VideoHelp Forum

+ Reply to Thread
Results 1 to 17 of 17
Thread
  1. I've demuxed a Blu-ray backup (so it wasn't from the disc) with MakeMKV and it was OK (I got no error messages) but while seeking I noticed some glitches. I tested with several players and I got the same result with every single one but if I extract the frames with ffmpeg all the image files look normal. Here is a clip: https://pixeldrain.com/u/JgM9f3SK



    With the clip if you pause the video and go here: 00:01.919 you will see the glitch.


    Here is a screenshot: https://i.imgur.com/9iGCG1D.png
    I have to clarify, this is ONLY noticeable when the video is paused and I'm seeking. If I play the video I don't see any glitch, even pausing and playing like crazy, but if I pause and then seek or go directly to the time it does appear.



    I also extracted all the frames from that clip and all the image files looked well too. This is not the only video file with which I'm having this problem, there is at least 1 or 2 more. What I want to know/understand here is:


    1. Is the video file corrupted?
    2. How is it that extracting all the frames doesn't show that corruption/glitch?
    3. If 1 is no, is my PC faulty, then? Or is the container?



    And no, the original m2ts file doesn't show this glitch, it looks normal.
    Quote Quote  
  2. It seems that the video can be decoded strictly sequentially only. Decoding may fail with random access. Maybe due to very long GOPs, or something odd with the timecodes.
    You can fix it by demuxing, re-encoding the video, remuxing
    Last edited by Sharc; 27th Jan 2022 at 03:12.
    Quote Quote  
  3. Hey, thanks for your help. I have tried remuxing with mkvtoolnix and the output doesn't have this problem, I guess this is worth reporting to MakeMKV? Or is random? Because as I said, there are other videos that have this problem, but there are others that don't too.
    And what I find more intriguing (maybe is because of my lack of knowledge about this subject) is that being MakeMKV just a remuxing tool (I mean, it can convert audio tracks but video tracks it just extracts them) how is it that is modifying the video tracks for them to have this problem?
    Quote Quote  
  4. I don't think that simple remuxing really helps. You should re-encode the video.
    But anyway, you may ask in the makeMKV forum. I don't have much experience with it.

    This should work:
    Code:
    ffmpeg.exe -i "your source" -map 0 -c:a copy -c:s copy -c:v libx264 -preset medium -bluray-compat 1 -crf 18 -g 48 -vf format=yuv420p "your_recoded.mkv"
    Image Attached Files
    Last edited by Sharc; 27th Jan 2022 at 06:48.
    Quote Quote  
  5. No, I didn't mean that I remuxed the mkv file, I remuxed the Blu-ray using MKVToolNix and the output file doesn't have any of these glitches. I found another user having similar issues than I and posted there.



    Thanks for the help.
    Quote Quote  
  6. Random seeking with out-of-order codecs (all the high compression codecs) can be very difficult, especially with variable frame rates. Many players don't bother trying, they just output corrupt video until a new keyframe comes along. Many have a "fast seek" setting which seeks to the nearest keyframe to prevent that corruption when seeking.
    Last edited by jagabo; 27th Jan 2022 at 13:21.
    Quote Quote  
  7. Code:
    Frame rate mode                          : Variable
    Frame rate                               : 31210.514
    Frame rate                               : 31 210.514 FPS
    Original frame rate                      : 23.976
    Original frame rate                      : 23.976 (24000/1001) FPS
    FrameRate_Original_Num                   : 24000
    FrameRate_Original_Den                   : 1001
    -> first thing I would try is remuxing the file while dropping all the time codes, otherwise any player follwing the container frame rate will fo funny things with that frame rate,...
    users currently on my ignore list: deadrats, Stears555
    Quote Quote  
  8. Originally Posted by Selur View Post
    Code:
    Frame rate mode                          : Variable
    Frame rate                               : 31210.514
    Frame rate                               : 31 210.514 FPS
    Original frame rate                      : 23.976
    Original frame rate                      : 23.976 (24000/1001) FPS
    FrameRate_Original_Num                   : 24000
    FrameRate_Original_Den                   : 1001
    -> first thing I would try is remuxing the file while dropping all the time codes, otherwise any player follwing the container frame rate will fo funny things with that frame rate,...
    I tried this forcing 24000/1001 fps and the framerate was then constant, but random seeking was still messy (MPC-HC), so I had to reencode it in order to fix it......
    Quote Quote  
  9. I don't see anything unsual in the time codes. From ffprobe:

    Code:
    entry, media_type, pkt_dts, pkt_dts_time, best_effort_Timestamp_time, duration, pkt_duration_time, ftype, coded# 
    frame,      video,       0,     0.000000,                   0.000000,       41,          0.041000,     I,      0
    frame,      video,      42,     0.042000,                   0.042000,       41,          0.041000,     B,      2
    frame,      video,      84,     0.084000,                   0.084000,       41,          0.041000,     P,      1
    frame,      video,     126,     0.126000,                   0.126000,       41,          0.041000,     B,      4
    frame,      video,     167,     0.167000,                   0.167000,       41,          0.041000,     P,      3
    frame,      video,     209,     0.209000,                   0.209000,       41,          0.041000,     B,      6
    frame,      video,     251,     0.251000,                   0.251000,       41,          0.041000,     P,      5
    frame,      video,     292,     0.292000,                   0.292000,       41,          0.041000,     B,      8
    frame,      video,     334,     0.334000,                   0.334000,       41,          0.041000,     P,      7
    frame,      video,     376,     0.376000,                   0.376000,       41,          0.041000,     B,      10
    frame,      video,     418,     0.418000,                   0.418000,       41,          0.041000,     P,      9
    frame,      video,     459,     0.459000,                   0.459000,       41,          0.041000,     B,      12
    frame,      video,     501,     0.501000,                   0.501000,       41,          0.041000,     P,      11
    frame,      video,     543,     0.543000,                   0.543000,       41,          0.041000,     B,      14
    frame,      video,     584,     0.584000,                   0.584000,       41,          0.041000,     P,      13
    frame,      video,     626,     0.626000,                   0.626000,       41,          0.041000,     B,      16
    frame,      video,     668,     0.668000,                   0.668000,       41,          0.041000,     P,      15
    frame,      video,     710,     0.710000,                   0.710000,       41,          0.041000,     B,      18
    frame,      video,     751,     0.751000,                   0.751000,       41,          0.041000,     P,      17
    frame,      video,     793,     0.793000,                   0.793000,       41,          0.041000,     B,      20
    frame,      video,     835,     0.835000,                   0.835000,       41,          0.041000,     I,      19
    It uses the usual method of alternating pkt_dts increments to achieve a non integer average.
    Quote Quote  
  10. For anyone reading this thread, check your remuxed videos if you used MakeMKV, I've been checking and I have several with this problem, from 10 I've remuxed recently with MakeMKV 8 have this issue and only 2 are glitch-free.


    And as I said, I don't understand how is it that MakeMKV is making these glitched files but MKVToolNix doesn't. I have also compared both video tracks and they are not the same, starting from the size (in bytes) both avc files are different, hashing also outputs different hashes.
    Aren't these two application supposed to make 1:1 copies of the video files?
    Quote Quote  
  11. Member
    Join Date
    Mar 2008
    Location
    United States
    Search Comp PM
    What was the source, a m2ts file, a blu-ray folder structure,
    Mpeg2, HEVC/AVC, VC-1, etc
    Quote Quote  
  12. Blu-ray folder so a m2ts file.
    Quote Quote  
  13. Here is a clip between same times as the other from the output file from MKVToolNix: https://pixeldrain.com/u/rZ67YAc2


    Tried remuxing with MKVToolNix another movie that has this glitches and again, the output from MTX doesn't have this issue.
    Last edited by orlydb; 27th Jan 2022 at 17:55.
    Quote Quote  
  14. Originally Posted by jagabo View Post
    I don't see anything unsual in the time codes. From ffprobe:

    Code:
    entry, media_type, pkt_dts, pkt_dts_time, best_effort_Timestamp_time, duration, pkt_duration_time, ftype, coded# 
    frame,      video,       0,     0.000000,                   0.000000,       41,          0.041000,     I,      0
    frame,      video,      42,     0.042000,                   0.042000,       41,          0.041000,     B,      2
    frame,      video,      84,     0.084000,                   0.084000,       41,          0.041000,     P,      1
    frame,      video,     126,     0.126000,                   0.126000,       41,          0.041000,     B,      4
    frame,      video,     167,     0.167000,                   0.167000,       41,          0.041000,     P,      3
    frame,      video,     209,     0.209000,                   0.209000,       41,          0.041000,     B,      6
    frame,      video,     251,     0.251000,                   0.251000,       41,          0.041000,     P,      5
    frame,      video,     292,     0.292000,                   0.292000,       41,          0.041000,     B,      8
    frame,      video,     334,     0.334000,                   0.334000,       41,          0.041000,     P,      7
    frame,      video,     376,     0.376000,                   0.376000,       41,          0.041000,     B,      10
    frame,      video,     418,     0.418000,                   0.418000,       41,          0.041000,     P,      9
    frame,      video,     459,     0.459000,                   0.459000,       41,          0.041000,     B,      12
    frame,      video,     501,     0.501000,                   0.501000,       41,          0.041000,     P,      11
    frame,      video,     543,     0.543000,                   0.543000,       41,          0.041000,     B,      14
    frame,      video,     584,     0.584000,                   0.584000,       41,          0.041000,     P,      13
    frame,      video,     626,     0.626000,                   0.626000,       41,          0.041000,     B,      16
    frame,      video,     668,     0.668000,                   0.668000,       41,          0.041000,     P,      15
    frame,      video,     710,     0.710000,                   0.710000,       41,          0.041000,     B,      18
    frame,      video,     751,     0.751000,                   0.751000,       41,          0.041000,     P,      17
    frame,      video,     793,     0.793000,                   0.793000,       41,          0.041000,     B,      20
    frame,      video,     835,     0.835000,                   0.835000,       41,          0.041000,     I,      19
    It uses the usual method of alternating pkt_dts increments to achieve a non integer average.
    Nothing unusual, yes. Maybe it's only the new I,P,B distribution which made the re-encode better seekable without producing glitches.

    On the other hand, when random stop/play/seeking within the OP's file using ffplay I am getting alerts like:
    Code:
    Seek to 55% ( 0:00:02) of total duration ( 0:00:04)       B f=0/0
    Seek to 42% ( 0:00:01) of total duration ( 0:00:04)       B f=0/0
    [h264 @ 000001519da8b180] top block unavailable for requested intra mode -1
    [h264 @ 000001519da8b180] error while decoding MB 0 0, bytestream 65896
    [h264 @ 000001519da8b180] left block unavailable for requested intra mode
    [h264 @ 000001519da8b180] error while decoding MB 0 16, bytestream 152837
    [h264 @ 000001519da8b180] top block unavailable for requested intra mode
    [h264 @ 000001519da8b180] error while decoding MB 0 34, bytestream 128748
    [h264 @ 000001519da8b180] top block unavailable for requested intra mode
    [h264 @ 000001519da8b180] error while decoding MB 0 50, bytestream 74817
    [h264 @ 000001519da8b180] concealing 8160 DC, 8160 AC, 8160 MV errors in I frame
    [h264 @ 000001519e76b0c0] top block unavailable for requested intra mode
    [h264 @ 000001519e76b0c0] error while decoding MB 53 34, bytestream 59007
    [h264 @ 000001519e76b0c0] top block unavailable for requested intra mode
    [h264 @ 000001519e76b0c0] error while decoding MB 62 50, bytestream 33570
    [h264 @ 000001519e76b0c0] concealing 7848 DC, 7848 AC, 7848 MV errors in P frame
    [h264 @ 000001519e76b7c0] top block unavailable for requested intra mode
    [h264 @ 000001519e76b7c0] error while decoding MB 11 34, bytestream 43880
    [h264 @ 000001519e76b7c0] top block unavailable for requested intra mode -1
    [h264 @ 000001519e76b7c0] error while decoding MB 4 50, bytestream 24374
    [h264 @ 000001519e1ed4c0] cabac decode of qscale diff failed at 35 1
    [h264 @ 000001519e1ed4c0] error while decoding MB 35 1, bytestream 36766
    [h264 @ 000001519e76b7c0] concealing 7083 DC, 7083 AC, 7083 MV errors in B frame
    [h264 @ 000001519e1ed4c0] top block unavailable for requested intra mode
    [h264 @ 000001519e1ed4c0] error while decoding MB 19 34, bytestream 67328
    [h264 @ 000001519e1ed4c0] top block unavailable for requested intra mode
    [h264 @ 000001519e1ed4c0] error while decoding MB 19 50, bytestream 35641
    [h264 @ 000001519e1ed4c0] concealing 7915 DC, 7915 AC, 7915 MV errors in P frame
    [h264 @ 000001519e1ed8c0] top block unavailable for requested intra mode
    [h264 @ 000001519e1ed8c0] error while decoding MB 19 0, bytestream 34250
    [h264 @ 000001519e1ecd00] top block unavailable for requested intra mode -1
    [h264 @ 000001519e1ecd00] error while decoding MB 1 16, bytestream 51346
    [h264 @ 000001519e1ed8c0] top block unavailable for requested intra mode -1
    [h264 @ 000001519e1ed8c0] error while decoding MB 93 16, bytestream 71937
    [h264 @ 000001519e1ecd00] top block unavailable for requested intra mode -1
    [h264 @ 000001519e1ecd00] error while decoding MB 27 50, bytestream 22253
    [h264 @ 000001519e1ed8c0] top block unavailable for requested intra mode
    [h264 @ 000001519e1ed8c0] error while decoding MB 46 34, bytestream 55991
    [h264 @ 000001519e1ecd00] concealing 7079 DC, 7079 AC, 7079 MV errors in B frame
    [h264 @ 000001519e1ed8c0] top block unavailable for requested intra mode
    [h264 @ 000001519e1ed8c0] error while decoding MB 4 50, bytestream 33836
    [h264 @ 000001519e1ed8c0] concealing 8116 DC, 8116 AC, 8116 MV errors in P frame
    [h264 @ 000001519da8b180] top block unavailable for requested intra mode
    [h264 @ 000001519da8b180] error while decoding MB 12 0, bytestream 34144
    [h264 @ 000001519e1ec140] top block unavailable for requested intra mode
    [h264 @ 000001519e1ec140] error while decoding MB 30 34, bytestream 39661
    [h264 @ 000001519e1ec140] top block unavailable for requested intra mode -1
    [h264 @ 000001519e1ec140] error while decoding MB 2 50, bytestream 21517
    [h264 @ 000001519da8b180] concealing 8118 DC, 8118 AC, 8118 MV errors in P frame
    [h264 @ 000001519e1ec140] concealing 6840 DC, 6840 AC, 6840 MV errors in B frame
    [h264 @ 000001519e76b7c0] top block unavailable for requested intra mode
    When I do the same with the re-encoded video I cannot provoke these alerts.
    No idea if this has some relevance. Just curious.
    Quote Quote  
  15. I don't see anything unsual in the time codes
    Then the question is: Why does MediaInfo report that frame rate....
    users currently on my ignore list: deadrats, Stears555
    Quote Quote  
  16. Originally Posted by Selur View Post
    I don't see anything unsual in the time codes
    Then the question is: Why does MediaInfo report that frame rate....
    It's obscure where it takes that huge framerate from. Even for the re-encoeded glitch-free seekable variant it reports
    Code:
    Frame rate mode : Variable 
    Frame rate : 31 203.320 FPS 
    Original frame rate : 23.976 (24000/1001) FPS
    Quote Quote  
  17. Even for the re-encoeded glitch-free seekable variant it reports
    Okay, that is really unexpected.
    users currently on my ignore list: deadrats, Stears555
    Quote Quote  



Similar Threads