VideoHelp Forum
+ Reply to Thread
Page 2 of 2
FirstFirst 1 2
Results 31 to 42 of 42
Thread
  1. @poisondeathray: Not really concerned about the colors since I don't see how a color shift cout lead to these artifacts, but here's the whole script:
    Code:
    # Imports
    import vapoursynth as vs
    # getting Vapoursynth core
    core = vs.core
    import site
    # Import libraries for onnxruntime
    import ctypes
    path = site.getsitepackages()[0]+'/onnxruntime_dlls/'
    ctypes.windll.kernel32.SetDllDirectoryW(path)
    # Loading Plugins
    core.std.LoadPlugin(path="i:/Hybrid/64bit/vsfilters/Support/fmtconv.dll")
    core.std.LoadPlugin(path="i:/Hybrid/64bit/vsfilters/SourceFilter/LSmashSource/vslsmashsource.dll")
    # source: 'C:\Users\Selur\Desktop\A_source_lagarith_RGB_720x480_1.78AR.avi'
    # current color space: RGB24, bit depth: 8, resolution: 720x480, fps: 29.97, color matrix: 470bg, yuv luminance scale: limited, scanorder: progressive
    # Loading C:\Users\Selur\Desktop\A_source_lagarith_RGB_720x480_1.78AR.avi using LWLibavSource
    clip = core.lsmas.LWLibavSource(source="C:/Users/Selur/Desktop/A_source_lagarith_RGB_720x480_1.78AR.avi", format="RGB24", stream_index=0, cache=0, fpsnum=30000, fpsden=1001, prefer_hw=0)
    clip = clip if not core.text.FrameProps(clip,'_Transfer') else core.std.SetFrameProps(clip, _Transfer=5)
    clip = clip if not core.text.FrameProps(clip,'_Primaries') else core.std.SetFrameProps(clip, _Primaries=5)
    # Setting color range to TV (limited) range.
    clip = core.std.SetFrameProp(clip=clip, prop="_ColorRange", intval=1)
    # making sure frame rate is set to 29.970
    clip = core.std.AssumeFPS(clip=clip, fpsnum=30000, fpsden=1001)
    # adjusting color space from RGB24 to RGBS for vsBasicVSRPP
    clip = core.resize.Bicubic(clip=clip, format=vs.RGBS, range_s="limited")
    # resizing using BasicVSR
    from vsbasicvsrpp import BasicVSRPP
    clip = BasicVSRPP(clip=clip, model=0, tile_x=144, tile_y=96, tile_pad=10) # 2880x1920
    # adjusting output color from: RGBS to YUV420P10 for x265Model
    clip = core.resize.Bicubic(clip=clip, format=vs.YUV420P10, matrix_s="470bg", range_s="limited", dither_type="none")
    # set output frame rate to 29.97fps
    clip = core.std.AssumeFPS(clip=clip, fpsnum=30000, fpsden=1001)
    # Output
    clip.set_output()
    I used. I suspect that this is some NVIDIA driver issue.
    users currently on my ignore list: deadrats, Stears555
    Quote Quote  
  2. Originally Posted by Selur View Post
    I suspect that this is some NVIDIA driver issue.
    Then why is model 0 only affected for you? I'd expect a driver issue to affect everything
    Quote Quote  
  3. No clue, might also be the Vapoursynth version somehow. I use R60RC1. Anything that comes to mind should cause issues on multiple fronts not just with BasicVSR++ and model=0. :/
    users currently on my ignore list: deadrats, Stears555
    Quote Quote  
  4. Video Damager VoodooFX's Avatar
    Join Date
    Oct 2021
    Location
    At Doom9
    Search PM
    Originally Posted by poisondeathray View Post
    ...
    It looks too good.
    Is it your personal sample or some from public, can you post original UHD for comparison? How long it took to generate one frame?
    Quote Quote  
  5. Originally Posted by VoodooFX View Post
    Originally Posted by poisondeathray View Post
    ...
    It looks too good.
    Is it your personal sample or some from public, can you post original UHD for comparison? How long it took to generate one frame?

    Original source was some "random" UHD video on this forum. Bicubic downsample to 720x480 .
    https://forum.videohelp.com/threads/402943-The-quality-of-SONY-FDR-Ax700-lossless-NVEN...ith-Camlink-4K

    But it wasn't a completely "random" choice - I chose one where there was fwd camera movement, to test the propogation and temporal SR aspects by using various intervals (or by trimming sections), and also one that had text/numbers because many machine learning algorithms fail miserably with text/numbers. And ideally one shot with a faster shutter to minimize motion blur. (There are other algorithms that reverse motion blur that you can test separately)

    I didn't measure FPS (more like SPF ), but as mentioned earlier, it's very very slow. Larger interval sizes and more tiles make it slower too.
    Quote Quote  
  6. Man of course it was Truthlers video, don't know any other person lacking so many camera-usinf skills. Tho a lack of low shutter speed can be a good thing
    Last edited by rrats; 6th Sep 2022 at 18:01.
    Quote Quote  
  7. What the heck is that stupid RGBS, it's been annoying me for over an hour

    Click image for larger version

Name:	UAlNKoO.png
Views:	609
Size:	62.2 KB
ID:	66633
    Last edited by rrats; 6th Sep 2022 at 18:00.
    Quote Quote  
  8. Originally Posted by s-mp View Post
    What the heck is that stupid RGBS, it's been annoying me for over an hour

    Image
    [Attachment 66633 - Click to enlarge]
    RGBS is single precision float (32-bit) RGB

    If the input is YUV, you have to specify the in matrix, e.g for "SD" or "601" you would use "170m"


    c = core.ffms2.Source(r'video.mp4')
    c = core.resize.Bicubic(c, format=vs.RGBS, matrix_in_s="170m")
    .
    .
    .
    Quote Quote  
  9. Thank you, will try tomorrow
    Quote Quote  
  10. Starndard script doesn't seem to be doing anything the way I wrote it. Anyone knows why?
    Quote Quote  
  11. Judging by your screenshot, the issue is probably because you assign the output of BasicVSR to ret, but output c,....
    users currently on my ignore list: deadrats, Stears555
    Quote Quote  
  12. Originally Posted by Selur View Post
    Judging by your screenshot, the issue is probably because you assign the output of BasicVSR to ret, but output c,....
    ah i see. im new to the vapoursynth thanks for the help
    Quote Quote  



Similar Threads

Visit our sponsor! Try DVDFab and backup Blu-rays!