f5ef30bc1b
Clean up some filter code
2012-04-16 00:39:22 -07:00
f3c2186137
Small fixups
2012-04-14 23:44:10 -07:00
44869cc9ea
Remove stray debugging statements
2012-04-14 23:42:38 -07:00
e7cb39b456
Fetch defaults when packing interped params
2012-04-14 22:57:11 -07:00
627bcf8921
Add blending with lookup. In theory.
2012-04-14 22:56:29 -07:00
a4178c60fb
Update the genome specs a bit
2012-04-14 22:55:00 -07:00
ee2d571e9d
Rename spec.py to specs.py (prevents clashes)
2012-04-14 22:48:10 -07:00
b53f703e6e
Checkpoint! Renders again. Many fixes outstanding.
2012-04-10 08:44:25 -07:00
9aa1a94aa1
Experimental HaloClip filter.
2012-04-06 21:33:17 -07:00
c57917abe6
Use a unified block and grid addressing scheme.
2012-04-06 21:24:25 -07:00
eca8a8e1d3
Factor texrefs out of bilateral filter.
...
This also corrects the fact that denblurlib wouldn't compile without
bilaterllib.
2012-04-06 21:17:31 -07:00
b1b09c4bde
Fix filter width/offset in density blur.
...
Fun fact: static initializers of the wrong length are filled with zeros
or whatever. Thanks, C! (This fixes some of the anisotropy which you can
kinda see all over the place with the bugged filtering. Very subtle but
still important to fix.)
2012-04-06 21:08:44 -07:00
5d3b290c43
Fix typo
2012-03-18 17:29:34 -07:00
460387ef91
Use alternate behavior in presence of NaNs
2012-03-18 17:29:02 -07:00
ea43e312e4
Avoid overflow in exponential var.
2012-03-18 14:58:25 -07:00
3ee4640001
Hack: add magnitude scaling to affine coeffecients
2012-03-16 20:51:50 -07:00
f5356b571b
Fix handling of optional values in XML import
2012-03-16 20:51:50 -07:00
e726511c5a
Fix print_interp_knots debugging helper
2012-03-16 20:51:50 -07:00
92b1a27e8f
Fix improper seed re-storage
2012-02-16 23:38:35 -05:00
66e02db155
Fix potential palette glitch on one-palette anims
2012-02-16 23:38:17 -05:00
5a91d9f96c
Hang on to old modules to avoid syncing
2012-02-15 10:07:47 -05:00
88abefa4f4
Fix rb_incr() when blockDim.y == 1.
2012-02-15 10:06:35 -05:00
b6dfd2d980
Fix (same) logic error in RenderManager.render()
2012-02-14 10:30:34 -08:00
60a45c9a20
Sweeping refactor. More bugs undoubtedly remain.
2012-02-14 07:40:58 -08:00
1f2f61fd3a
Fix SplEval.insert_knot; add .update
2012-01-31 12:13:45 -05:00
c01d52ba24
Fix scaling spatial std on output width
2012-01-31 12:10:15 -05:00
6fba14e2f7
Okay, now I'm satisfied.
2012-01-29 18:49:19 -05:00
387dfd9f8c
More experiments
2012-01-28 13:20:48 -05:00
7c37b2b688
Use box filtering
2012-01-28 13:20:20 -05:00
899b8082da
Expose options on filter
2012-01-26 05:34:43 -05:00
a4b482f61a
Fix reversed angle factor. Ugh.
2012-01-25 12:25:56 -05:00
b4132c7cd9
Absurdly complicated enhancements to filtering.
2012-01-22 23:57:03 -05:00
c572f62d7d
Use YUV during accumulation
2012-01-22 23:56:16 -05:00
c054c757bd
Limit the maximum number of separate xf buffers
2012-01-22 23:52:09 -05:00
45b75d3fa5
Experimental bilateral filtering.
2012-01-21 00:06:15 -05:00
a803216551
Move argset to code.util
2012-01-21 00:03:28 -05:00
acbde65b9f
Don't call set_format after set_address_2d
2012-01-20 11:22:27 -05:00
1398706886
Remove SS from DE, and improve performance.
2012-01-20 11:17:07 -05:00
e1914a9c87
Clamp min to 0 in colorclip (for wavelet et al)
2012-01-20 09:23:46 -05:00
1ff0deb16b
Fix 'waves' variation
2012-01-15 20:25:37 -05:00
a97b53d4bd
Allow use of 'prex' in precalc statements
2012-01-15 20:25:12 -05:00
933e5e50b3
Support arc, offset when importing XML loops
2012-01-15 12:49:26 -05:00
830c788263
Use newer-format 'link' property
2012-01-15 12:47:56 -05:00
f3bf16f18a
Reduce the minimum blur radius
2012-01-10 10:36:26 -05:00
7825b006b4
Stamp out pre-echo (hopefully)
2012-01-09 23:39:07 -05:00
8c29212821
Experimental supersampling and DE changes
2012-01-09 21:15:05 -05:00
11c729d370
Apply edge clamping a little more forcefully
2012-01-09 14:08:32 -05:00
7f745d74cd
Use np.unwrap (same idea as monotonify())
2012-01-09 10:17:53 -05:00
728df3ace6
Add test routine for splines
2012-01-09 10:17:30 -05:00
808bd66138
Add SplEval.insert_knot()
2011-12-28 15:39:17 -05:00
