wolframrhodium / vapoursynth-bilateralgpu Goto Github PK

Would be good to have one like the non-cuda VS-Bilateral

This code fails with the following error:

import vapoursynth as vs
core = vs.core

clip = core.std.BlankClip(width=1440, height=810)
clip = clip.bilateralgpu.Bilateral()

clip.set_output(0)

>>>vspipe -p test.py .   
Script evaluation done in 1.16 seconds
Error: Failed to retrieve frame 0 with error: BilateralGPU: 'cuStreamSynchronize(stream)' failed: an illegal memory access was encountered
Output 24 frames in 2.19 seconds (10.98 fps)
Warning: Core freed but 84049920 bytes still allocated in framebuffers

Using clip = clip.bilateralgpu_rtc.Bilateral(block_y=10) works though. Is the default block_y the cause?
Tested with r7. r6 works fine.

If the input clip is 16 bit then the output is identical to the input. 8 bit works fine for me.

Confirmed with the following script:

from vapoursynth import core, YUV420P16

src = core.ffms2.Source("Path/to/some/file.mkv")
src_16 = src.resize.Bicubic(format=YUV420P16)
bilateral = core.bilateralgpu.Bilateral(src_16, sigma_spatial=10, sigma_color=2.25, kernel_size=0, borderMode=4, planes=[0,1,2])
bilateral = core.std.PlaneStats(bilateral, src_16)

for x in bilateral[300:340].frames():
    print(x.props["PlaneStatsDiff"])

This script will always print 0.0.

I'm on Linux by the way.

I seen you had BM3DCUDA & EEDI2CUDA can work with vs & avs+. Can you possible port BilateralGPU to avs+?? Thanks

Hello,
I noticed the top row of this sample are like mirrored from the bottom row. You can see a darker blurred row at the top of bilateralgpu and bilateralgpu_rtc while tcanny doesn't have this issue.

clip.tcanny.TCanny(1, 1, mode=-1).set_output(0)
clip.bilateralgpu.Bilateral(1, 1.0).set_output(1)
clip.bilateralgpu_rtc.Bilateral(1, 1.0).set_output(2)

https://slow.pics/c/SWQNdew5