mirror-github-bspeice/cuburn
1
0
Fork 0
mirror of https://github.com/stevenrobertson/cuburn.git synced 2025-02-05 11:40:04 -05:00
Code Issues Actions Packages Projects Releases Wiki Activity

Fix clamping issues

Browse Source
This commit is contained in:
Steven Robertson 2014-12-25 06:31:09 -08:00
parent b073b7e22e
commit 0988f8d5ed
1 changed files with 22 additions and 10 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

32
cuburn/code/output.py
View File

@ -1,7 +1,19 @@
from util import devlib, ringbuflib
from mwc import mwclib
rgba8lib = devlib(deps=[ringbuflib, mwclib], defs=r'''
ditherlib = devlib(deps=[mwclib], defs=r'''
// Clamp an input between 0 and a given peak (inclusive), dithering its output,
// with full clamping for pixels that are true-black for compressibility.
__device__ float dclampf(mwc_st &rctx, float peak, float in) {
float ret = 0.0f;
if (in > 0.0f) {
ret = fminf(peak, fmaxf(0.0f, in * peak + 0.49f * mwc_next_11(rctx)));
}
return ret;
}
''')
rgba8lib = devlib(deps=[ringbuflib, mwclib, ditherlib], defs=r'''
// Perform a conversion from float32 values to uint8 ones, applying
// pixel- and channel-independent dithering to reduce suprathreshold banding
// artifacts. Clamps values larger than 1.0f.
@ -22,10 +34,10 @@ __global__ void f32_to_rgba_u8(
float4 in = src[isrc];
uchar4 out = make_uchar4(
fminf(1.0f, in.x) * 255.0f + 0.49f * mwc_next_11(rctx),
fminf(1.0f, in.y) * 255.0f + 0.49f * mwc_next_11(rctx),
fminf(1.0f, in.z) * 255.0f + 0.49f * mwc_next_11(rctx),
fminf(1.0f, in.w) * 255.0f + 0.49f * mwc_next_11(rctx)
dclampf(rctx, 255.0f, in.x),
dclampf(rctx, 255.0f, in.y),
dclampf(rctx, 255.0f, in.z),
dclampf(rctx, 255.0f, in.w)
);
int idst = dstride * y + x;
@ -34,7 +46,7 @@ __global__ void f32_to_rgba_u8(
}
''')
rgba16lib = devlib(deps=[ringbuflib, mwclib], defs=r'''
rgba16lib = devlib(deps=[ringbuflib, mwclib, ditherlib], defs=r'''
// Perform a conversion from float32 values to uint16 ones, as above.
__global__ void f32_to_rgba_u16(
ushort4 *dst, const float4 *src,
@ -51,10 +63,10 @@ __global__ void f32_to_rgba_u16(
float4 in = src[isrc];
ushort4 out = make_ushort4(
fminf(1.0f, in.x) * 65535.0f + 0.49f * mwc_next_11(rctx),
fminf(1.0f, in.y) * 65535.0f + 0.49f * mwc_next_11(rctx),
fminf(1.0f, in.z) * 65535.0f + 0.49f * mwc_next_11(rctx),
fminf(1.0f, in.w) * 65535.0f + 0.49f * mwc_next_11(rctx)
dclampf(rctx, 65535.0f, in.x),
dclampf(rctx, 65535.0f, in.y),
dclampf(rctx, 65535.0f, in.z),
dclampf(rctx, 65535.0f, in.w)
);
int idst = dstride * y + x;