Bill-Spitzak
  • Joined on 2024-02-16
Bill-Spitzak pushed to main at Bill-Spitzak/blender 2024-07-24 21:27:18 +02:00
cb1c325d45 Fix #124133: Compute actual transform between domains outside glsl
b86eddf6ee Fix #124133: Put the .5 shift into RealizeOnDomainOperation
3f6a6db79b Fix #124133: Transforms shift due to incorrect floor()
ab51d879c3 Fix #124339: Compositor artifacts with 0.5 translation
6041bdd93a Fix #125340: Div by zero in LineArt shadow stage
Compare 80 commits »
Bill-Spitzak commented on pull request blender/blender#125342 2024-07-24 18:49:47 +02:00
Fix #124339: Add a slight offset to Nearest sampling to make 1/2 indexes less likely

I think your version is moving by epsilon in the output space, rather than in the input space. Not sure if this makes much difference though. Bit concerned that mirroring will cause the shift to…

Bill-Spitzak created pull request blender/blender#125342 2024-07-23 23:35:09 +02:00
Fix #124339: Add a slight offset to Nearest sampling to make 1/2 indexes less likely
Bill-Spitzak pushed to round_to_even at Bill-Spitzak/blender 2024-07-23 23:34:24 +02:00
c82c94a753 Fix #124339: Add a slight offset to Nearest sampling to make 1/2 indexes less likely
73e519ded2 UI: When Docking Use Movement Threshold for Dupli & Swap
12f3aec3eb BLO readfile: report error when reading invalid 'data' chunks.
5cbc540a64 Fix #125219: Transmissive BSDF different in Blended and Dithered mode
6e152df547 Fix #125319: Can't drop images in uv editor
Compare 10 commits »
Bill-Spitzak created branch round_to_even in Bill-Spitzak/blender 2024-07-23 23:34:23 +02:00
Bill-Spitzak commented on issue blender/blender#125339 2024-07-23 23:05:20 +02:00
GPU Compositor Nearest sampling rounds incorrectly

I have a patch that "fixes" it, but it is a hack. Proper sampling code should be using texelFetch(), not texture(), so it has total control over the indices.

Bill-Spitzak opened issue blender/blender#125339 2024-07-23 23:02:58 +02:00
GPU Compositor Nearest sampling rounds incorrectly
Bill-Spitzak commented on pull request blender/blender#125267 2024-07-23 20:47:23 +02:00
Fix #124842: Made CPU transforms match the GPU ones

Push is a trivial code change to the nearest sampling that does not change behavior. I did not want xy to have different values in the different case statements.

Bill-Spitzak pushed to resolveshift at Bill-Spitzak/blender 2024-07-23 20:43:20 +02:00
ff54811f38 Fix #124842: Remove unused wrap_pixel(int,int...) function
6c6f77e531 Fix #124842: Fix Clipping in MemoryBuffer::read_elem_bilinear
06c10b7dde Fix #124842: Fix clipping and Nearest sampling in COM_MemoryBuffer::read/read_elem_sampled
8e97855d7b Fix #124842: Made CPU ScaleOperation match the GPU
5aa9007350 Fix #124842: Fix COM_ScaleOperation to produce correct bounding box
Compare 40 commits »
Bill-Spitzak commented on pull request blender/blender#124381 2024-07-23 03:37:27 +02:00
Fix #124133: Transforms shift due to incorrect floor()

Patch updated to split into more steps with explanation.

I do want to keep the fixes to the matrix calculation. It is possible to patch up the old calculation but it seems greately preferrable…

Bill-Spitzak pushed to main at Bill-Spitzak/blender 2024-07-23 03:35:12 +02:00
b89e504a23 Fix #124133: Simplify matrix calculation
e9b74f35e3 Fix #124133: Compute actual transform between domains outside glsl
358b5d4490 Fix #124133: Put the .5 shift into RealizeOnDomainOperation
Compare 3 commits »
Bill-Spitzak pushed to main at Bill-Spitzak/blender 2024-07-23 03:22:26 +02:00
075efce2b3 Fix #124133: Simplify matrix calculation
8f2828ce9b Fix #124133: Compute actual transform between domains outside glsl
e06e6b720b Fix #124133: Put the .5 shift into RealizeOnDomainOperation
99da48a582 Fix #124133: Transforms shift due to incorrect floor()
0fe0c04d2a Fix #125074: MEM_new/MEM_freeN mismatch in selection pick code
Compare 61 commits »
Bill-Spitzak created pull request blender/blender#125267 2024-07-23 02:08:18 +02:00
Fix #124842: Made CPU transforms match the GPU ones
Bill-Spitzak closed pull request blender/blender#125149 2024-07-23 02:04:40 +02:00
Fix #124842: Fix CPU compositor sampling errors and make it match GPU
Bill-Spitzak commented on pull request blender/blender#125149 2024-07-23 02:04:38 +02:00
Fix #124842: Fix CPU compositor sampling errors and make it match GPU

I tried to update this patch but it does not seem to have worked correctly. Closing this and making a new pull request.

Bill-Spitzak pushed to resolveshift at Bill-Spitzak/blender 2024-07-23 01:59:08 +02:00
a0e5840e47 Fix #124842: Remove unused wrap_pixel(int,int...) function
3eae0eeb97 Fix #124842: Fix Clipping in MemoryBuffer::read_elem_bilinear
8ff8504899 Fix #124842: Fix clipping and Nearest sampling in COM_MemoryBuffer::read/read_elem_sampled
724ebb1ab4 Fix #124842: Made CPU ScaleOperation match the GPU
eac0d8dced Fix #124842: Fix COM_ScaleOperation to produce correct bounding box
Compare 29 commits »
Bill-Spitzak commented on pull request blender/blender#125149 2024-07-23 00:33:27 +02:00
Fix #124842: Fix CPU compositor sampling errors and make it match GPU

This is a mistake, they only fixed the Scale. It appears the errors in Rotate are much larger but I also think it is less vital to fix them.

Bill-Spitzak commented on pull request blender/blender#125149 2024-07-23 00:32:31 +02:00
Fix #124842: Fix CPU compositor sampling errors and make it match GPU

The translation is wrong without these .5 offsets and did not match the GPU, and made the edges up to .5 too dark.