helper.m_PrecalcSumSquares + 1;
T x2 = 2 * helper.In.x;
T y2 = 2 * helper.In.y;
T xmax = T(0.5) * (std::sqrt(tmp + x2) + std::sqrt(tmp - x2));
T ymax = T(0.5) * (std::sqrt(tmp + y2) + std::sqrt(tmp - y2));
T a = helper.In.x / Zeps(xmax);
T b = VarFuncs<T>::SafeSqrt(1 - SQR(a));
helper.Out.x = m_WightDivPiOver2 * std::atan2(a, b);
a = helper.In.y / Zeps(ymax);
b = VarFuncs<T>::SafeSqrt(1 - SQR(a));
helper.Out.y = m_WightDivPiOver2 * std::atan2(a, b);
helper.Out.z = DefaultZ(helper);
}
virtual string OpenCLString() const override
{
ostringstream ss, ss2;
int i = 0;
ss2 << "_" << XformIndexInEmber() << "]";
string index = ss2.str();
string weight = WeightDefineString();
string wdpio2 = "parVars[" + ToUpper(m_Params[i++].Name()) + index;//Precalcs only, no params.
ss << "\t{\n"
<< "\t\treal_t tmp = precalcSumSquares + 1;\n"
<< "\t\treal_t x2 = 2 * vIn.x;\n"
<< "\t\treal_t y2 = 2 * vIn.y;\n"
<< "\t\treal_t xmax = 0.5 * (sqrt(tmp + x2) + sqrt(tmp - x2));\n"
<< "\t\treal_t ymax = 0.5 * (sqrt(tmp + y2) + sqrt(tmp - y2));\n"
<< "\t\treal_t a = vIn.x / Zeps(xmax);\n"
<< "\t\treal_t b = SafeSqrt(1 - SQR(a));\n"
<< "\t\tvOut.x = " << wdpio2 << " * atan2(a, b);\n"
<< "\t\ta = vIn.y / Zeps(ymax);\n"
<< "\t\tb = SafeSqrt(1 - SQR(a));\n"
<< "\t\tvOut.y = " << wdpio2 << " * atan2(a, b);\n"
<< "\t\tvOut.z = " << DefaultZCl()
<< "\t}\n";
return ss.str();
}
virtual vector<string> OpenCLGlobalFuncNames() const override
{
return vector<string> { "SafeSqrt", "Zeps" };
}
virtual void Precalc() override
{
m_WightDivPiOver2 = m_Weight * T(M_2_PI);
}
protected:
void Init()
{
string prefix = Prefix();
m_Params.clear();
m_Params.push_back(ParamWithName<T>(true, &m_WightDivPiOver2, prefix + "block_weightdivpiover2"));//Precalcs only, no params.
}
private:
T m_WightDivPiOver2;
};
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
MAKEPREPOSTPARVAR(ESwirl, eSwirl, ESWIRL)
--User changes: -Show common folder locations such as documents, downloads, pictures in the sidebar in all file dialogs. -Warning message about exceeding memory in final render dialog now suggests strips as the solution to the problem. -Strips now has a tooltip explaining what it does. -Allow more digits in the spinners on the color section the flame tab. -Add manually adjustable size spinners in the final render dialog. Percentage scale and absolute size are fully synced. -Default prefix in final render is now the filename when doing animations (coming from sequence section of the library tab). -Changed the elliptic variation back to using a less precise version for float, and a more precise version for double. The last release had it always using double. -New applied xaos table that shows a read-only view of actual weights by taking the base xform weights and multiplying them by the xaos values. -New table in the xaos tab that gives a graphical representation of the probability that each xform is chosen, with and without xaos. -Add button to transpose the xaos rows and columns. -Add support for importing .chaos files from Chaotica. --Pasting back to Chaotica will work for most, but not all, variations due to incompatible parameter names in some. -Curves are now splines instead of Bezier. This adds compatibility with Chaotica, but breaks it for Apophysis. Xmls are still pastable, but the color curves will look different. --The curve editor on the palette tab can now add points by clicking on the lines and remove points by clicking on the points themselves, just like Chaotica. --Splines are saved in four new xml fields: overall_curve, red_curve, green_curve and blue_curve. -Allow for specifying the percentage of a sub batch each thread should iterate through per kernel call when running with OpenCL. This gives a roughly 1% performance increase due to having to make less kernel calls while iterating. --This field is present for interactive editing (where it's not very useful) and in the final render dialog. --On the command line, this is specified as --sbpctth for EmberRender and EmberAnimate. -Allow double clicking to toggle the supersample field in the flame tab between 1 and 2 for easily checking the effect of the field. -When showing affine values as polar coordinates, show angles normalized to 360 to match Chaotica. -Fuse Count spinner now toggles between 15 and 100 when double clicking for easily checking the effect of the field. -Added field for limiting the range in the x and y direction that the initial points are chosen from. -Added a field called K2 which is an alternative way to set brightness, ignored when zero. --This has no effect for many variations, but hs a noticeable effect for some. -Added new variations: arcsech arcsech2 arcsinh arctanh asteria block bwraps_rand circlecrop2 coth_spiral crackle2 depth_blur depth_blur2 depth_gaussian depth_gaussian2 depth_ngon depth_ngon2 depth_sine depth_sine2 dragonfire dspherical dust excinis exp2 flipx flowerdb foci_p gaussian glynnia2 glynnsim4 glynnsim5 henon henon hex_rand hex_truchet hypershift lazyjess lens lozi lozi modulusx modulusy oscilloscope2 point_symmetry pointsymmetry projective pulse rotate scry2 shift smartshape spher squares starblur2 swirl3 swirl3r tanh_spiral target0 target2 tile_hlp truchet_glyph truchet_inv truchet_knot unicorngaloshen vibration vibration2 --hex_truchet, hex_rand should always use double. They are extremely sensitive. --Bug fixes: -Bounds sign was flipped for x coordinate of world space when center was not zero. -Right clicking and dragging spinner showed menu on mouse up, even if it was very far away. -Text boxes for size in final render dialog were hard to type in. Same bug as xform weight used to be so fix the same way. -Fix spelling to be plural in toggle color speed box. -Stop using the blank user palette to generate flames. Either put colored palettes in it, or exclude it from randoms. -Clicking the random palette button for a palette file with only one palette in it would freeze the program. -Clicking none scale in final render did not re-render the preview. -Use less precision on random xaos. No need for 12 decimal places. -The term sub batch is overloaded in the options dialog. Change the naming and tooltip of those settings for cpu and opencl. --Also made clear in the tooltip for the default opencl quality setting that the value is per device. -The arrows spinner in palette editor appears like a read-only label. Made it look like a spinner. -Fix border colors for various spin boxes and table headers in the style sheet. Requires reload. -Fix a bug in the bwraps variation which would produce different results than Chaotica and Apophysis. -Synth was allowed to be selected for random flame generation when using an Nvidia card but it shouldn't have been because Nvidia has a hard time compiling synth. -A casting bug in the OpenCL kernels for log scaling and density filtering was preventing successful compilations on Intel iGPUs. Fixed even though we don't support anything other than AMD and Nvidia. -Palette rotation (click and drag) position was not being reset when loading a new flame. -When the xform circles were hidden, opening and closing the options dialog would improperly reshow them. -Double click toggle was broken on integer spin boxes. -Fixed tab order of some controls. -Creating a palette from a jpg in the palette editor only produced a single color. --Needed to package imageformats/qjpeg.dll with the Windows installer. -The basic memory benchmark test flame was not really testing memory. Make it more spread out. -Remove the temporal samples field from the flame tab, it was never used because it's only an animation parameter which is specified in the final render dialog or on the command line with EmberAnimate. --Code changes: -Add IsEmpty() to Palette to determine if a palette is all black. -Attempt to avoid selecting a blank palette in PaletteList::GetRandomPalette(). -Add function ScanForChaosNodes() and some associated helper functions in XmlToEmber. -Make variation param name correction be case insensitive in XmlToEmber. -Report error when assigning a variation param value in XmlToEmber. -Add SubBatchPercentPerThread() method to RendererCL. -Override enterEvent() and leaveEvent() in DoubleSpinBox and SpinBox to prevent the context menu from showing up on right mouse up after already leaving the spinner. -Filtering the mouse wheel event in TableWidget no longer appears to be needed. It was probably an old Qt bug that has been fixed. -Gui/ember syncing code in the final render dialog needed to be reworked to accommodate absolute sizes.
2019-04-13 22:00:46 -04:00
MAKEPREPOSTPARVAR(LazyJess, lazyjess, LAZYJESS)
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
MAKEPREPOSTPARVAR(LazyTravis, lazyTravis, LAZY_TRAVIS)
MAKEPREPOSTPARVAR(Squish, squish, SQUISH)
MAKEPREPOSTPARVAR(Circus, circus, CIRCUS)
MAKEPREPOSTVAR(Tancos, tancos, TANCOS)
MAKEPREPOSTVAR(Rippled, rippled, RIPPLED)
MAKEPREPOSTPARVAR(RotateX, rotate_x, ROTATE_X)
MAKEPREPOSTPARVAR(RotateY, rotate_y, ROTATE_Y)
MAKEPREPOSTPARVAR(RotateZ, rotate_z, ROTATE_Z)
MAKEPREPOSTVAR(MirrorX, mirror_x, MIRROR_X)
MAKEPREPOSTVAR(MirrorY, mirror_y, MIRROR_Y)
MAKEPREPOSTVAR(MirrorZ, mirror_z, MIRROR_Z)
MAKEPREPOSTPARVAR(RBlur, rblur, RBLUR)
MAKEPREPOSTPARVAR(JuliaNab, juliaNab, JULIANAB)
MAKEPREPOSTPARVAR(Sintrange, sintrange, SINTRANGE)
MAKEPREPOSTPARVAR(Voron, Voron, VORON)
--Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums.
2016-01-04 19:50:15 -05:00
MAKEPREPOSTPARVARASSIGN(Waffle, waffle, WAFFLE, eVariationAssignType::ASSIGNTYPE_SUM)
MAKEPREPOSTVARASSIGN(Square3D, square3D, SQUARE3D, eVariationAssignType::ASSIGNTYPE_SUM)
MAKEPREPOSTPARVARASSIGN(SuperShape3D, SuperShape3D, SUPER_SHAPE3D, eVariationAssignType::ASSIGNTYPE_SUM)
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
MAKEPREPOSTPARVAR(Sphyp3D, sphyp3D, SPHYP3D)
MAKEPREPOSTPARVAR(Circlecrop, circlecrop, CIRCLECROP)
--User changes: -Show common folder locations such as documents, downloads, pictures in the sidebar in all file dialogs. -Warning message about exceeding memory in final render dialog now suggests strips as the solution to the problem. -Strips now has a tooltip explaining what it does. -Allow more digits in the spinners on the color section the flame tab. -Add manually adjustable size spinners in the final render dialog. Percentage scale and absolute size are fully synced. -Default prefix in final render is now the filename when doing animations (coming from sequence section of the library tab). -Changed the elliptic variation back to using a less precise version for float, and a more precise version for double. The last release had it always using double. -New applied xaos table that shows a read-only view of actual weights by taking the base xform weights and multiplying them by the xaos values. -New table in the xaos tab that gives a graphical representation of the probability that each xform is chosen, with and without xaos. -Add button to transpose the xaos rows and columns. -Add support for importing .chaos files from Chaotica. --Pasting back to Chaotica will work for most, but not all, variations due to incompatible parameter names in some. -Curves are now splines instead of Bezier. This adds compatibility with Chaotica, but breaks it for Apophysis. Xmls are still pastable, but the color curves will look different. --The curve editor on the palette tab can now add points by clicking on the lines and remove points by clicking on the points themselves, just like Chaotica. --Splines are saved in four new xml fields: overall_curve, red_curve, green_curve and blue_curve. -Allow for specifying the percentage of a sub batch each thread should iterate through per kernel call when running with OpenCL. This gives a roughly 1% performance increase due to having to make less kernel calls while iterating. --This field is present for interactive editing (where it's not very useful) and in the final render dialog. --On the command line, this is specified as --sbpctth for EmberRender and EmberAnimate. -Allow double clicking to toggle the supersample field in the flame tab between 1 and 2 for easily checking the effect of the field. -When showing affine values as polar coordinates, show angles normalized to 360 to match Chaotica. -Fuse Count spinner now toggles between 15 and 100 when double clicking for easily checking the effect of the field. -Added field for limiting the range in the x and y direction that the initial points are chosen from. -Added a field called K2 which is an alternative way to set brightness, ignored when zero. --This has no effect for many variations, but hs a noticeable effect for some. -Added new variations: arcsech arcsech2 arcsinh arctanh asteria block bwraps_rand circlecrop2 coth_spiral crackle2 depth_blur depth_blur2 depth_gaussian depth_gaussian2 depth_ngon depth_ngon2 depth_sine depth_sine2 dragonfire dspherical dust excinis exp2 flipx flowerdb foci_p gaussian glynnia2 glynnsim4 glynnsim5 henon henon hex_rand hex_truchet hypershift lazyjess lens lozi lozi modulusx modulusy oscilloscope2 point_symmetry pointsymmetry projective pulse rotate scry2 shift smartshape spher squares starblur2 swirl3 swirl3r tanh_spiral target0 target2 tile_hlp truchet_glyph truchet_inv truchet_knot unicorngaloshen vibration vibration2 --hex_truchet, hex_rand should always use double. They are extremely sensitive. --Bug fixes: -Bounds sign was flipped for x coordinate of world space when center was not zero. -Right clicking and dragging spinner showed menu on mouse up, even if it was very far away. -Text boxes for size in final render dialog were hard to type in. Same bug as xform weight used to be so fix the same way. -Fix spelling to be plural in toggle color speed box. -Stop using the blank user palette to generate flames. Either put colored palettes in it, or exclude it from randoms. -Clicking the random palette button for a palette file with only one palette in it would freeze the program. -Clicking none scale in final render did not re-render the preview. -Use less precision on random xaos. No need for 12 decimal places. -The term sub batch is overloaded in the options dialog. Change the naming and tooltip of those settings for cpu and opencl. --Also made clear in the tooltip for the default opencl quality setting that the value is per device. -The arrows spinner in palette editor appears like a read-only label. Made it look like a spinner. -Fix border colors for various spin boxes and table headers in the style sheet. Requires reload. -Fix a bug in the bwraps variation which would produce different results than Chaotica and Apophysis. -Synth was allowed to be selected for random flame generation when using an Nvidia card but it shouldn't have been because Nvidia has a hard time compiling synth. -A casting bug in the OpenCL kernels for log scaling and density filtering was preventing successful compilations on Intel iGPUs. Fixed even though we don't support anything other than AMD and Nvidia. -Palette rotation (click and drag) position was not being reset when loading a new flame. -When the xform circles were hidden, opening and closing the options dialog would improperly reshow them. -Double click toggle was broken on integer spin boxes. -Fixed tab order of some controls. -Creating a palette from a jpg in the palette editor only produced a single color. --Needed to package imageformats/qjpeg.dll with the Windows installer. -The basic memory benchmark test flame was not really testing memory. Make it more spread out. -Remove the temporal samples field from the flame tab, it was never used because it's only an animation parameter which is specified in the final render dialog or on the command line with EmberAnimate. --Code changes: -Add IsEmpty() to Palette to determine if a palette is all black. -Attempt to avoid selecting a blank palette in PaletteList::GetRandomPalette(). -Add function ScanForChaosNodes() and some associated helper functions in XmlToEmber. -Make variation param name correction be case insensitive in XmlToEmber. -Report error when assigning a variation param value in XmlToEmber. -Add SubBatchPercentPerThread() method to RendererCL. -Override enterEvent() and leaveEvent() in DoubleSpinBox and SpinBox to prevent the context menu from showing up on right mouse up after already leaving the spinner. -Filtering the mouse wheel event in TableWidget no longer appears to be needed. It was probably an old Qt bug that has been fixed. -Gui/ember syncing code in the final render dialog needed to be reworked to accommodate absolute sizes.
2019-04-13 22:00:46 -04:00
MAKEPREPOSTPARVAR(Circlecrop2, circlecrop2, CIRCLECROP2)
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
MAKEPREPOSTPARVAR(Julian3Dx, julian3Dx, JULIAN3DX)
MAKEPREPOSTPARVAR(Fourth, fourth, FOURTH)
MAKEPREPOSTPARVAR(Mobiq, mobiq, MOBIQ)
MAKEPREPOSTPARVAR(Spherivoid, spherivoid, SPHERIVOID)
MAKEPREPOSTPARVAR(Farblur, farblur, FARBLUR)
MAKEPREPOSTPARVAR(CurlSP, curl_sp, CURL_SP)
MAKEPREPOSTPARVAR(Heat, heat, HEAT)
MAKEPREPOSTPARVAR(Interference2, interference2, INTERFERENCE2)
MAKEPREPOSTVAR(Sinq, sinq, SINQ)
MAKEPREPOSTVAR(Sinhq, sinhq, SINHQ)
MAKEPREPOSTVAR(Secq, secq, SECQ)
MAKEPREPOSTVAR(Sechq, sechq, SECHQ)
MAKEPREPOSTVAR(Tanq, tanq, TANQ)
MAKEPREPOSTVAR(Tanhq, tanhq, TANHQ)
MAKEPREPOSTVAR(Cosq, cosq, COSQ)
MAKEPREPOSTVAR(Coshq, coshq, COSHQ)
MAKEPREPOSTVAR(Cotq, cotq, COTQ)
MAKEPREPOSTVAR(Cothq, cothq, COTHQ)
MAKEPREPOSTVAR(Cscq, cscq, CSCQ)
MAKEPREPOSTVAR(Cschq, cschq, CSCHQ)
MAKEPREPOSTVAR(Estiq, estiq, ESTIQ)
MAKEPREPOSTPARVAR(Loq, loq, LOQ)
MAKEPREPOSTVAR(Curvature, curvature, CURVATURE)
MAKEPREPOSTPARVAR(Qode, q_ode, Q_ODE)
--Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums.
2016-01-04 19:50:15 -05:00
MAKEPREPOSTPARVARASSIGN(BlurHeart, blur_heart, BLUR_HEART, eVariationAssignType::ASSIGNTYPE_SUM)
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
MAKEPREPOSTPARVAR(Truchet, Truchet, TRUCHET)
--User changes: -Show common folder locations such as documents, downloads, pictures in the sidebar in all file dialogs. -Warning message about exceeding memory in final render dialog now suggests strips as the solution to the problem. -Strips now has a tooltip explaining what it does. -Allow more digits in the spinners on the color section the flame tab. -Add manually adjustable size spinners in the final render dialog. Percentage scale and absolute size are fully synced. -Default prefix in final render is now the filename when doing animations (coming from sequence section of the library tab). -Changed the elliptic variation back to using a less precise version for float, and a more precise version for double. The last release had it always using double. -New applied xaos table that shows a read-only view of actual weights by taking the base xform weights and multiplying them by the xaos values. -New table in the xaos tab that gives a graphical representation of the probability that each xform is chosen, with and without xaos. -Add button to transpose the xaos rows and columns. -Add support for importing .chaos files from Chaotica. --Pasting back to Chaotica will work for most, but not all, variations due to incompatible parameter names in some. -Curves are now splines instead of Bezier. This adds compatibility with Chaotica, but breaks it for Apophysis. Xmls are still pastable, but the color curves will look different. --The curve editor on the palette tab can now add points by clicking on the lines and remove points by clicking on the points themselves, just like Chaotica. --Splines are saved in four new xml fields: overall_curve, red_curve, green_curve and blue_curve. -Allow for specifying the percentage of a sub batch each thread should iterate through per kernel call when running with OpenCL. This gives a roughly 1% performance increase due to having to make less kernel calls while iterating. --This field is present for interactive editing (where it's not very useful) and in the final render dialog. --On the command line, this is specified as --sbpctth for EmberRender and EmberAnimate. -Allow double clicking to toggle the supersample field in the flame tab between 1 and 2 for easily checking the effect of the field. -When showing affine values as polar coordinates, show angles normalized to 360 to match Chaotica. -Fuse Count spinner now toggles between 15 and 100 when double clicking for easily checking the effect of the field. -Added field for limiting the range in the x and y direction that the initial points are chosen from. -Added a field called K2 which is an alternative way to set brightness, ignored when zero. --This has no effect for many variations, but hs a noticeable effect for some. -Added new variations: arcsech arcsech2 arcsinh arctanh asteria block bwraps_rand circlecrop2 coth_spiral crackle2 depth_blur depth_blur2 depth_gaussian depth_gaussian2 depth_ngon depth_ngon2 depth_sine depth_sine2 dragonfire dspherical dust excinis exp2 flipx flowerdb foci_p gaussian glynnia2 glynnsim4 glynnsim5 henon henon hex_rand hex_truchet hypershift lazyjess lens lozi lozi modulusx modulusy oscilloscope2 point_symmetry pointsymmetry projective pulse rotate scry2 shift smartshape spher squares starblur2 swirl3 swirl3r tanh_spiral target0 target2 tile_hlp truchet_glyph truchet_inv truchet_knot unicorngaloshen vibration vibration2 --hex_truchet, hex_rand should always use double. They are extremely sensitive. --Bug fixes: -Bounds sign was flipped for x coordinate of world space when center was not zero. -Right clicking and dragging spinner showed menu on mouse up, even if it was very far away. -Text boxes for size in final render dialog were hard to type in. Same bug as xform weight used to be so fix the same way. -Fix spelling to be plural in toggle color speed box. -Stop using the blank user palette to generate flames. Either put colored palettes in it, or exclude it from randoms. -Clicking the random palette button for a palette file with only one palette in it would freeze the program. -Clicking none scale in final render did not re-render the preview. -Use less precision on random xaos. No need for 12 decimal places. -The term sub batch is overloaded in the options dialog. Change the naming and tooltip of those settings for cpu and opencl. --Also made clear in the tooltip for the default opencl quality setting that the value is per device. -The arrows spinner in palette editor appears like a read-only label. Made it look like a spinner. -Fix border colors for various spin boxes and table headers in the style sheet. Requires reload. -Fix a bug in the bwraps variation which would produce different results than Chaotica and Apophysis. -Synth was allowed to be selected for random flame generation when using an Nvidia card but it shouldn't have been because Nvidia has a hard time compiling synth. -A casting bug in the OpenCL kernels for log scaling and density filtering was preventing successful compilations on Intel iGPUs. Fixed even though we don't support anything other than AMD and Nvidia. -Palette rotation (click and drag) position was not being reset when loading a new flame. -When the xform circles were hidden, opening and closing the options dialog would improperly reshow them. -Double click toggle was broken on integer spin boxes. -Fixed tab order of some controls. -Creating a palette from a jpg in the palette editor only produced a single color. --Needed to package imageformats/qjpeg.dll with the Windows installer. -The basic memory benchmark test flame was not really testing memory. Make it more spread out. -Remove the temporal samples field from the flame tab, it was never used because it's only an animation parameter which is specified in the final render dialog or on the command line with EmberAnimate. --Code changes: -Add IsEmpty() to Palette to determine if a palette is all black. -Attempt to avoid selecting a blank palette in PaletteList::GetRandomPalette(). -Add function ScanForChaosNodes() and some associated helper functions in XmlToEmber. -Make variation param name correction be case insensitive in XmlToEmber. -Report error when assigning a variation param value in XmlToEmber. -Add SubBatchPercentPerThread() method to RendererCL. -Override enterEvent() and leaveEvent() in DoubleSpinBox and SpinBox to prevent the context menu from showing up on right mouse up after already leaving the spinner. -Filtering the mouse wheel event in TableWidget no longer appears to be needed. It was probably an old Qt bug that has been fixed. -Gui/ember syncing code in the final render dialog needed to be reworked to accommodate absolute sizes.
2019-04-13 22:00:46 -04:00
MAKEPREPOSTVAR(TruchetKnot, truchet_knot, TRUCHET_KNOT)
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
MAKEPREPOSTPARVAR(Gdoffs, gdoffs, GDOFFS)
MAKEPREPOSTPARVAR(Octagon, octagon, OCTAGON)
MAKEPREPOSTPARVAR(Trade, trade, TRADE)
MAKEPREPOSTPARVAR(Juliac, Juliac, JULIAC)
MAKEPREPOSTVAR(Blade3D, blade3D, BLADE3D)
MAKEPREPOSTPARVAR(Blob3D, blob3D, BLOB3D)
MAKEPREPOSTPARVAR(Blocky, blocky, BLOCKY)
--User changes: -Show common folder locations such as documents, downloads, pictures in the sidebar in all file dialogs. -Warning message about exceeding memory in final render dialog now suggests strips as the solution to the problem. -Strips now has a tooltip explaining what it does. -Allow more digits in the spinners on the color section the flame tab. -Add manually adjustable size spinners in the final render dialog. Percentage scale and absolute size are fully synced. -Default prefix in final render is now the filename when doing animations (coming from sequence section of the library tab). -Changed the elliptic variation back to using a less precise version for float, and a more precise version for double. The last release had it always using double. -New applied xaos table that shows a read-only view of actual weights by taking the base xform weights and multiplying them by the xaos values. -New table in the xaos tab that gives a graphical representation of the probability that each xform is chosen, with and without xaos. -Add button to transpose the xaos rows and columns. -Add support for importing .chaos files from Chaotica. --Pasting back to Chaotica will work for most, but not all, variations due to incompatible parameter names in some. -Curves are now splines instead of Bezier. This adds compatibility with Chaotica, but breaks it for Apophysis. Xmls are still pastable, but the color curves will look different. --The curve editor on the palette tab can now add points by clicking on the lines and remove points by clicking on the points themselves, just like Chaotica. --Splines are saved in four new xml fields: overall_curve, red_curve, green_curve and blue_curve. -Allow for specifying the percentage of a sub batch each thread should iterate through per kernel call when running with OpenCL. This gives a roughly 1% performance increase due to having to make less kernel calls while iterating. --This field is present for interactive editing (where it's not very useful) and in the final render dialog. --On the command line, this is specified as --sbpctth for EmberRender and EmberAnimate. -Allow double clicking to toggle the supersample field in the flame tab between 1 and 2 for easily checking the effect of the field. -When showing affine values as polar coordinates, show angles normalized to 360 to match Chaotica. -Fuse Count spinner now toggles between 15 and 100 when double clicking for easily checking the effect of the field. -Added field for limiting the range in the x and y direction that the initial points are chosen from. -Added a field called K2 which is an alternative way to set brightness, ignored when zero. --This has no effect for many variations, but hs a noticeable effect for some. -Added new variations: arcsech arcsech2 arcsinh arctanh asteria block bwraps_rand circlecrop2 coth_spiral crackle2 depth_blur depth_blur2 depth_gaussian depth_gaussian2 depth_ngon depth_ngon2 depth_sine depth_sine2 dragonfire dspherical dust excinis exp2 flipx flowerdb foci_p gaussian glynnia2 glynnsim4 glynnsim5 henon henon hex_rand hex_truchet hypershift lazyjess lens lozi lozi modulusx modulusy oscilloscope2 point_symmetry pointsymmetry projective pulse rotate scry2 shift smartshape spher squares starblur2 swirl3 swirl3r tanh_spiral target0 target2 tile_hlp truchet_glyph truchet_inv truchet_knot unicorngaloshen vibration vibration2 --hex_truchet, hex_rand should always use double. They are extremely sensitive. --Bug fixes: -Bounds sign was flipped for x coordinate of world space when center was not zero. -Right clicking and dragging spinner showed menu on mouse up, even if it was very far away. -Text boxes for size in final render dialog were hard to type in. Same bug as xform weight used to be so fix the same way. -Fix spelling to be plural in toggle color speed box. -Stop using the blank user palette to generate flames. Either put colored palettes in it, or exclude it from randoms. -Clicking the random palette button for a palette file with only one palette in it would freeze the program. -Clicking none scale in final render did not re-render the preview. -Use less precision on random xaos. No need for 12 decimal places. -The term sub batch is overloaded in the options dialog. Change the naming and tooltip of those settings for cpu and opencl. --Also made clear in the tooltip for the default opencl quality setting that the value is per device. -The arrows spinner in palette editor appears like a read-only label. Made it look like a spinner. -Fix border colors for various spin boxes and table headers in the style sheet. Requires reload. -Fix a bug in the bwraps variation which would produce different results than Chaotica and Apophysis. -Synth was allowed to be selected for random flame generation when using an Nvidia card but it shouldn't have been because Nvidia has a hard time compiling synth. -A casting bug in the OpenCL kernels for log scaling and density filtering was preventing successful compilations on Intel iGPUs. Fixed even though we don't support anything other than AMD and Nvidia. -Palette rotation (click and drag) position was not being reset when loading a new flame. -When the xform circles were hidden, opening and closing the options dialog would improperly reshow them. -Double click toggle was broken on integer spin boxes. -Fixed tab order of some controls. -Creating a palette from a jpg in the palette editor only produced a single color. --Needed to package imageformats/qjpeg.dll with the Windows installer. -The basic memory benchmark test flame was not really testing memory. Make it more spread out. -Remove the temporal samples field from the flame tab, it was never used because it's only an animation parameter which is specified in the final render dialog or on the command line with EmberAnimate. --Code changes: -Add IsEmpty() to Palette to determine if a palette is all black. -Attempt to avoid selecting a blank palette in PaletteList::GetRandomPalette(). -Add function ScanForChaosNodes() and some associated helper functions in XmlToEmber. -Make variation param name correction be case insensitive in XmlToEmber. -Report error when assigning a variation param value in XmlToEmber. -Add SubBatchPercentPerThread() method to RendererCL. -Override enterEvent() and leaveEvent() in DoubleSpinBox and SpinBox to prevent the context menu from showing up on right mouse up after already leaving the spinner. -Filtering the mouse wheel event in TableWidget no longer appears to be needed. It was probably an old Qt bug that has been fixed. -Gui/ember syncing code in the final render dialog needed to be reworked to accommodate absolute sizes.
2019-04-13 22:00:46 -04:00
MAKEPREPOSTPARVAR(Block, block, BLOCK)
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
///// <summary>
///// LinearXZ.
///// </summary>
//template <typename T>
--Code changes -Split VariationList into .h/cpp files. --Make VariationList.cpp the only place where the variations files are included from. This speeds up the build, but variations can no longer be directly instantiated. --Variations are no longer exported.
2016-02-18 21:58:24 -05:00
//class LinearXZVariation : public Variation<T>
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
//{
//public:
--Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums.
2016-01-04 19:50:15 -05:00
// LinearXZVariation(T weight = 1.0) : Variation<T>("linearxz", eVariationId::VAR_LINEAR_XZ, weight) { }
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
//
// VARCOPY(LinearXZVariation)
//
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
// virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
// {
// helper.Out.x = m_Weight * helper.In.x;
// helper.Out.z = m_Weight * helper.In.z;
// }
//
--Code changes -Better use of const functions. -Ember.pro should not define BUILDING_EMBERCL, instead define BUILDING_EMBER.
2015-06-29 23:13:53 -04:00
// virtual string OpenCLString() const override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
// {
// ostringstream ss;
0.4.1.3 Beta 10/14/2014 --User Changes Size is no longer fixed to the window size. Size scaling is done differently in the final render dialog. This fixes several bugs. Remove Xml saving size from settings and options dialog, it no longer applies. Final render can be broken into strips. Set default save path to the desktop if none is found in the settings file. Set default output size to 1920x1080 if none is found in the settings file. --Bug Fixes Better memory size reporting in final render dialog. --Code Changes Migrate to C++11, Qt 5.3.1, and Visual Studio 2013. Change most instances of unsigned int to size_t, and int to intmax_t. Add m_OrigPixPerUnit and m_ScaleType to Ember for scaling purposes. Replace some sprintf_s() calls in XmlToEmber with ostringstream. Move more non-templated members into RendererBase. Add CopyVec() overload that takes a per element function pointer. Add vector Memset(). Replace '&' with '+' instead of "&amp;" in XmlToEmber for much faster parsing. Break strips rendering out into EmberCommon and call from EmberRender and Fractorium. Make AddAndWriteBuffer() just call WriteBuffer(). Make AddAndWriteImage() delete the existing image first before replacing it. Add SetOutputTexture() to RendererCL to support making new textures in response to resize events. Remove multiple return statements in RendererCL, and replace with a bool that tracks results. Add ToDouble(), MakeEnd(), ToString() and Exists() wrappers in Fractorium. Add Size() wrapper in EmberFile. Make QString function arguments const QString&, and string with const string&. Make ShowCritical() wrapper for invoking a message box from another thread. Add combo box to TwoButtonWidget and rename.
2014-10-14 11:53:15 -04:00
// intmax_t varIndex = IndexInXform();
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
// string weight = WeightDefineString();
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
//
// ss << "\t{\n"
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
// << "\t\tvOut.x = " << weight << " * vIn.x;\n"
// << "\t\tvOut.z = " << weight << " * vIn.z;\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
// << "\t}\n";
//
// return ss.str();
// }
//};
//
///// <summary>
///// LinearYZ.
///// </summary>
//template <typename T>
--Code changes -Split VariationList into .h/cpp files. --Make VariationList.cpp the only place where the variations files are included from. This speeds up the build, but variations can no longer be directly instantiated. --Variations are no longer exported.
2016-02-18 21:58:24 -05:00
//class LinearYZVariation : public Variation<T>
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
//{
//public:
--Code changes -Convert all enums to class enum to be consistent with C++11 style. -Convert some if/else statements in filter classes to case statements. -Add overloaded stream operators to print various enums.
2016-01-04 19:50:15 -05:00
// LinearYZVariation(T weight = 1.0) : Variation<T>("linearyz", eVariationId::VAR_LINEAR_YZ, weight) { }
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
//
// VARCOPY(LinearYZVariation)
//
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
// virtual void Func(IteratorHelper<T>& helper, Point<T>& outPoint, QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand) override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
// {
// helper.Out.y = m_Weight * helper.In.y;
// helper.Out.z = m_Weight * helper.In.z;
// }
//
--Code changes -Better use of const functions. -Ember.pro should not define BUILDING_EMBERCL, instead define BUILDING_EMBER.
2015-06-29 23:13:53 -04:00
// virtual string OpenCLString() const override
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
// {
// ostringstream ss;
0.4.1.3 Beta 10/14/2014 --User Changes Size is no longer fixed to the window size. Size scaling is done differently in the final render dialog. This fixes several bugs. Remove Xml saving size from settings and options dialog, it no longer applies. Final render can be broken into strips. Set default save path to the desktop if none is found in the settings file. Set default output size to 1920x1080 if none is found in the settings file. --Bug Fixes Better memory size reporting in final render dialog. --Code Changes Migrate to C++11, Qt 5.3.1, and Visual Studio 2013. Change most instances of unsigned int to size_t, and int to intmax_t. Add m_OrigPixPerUnit and m_ScaleType to Ember for scaling purposes. Replace some sprintf_s() calls in XmlToEmber with ostringstream. Move more non-templated members into RendererBase. Add CopyVec() overload that takes a per element function pointer. Add vector Memset(). Replace '&' with '+' instead of "&amp;" in XmlToEmber for much faster parsing. Break strips rendering out into EmberCommon and call from EmberRender and Fractorium. Make AddAndWriteBuffer() just call WriteBuffer(). Make AddAndWriteImage() delete the existing image first before replacing it. Add SetOutputTexture() to RendererCL to support making new textures in response to resize events. Remove multiple return statements in RendererCL, and replace with a bool that tracks results. Add ToDouble(), MakeEnd(), ToString() and Exists() wrappers in Fractorium. Add Size() wrapper in EmberFile. Make QString function arguments const QString&, and string with const string&. Make ShowCritical() wrapper for invoking a message box from another thread. Add combo box to TwoButtonWidget and rename.
2014-10-14 11:53:15 -04:00
// intmax_t varIndex = IndexInXform();
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
// string weight = WeightDefineString();
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
//
// ss << "\t{\n"
--User changes -Support 4k monitors, and in general, properly scale any monitor that is not HD. -Allow for a spatial filter of radius zero, which means do not use a spatial filter. -Add new variations: concentric, cpow3, helicoid, helix, rand_cubes, sphereblur. -Use a new method for computing elliptic which is more precise. Developed by Discord user Claude. -Remove the 8 variation per xform limitation on the GPU. -Allow for loading the last flame file on startup, rather than randoms. -Use two different default quality values in the interactive renderer, one each for CPU and GPU. -Creating linked xforms was using non-standard behavior. Make it match Apo and also support creating multiple linked xforms at once. --Bug fixes -No variations in an xform used to have the same behavior as a single linear variation with weight 1. While sensible, this breaks backward compatibility. No variations now sets the output point to zeroes. -Prevent crashing the program when adjusting a value on the main window while a final render is in progress. -The xaos table was inverted. --Code changes -Convert projects to Visual Studio 2017. -Change bad vals from +- 1e10 to +-1e20. -Reintroduce the symmetry tag in xforms for legacy support in programs that do not use color_speed. -Compiler will not let us use default values in templated member functions anymore.
2017-11-26 20:27:00 -05:00
// << "\t\tvOut.y = " << weight << " * vIn.y;\n"
// << "\t\tvOut.z = " << weight << " * vIn.z;\n"
Initial source commit Initial source commit
2014-07-08 03:11:14 -04:00
// << "\t}\n";
//
// return ss.str();
// }
//};
Initial VS2013, C++11 and linux commit. This most likely won't build and suffers from some compiler issues. More commits to follow.
2014-09-01 00:25:15 -04:00
}
Reference in New Issue Copy Permalink