Recherche avancée

Sur d’autres sites (6106)

• Need help configuring FFMPEG to work with a webcams h264 stream

  9 août 2020, par The Welsh Dragon

  I have been trying to get a H264 stream from a H264 usb webcam working but I am not making much progress so I'm hoping someone knows FFMPEG better than me !

  


  There are dozens of questions/answers on SO but none solve my problem.

  


  In short, I get a very pixelated (or sometimes mostly green) screen. I am using VLC to test the stream which is coming via an RTSP server. I am using FFMPEG to copy the webcam stream to the local RTSP server.

  


  The webcam also supports YUYV which I can get working - it is just the h264 stream causing me problems.

  


  So this is how the device is presented :

  


  H264 USB Camera: USB Camera (usb-20980000.usb-1):
        /dev/video0
        /dev/video1
        /dev/video2
        /dev/video3


  


  /dev/video0 is the YUYV and MPEG stream
/dev/video2 is the h264 stream that has the following capabilities :

  


  ioctl: VIDIOC_ENUM_FMT
        Type: Video Capture

        [0]: 'H264' (H.264, compressed)
                Size: Discrete 1920x1080
                        Interval: Discrete 0.033s (30.000 fps)
                        Interval: Discrete 0.040s (25.000 fps)
                        Interval: Discrete 0.067s (15.000 fps)
                        Interval: Discrete 0.033s (30.000 fps)
                        Interval: Discrete 0.040s (25.000 fps)
                        Interval: Discrete 0.067s (15.000 fps)
                Size: Discrete 1280x720
                        Interval: Discrete 0.033s (30.000 fps)
                        Interval: Discrete 0.040s (25.000 fps)
                        Interval: Discrete 0.067s (15.000 fps)
                Size: Discrete 800x600
                        Interval: Discrete 0.033s (30.000 fps)
                        Interval: Discrete 0.040s (25.000 fps)
                        Interval: Discrete 0.067s (15.000 fps)
                Size: Discrete 640x480
                        Interval: Discrete 0.033s (30.000 fps)
                        Interval: Discrete 0.040s (25.000 fps)
                        Interval: Discrete 0.067s (15.000 fps)
                Size: Discrete 640x360
                        Interval: Discrete 0.033s (30.000 fps)
                        Interval: Discrete 0.040s (25.000 fps)
                        Interval: Discrete 0.067s (15.000 fps)
                Size: Discrete 352x288
                        Interval: Discrete 0.033s (30.000 fps)
                        Interval: Discrete 0.040s (25.000 fps)
                        Interval: Discrete 0.067s (15.000 fps)
                Size: Discrete 320x240
                        Interval: Discrete 0.033s (30.000 fps)
                        Interval: Discrete 0.040s (25.000 fps)
                        Interval: Discrete 0.067s (15.000 fps)
                Size: Discrete 1920x1080
                        Interval: Discrete 0.033s (30.000 fps)
                        Interval: Discrete 0.040s (25.000 fps)
                        Interval: Discrete 0.067s (15.000 fps)
                        Interval: Discrete 0.033s (30.000 fps)
                        Interval: Discrete 0.040s (25.000 fps)
                        Interval: Discrete 0.067s (15.000 fps)


  


  I have tried various resolutions, the smaller giving slightly less pixelated images but none are usable and definitely dont compare to the YUYV high resolution results.

  


  This (YUYV) command works :

  


  ffmpeg -input_format yuyv422 -f video4linux2 -s 1280x720 -r 10 -i /dev/video0 -c:v h264_omx -r 10 -b:v 2M -an -f rtsp rtsp://localhost:80/live/stream


  


  These two h264 options dont work :

  


  ffmpeg -input_format h264 -f video4linux2 -video_size 1920x1080 -framerate 30 -i /dev/video0 -c:v copy -an -f rtsp rtsp://localhost:80/live/stream


  


  ffmpeg -re -i /dev/video2 -video_size 800x600 -framerate 15 -pix_fmt yuv420p -tune zerolatency -c:v copy -an -f rtsp rtsp://localhost:80/live/stream


  


  For that last command the FFMPEG output looks like this :

  


  ffmpeg version git-2020-08-07-6fdf3cc Copyright (c) 2000-2020 the FFmpeg developers
  built with gcc 8 (Raspbian 8.3.0-6+rpi1)
  configuration: --extra-ldflags=-latomic --arch=armel --target-os=linux --enable-gpl --enable-omx --enable-omx-rpi --enable-nonfree --enable-libfreetype --enable-libx264 --enable-libmp3lame --enable-mmal --enable-indev=alsa --enable-outdev=alsa
  libavutil      56. 58.100 / 56. 58.100
  libavcodec     58.100.100 / 58.100.100
  libavformat    58. 50.100 / 58. 50.100
  libavdevice    58. 11.101 / 58. 11.101
  libavfilter     7. 87.100 /  7. 87.100
  libswscale      5.  8.100 /  5.  8.100
  libswresample   3.  8.100 /  3.  8.100
  libpostproc    55.  8.100 / 55.  8.100
Input #0, video4linux2,v4l2, from '/dev/video2':
  Duration: N/A, start: 1353.265049, bitrate: N/A
    Stream #0:0: Video: h264 (Main), yuv420p(progressive), 1920x1080, 30 fps, 30 tbr, 1000k tbn, 2000k tbc
[udp @ 0x38c29f0] attempted to set receive buffer to size 393216 but it only ended up set as 360448
[udp @ 0x38d7b50] attempted to set receive buffer to size 393216 but it only ended up set as 360448
Output #0, rtsp, to 'rtsp://localhost:80/live/stream':
  Metadata:
    encoder         : Lavf58.50.100
    Stream #0:0: Video: h264 (Main), yuv420p(progressive), 1920x1080, q=2-31, 30 fps, 30 tbr, 90k tbn, 1000k tbc
Stream mapping:
  Stream #0:0 -> #0:0 (copy)
Press [q] to stop, [?] for help
[rtsp @ 0x38fd890] Timestamps are unset in a packet for stream 0. This is deprecated and will stop working in the future. Fix your code to set the timestamps properly
[rtsp @ 0x38fd890] Non-monotonous DTS in output stream 0:0; previous: 0, current: 0; changing to 1. This may result in incorrect timestamps in the output file.
frame=  348 fps= 18 q=-1.0 size=N/A time=00:00:21.03 bitrate=N/A speed=1.09x


  


  The issue looks like it is bandwidth related or the lack of processing power in the device being used BUT the YUYV works at a high resolution and (taking a completely different approach i.e. not using FFMPEG) I can get a very decent MPEG stream working on the same device.

  


  So any FFMPEG experts out there who can help me with getting the correct parameters for a h264 stream ?

  


• ffmpeg : stream copy from .mxf into NLE-compatible format

  9 juin 2013, par David

  Because my NLE software does not support the .mxf-files from Canon XF100 I need to convert them into a supported format.

  As far as I know, mxf-files are just another container format for mpeg2 streams, so it would be really nice to extract the streams and place them into another container (without reencoding).

  I think ffmpeg can do this – correct me if I'm wrong – by running the following command :

  ffmpeg -i in.mxf -vcodec copy out.m2ts (or .ts, .mts, ...)

  ffmpeg finishes without errors after about 2 seconds (in.mxf is abut 170mb) :

  c:\video>c:\ffmpeg\bin\ffmpeg -i in.MXF -vcodec copy out.m2ts
  
  ffmpeg version N-53680-g0ab9362 Copyright (c) 2000-2013 the FFmpeg developers
  
    built on May 30 2013 12:14:03 with gcc 4.7.3 (GCC)
  
    configuration: --enable-gpl --enable-version3 --disable-w32threads --enable-av
  
  isynth --enable-bzlib --enable-fontconfig --enable-frei0r --enable-gnutls --enab
  
  le-iconv --enable-libass --enable-libbluray --enable-libcaca --enable-libfreetyp
  
  e --enable-libgsm --enable-libilbc --enable-libmodplug --enable-libmp3lame --ena
  
  ble-libopencore-amrnb --enable-libopencore-amrwb --enable-libopenjpeg --enable-l
  
  ibopus --enable-librtmp --enable-libschroedinger --enable-libsoxr --enable-libsp
  
  eex --enable-libtheora --enable-libtwolame --enable-libvo-aacenc --enable-libvo-
  
  amrwbenc --enable-libvorbis --enable-libvpx --enable-libx264 --enable-libxavs --
  
  enable-libxvid --enable-zlib
  
    libavutil      52. 34.100 / 52. 34.100
  
    libavcodec     55. 12.102 / 55. 12.102
  
    libavformat    55.  8.100 / 55.  8.100
  
    libavdevice    55.  2.100 / 55.  2.100
  
    libavfilter     3. 73.100 /  3. 73.100
  
    libswscale      2.  3.100 /  2.  3.100
  
    libswresample   0. 17.102 /  0. 17.102
  
    libpostproc    52.  3.100 / 52.  3.100
  
  Guessed Channel Layout for  Input Stream #0.1 : mono
  
  Guessed Channel Layout for  Input Stream #0.2 : mono
  
  Input #0, mxf, from 'in.MXF':
  
    Metadata:
  
      uid             : 1bb23c97-6205-4800-80a2-e00002244ba7
  
      generation_uid  : 1bb23c97-6205-4800-8122-e00002244ba7
  
      company_name    : CANON
  
      product_name    : XF100
  
      product_version : 1.00
  
      product_uid     : 060e2b34-0401-010d-0e15-005658460100
  
      modification_date: 2013-01-06 11:05:02
  
      timecode        : 01:42:14:22
  
    Duration: 00:00:28.32, start: 0.000000, bitrate: 51811 kb/s
  
      Stream #0:0: Video: mpeg2video (4:2:2), yuv422p, 1920x1080 [SAR 1:1 DAR 16:9
  
  ], 25 fps, 25 tbr, 25 tbn, 50 tbc
  
      Stream #0:1: Audio: pcm_s16le, 48000 Hz, mono, s16, 768 kb/s
  
      Stream #0:2: Audio: pcm_s16le, 48000 Hz, mono, s16, 768 kb/s
  
  Output #0, mpegts, to 'out.m2ts':
  
    Metadata:
  
      uid             : 1bb23c97-6205-4800-80a2-e00002244ba7
  
      generation_uid  : 1bb23c97-6205-4800-8122-e00002244ba7
  
      company_name    : CANON
  
      product_name    : XF100
  
      product_version : 1.00
  
      product_uid     : 060e2b34-0401-010d-0e15-005658460100
  
      modification_date: 2013-01-06 11:05:02
  
      timecode        : 01:42:14:22
  
      encoder         : Lavf55.8.100
  
      Stream #0:0: Video: mpeg2video, yuv422p, 1920x1080 [SAR 1:1 DAR 16:9], q=2-3
  
  1, 25 fps, 90k tbn, 25 tbc
  
      Stream #0:1: Audio: mp2, 48000 Hz, mono, s16, 128 kb/s
  
  Stream mapping:
  
    Stream #0:0 -> #0:0 (copy)
  
    Stream #0:1 -> #0:1 (pcm_s16le -> mp2)
  
  Press [q] to stop, [?] for help
  
  frame=  532 fps=0.0 q=-1.0 size=  143511kB time=00:00:21.25 bitrate=55314.1kbits
  
  frame=  561 fps=435 q=-1.0 size=  151254kB time=00:00:22.42 bitrate=55242.0kbits
  
  frame=  586 fps=314 q=-1.0 size=  158021kB time=00:00:23.41 bitrate=55288.0kbits
  
  frame=  609 fps=255 q=-1.0 size=  164182kB time=00:00:24.34 bitrate=55235.4kbits
  
  frame=  636 fps=217 q=-1.0 size=  171463kB time=00:00:25.42 bitrate=55235.1kbits
  
  frame=  669 fps=194 q=-1.0 size=  180133kB time=00:00:26.72 bitrate=55226.3kbits
  
  frame=  699 fps=173 q=-1.0 size=  188326kB time=00:00:27.92 bitrate=55256.6kbits
  
  frame=  708 fps=169 q=-1.0 Lsize=  190877kB time=00:00:28.30 bitrate=55233.6kbit
  
  s/s
  
  video:172852kB audio:442kB subtitle:0 global headers:0kB muxing overhead 10.1461
  
  18%

  Unfortunately the output file turns out to be displayed correctly only by vlc player.
  My NLE-software (Cyberlink Power Director) is able to open the file but most of the picture is green. Only a few pixels on the left edge show the original video :

  output file

  Any ideas how to solve that problem ? Is there a better way to use .mxf-files in NLE-software without native support ?

  thanks in advance

• How to batch process a series of video files with powershell and other-transcode/ffmpeg

  7 juin 2022, par DarkDiamond

  TL ;DR

  


  What did I do wrong in the following PowerShell-Script ? It does not work as expected.

  


  ---

  


  I am recording some of my lectures in my university with a photo camera. This works pretty well although I have to split the single lecture into three to four parts because the camera can only record 29 minutes of video in one take. I know that this is a common issue related to some licensensing problem that most photo cameras simply don't have the right license to record longer videos. But it confronts me with the problem that I later have to edit the files together after I did some post processing on them.

  


  With the camera I produce up to four video files with sizes around 3.5 GB which is way to big in order to be of any use because our IT department understandably doesn't want to host so much data, as I produce around 22 GB of video material each week.

  


  Some time ago I came across a very useful tool called "other-video-transcoding" by Don Melton over on GitHub, written in ruby, that allows me to compress the files to a reasonable file size without any visual loss. In addition I crop the videos to remove the part of each frame that is neither the board nor a place where my professor stands in order to decrease the filesize even further and do some privacy protection by cutting out most of the students.

  


  As the tools are accessable via the command line, it is relatively easy to configure and does not cost additional computational power to render a nice gui, so I can edit one of the 29 minute clips in less than 10 minutes.

  


  Now I wanted to optimize my workflow by writing a PowerShell script that only takes the parameters what to crop and which files to work on and then does the rest on its own so I can just start the script and then do something else while my laptop renders the new files.

  


  So far I have the following :

  


  $video_path = Get-ChildItem ..\ -Directory | findstr "SoSe"

Get-ChildItem $video_path -name | findstr ".MP4" | Out-File temp.txt -Append 
Get-Content temp.txt | ForEach-Object {"file " + $_} >> .\files.txt

Get-ChildItem $video_path |
Foreach-Object {
other-transcode --hevc --mp4 --target 3000 --crop 1920:780:0:0 $_.FullName
}

#other-transcode --hevc --mp4 --crop 1920:720:60:0 ..\SoSe22_Theo1_videos_v14_RAW\
ffmpeg -f concat -i files.txt -c copy merged.mp4
Remove-Item .\temp.txt


  


  but it does not quite do what I it expect to do.
This is my file system :

  


  sciebo/
└── SoSe22_Theo1_videos/
    ├── SoSe22_Theo1_videos_v16/
    │   ├── SoSe22_Theo1_videos_v16_KOMPR/
    │   │   ├── C0001.mp4
    │   │   ├── C0002.mp4
    │   │   ├── C0003.mp4
    │   │   ├── C0004.mp4
    │   │   ├── temp.txt
    │   │   ├── files.txt
    │   │   └── merged.mp4
    │   └── SoSe22_Theo1_videos_v16_RAW/
    │       ├── C0001.mp4
    │       ├── C0002.mp4
    │       ├── C0003.mp4
    │       └── C0004.mp4
    └── SoSe22_Theo1_videos_v17/
        ├── SoSe22_Theo1_videos_v17_KOMPR
        └── SoSe22_Theo1_videos_v17_RAW/
            ├── C0006.mp4
            ├── C0007.mp4
            ├── C0008.mp4
            └── C0009.mp4


  


  where the 16th lecture is already processed and the 17th is not. I always have the raw video data in the folders ending on RAW and the edited/compressed output files in the one ending on KOMPR. Note that the video files in the KOMPR folder are the output files of the other-transcode tool.

  


  The real work happens in the line where it says

  


  other-transcode --hevc --mp4 --target 3000 --crop 1920:780:0:0 $_.FullName


  


  and in the line

  


  ffmpeg -f concat -i files.txt -c copy merged.mp4


  


  where I concat the output files into the final version I can upload to our online learning platform.
What is wrong with my script ? In the end I'd like to pass the --crop parameter just to my script, but that is not the primary problem.

  


  ---

  


  A little information on the transcoding script so you don't have to look into the documentation :
  
As the last argument the tool takes the location of the video files to work on, be it relative or absolute file paths. The output is placed in the folder the script is called in, so if I cd into one of the KOMPR directories and then call

  


  other-transcode --mp4 ../SoSe22_Theo1_videos_v16_RAW/C0001.mp4


  


  a new file C0001.mp4 is created in the KOMPR directory and the transcoded video and old audio are written to that new video file.