Sur d’autres sites (4996)

• Recording video in C#

  31 juillet 2014, par pfedotovsky

  I have to do the following : record video from camera using C#. Camera I use produces video frames (framerate is not fixed) and I have to somehow put all the frames together and create video file. Also I need to use different codecs such as AVI or MPEG4 (these codecs are required, others are optional).

  The main problem I faced is how to create video in which framerate is not fixed. I have a stream of frames, for example I can recive the first frame after 1ms, then after 20ms, then 36ms and so on. If I create video with 25 frames/second the result will be wrong because it means that frames are added after 40ms.

  I tried to use Aforge.Video library, it has a method http://www.aforgenet.com/framework/docs/html/84a560df-bfd5-e0d6-2812-f810b56a254d.htm which adds a frame according to timespan. But this method has problems with setting bitrate - the bitrate value I pass to the method simply ignored (http://www.aforgenet.com/forum/viewtopic.php?f=2&t=2665).

  So my questions is : is there some C# library which I can use to video recording ? I have to support AVI and MPEG4, also possibility to set bitrate and the last but not the least - record video with variable framerate.

  Thanks in advance. Any suggestions would be very helpful.

  UPD1 : I can’t connect to camera directly - all I have is a stream of frames and I need to convert this stream to video at run-time.

  UPD2 : A library I’m looking for should satisfy the following properties. It has to contain a method (or some way how to do the same) to add next frame with a timestamp just like in Aforge.Video.FFMPEG :

  public void WriteVideoFrame(Bitmap frame, TimeSpan timestamp)

  And it should be possible to choose different codecs (at least AVI and MPEG4) and also to set bitrate.
  Is there some alternatives to Aforge.Video.FFMPEG ? Because Aforge doesn’t work properly - the bitrate value is ignored, also some codecs are not supported (MPEG2 for example).

  UPD3 : One more questions about codec license. If I use open source library, should I worry about the codec license ?

• How to replace a snippet in a video without reencoding the whole video

  3 novembre 2022, par Simon Streicher

  I am trying to edit and replace a section of a video without reencoding the whole video.
Here are the steps I think I need to take :

  


  1. Find keyframes

  


  input :

  


  ffprobe -v error -select_streams v:0 -skip_frame nokey -show_entries frame=pkt_pts_time -of csv=p=0 'example.mkv'


  


  output :

  


  0.000000
1.001000
11.011000
13.430000
20.812000
30.822000
40.832000
50.842000
⋮


  


  2. Export relevant section

  


  For example, export the snippet 40.832000 → 50.842000.

  


  3. Edit and reencode

  


  After editing that section, I need to reencode it to the original codecs for compatibility with the surrounding video. For example, this is the original codecs :

  


  Stream #0:0: Video: hevc (Main 10), yuv420p10le(tv, bt2020nc/bt2020/smpte2084), 3840x1600 [SAR 1:1 DAR 12:5], 23.98 fps, 23.98 tbr, 1k tbn, 23.98 tbc (default)


  


  4. Inject edit back into the video stream

  


  Finally, I need to use FFmpeg somehow to retain everything from the original file (audio, subtitles, chapters, etc.) and to construct a new video stream (for example, Stream #0:0) that is exactly 0 → 40.832000 of the original, the whole edited section, and 50.842000 → end of the original section.

  


  My main questions are :

  


  

  • A. How do I trim the video stream in 2. without reencoding ?

  


  • B. Assuming that the edited video's resolution will remain the same, what is the command for FFmpeg to encode my edit to the codecs in 3. (and would the video be concatenable with the original video) ?

  


  • C. How should I go about glueing the sections together ? Should I simply trim Stream #0:0 into sections v1 = 0 → 40.832 and v2 = 50.842 → end and then concatenate a new stream as new = v1 + edited + v2 ?

  


  • D. How do I replace Stream #0:0 with new ?

  


  


  And probably the most important question : are my assumptions correct and can this be achieved ?

  


• avcodec : add option to make is_intra_more_likely() from error concealment return...

  2 mai 2014, par Michael Niedermayer

  avcodec : add option to make is_intra_more_likely() from error concealment return "no"
  Signed-off-by : Michael Niedermayer <michaelni@gmx.at>
  

  • [DH] doc/codecs.texi
  • [DH] libavcodec/avcodec.h
  • [DH] libavcodec/error_resilience.c
  • [DH] libavcodec/options_table.h
  • [DH] libavcodec/version.h