Recherche avancée

Sur d’autres sites (7928)

• accurate cutting of video (+ audio) with ffmpeg [migrated]

  4 novembre 2012, par Pete Oakey

  I want my website to allow users to accurately create their own clips from a source video I provide.

  I have a source video file that I first want to convert to something suitable for a website :

  Input #0, matroska,webm, from 'source.mkv': 
  
  Duration: 00:28:18.57, start: 0.000000, bitrate: 10183 kb/s 
  
  Stream #0:0: Video: h264 (Constrained Baseline), yuv420p, 1920x1080 [SAR 1:1 DAR 16:9], 23.98 fps, 23.98 tbr, 1k tbn, 48 tbc (default)
  
  Stream #0:1: Audio: mp3, 44100 Hz, stereo, s16, 128 kb/s (default)

  So I use ffmpeg to convert it like so :

  ffmpeg -i source.mkv -c:v libx264 -c:a aac -strict experimental -vf scale="960:-1" source.mp4

  Watching this video back it is good enough quality and small enough file size for my needs, and loads/plays on my website.

  So I have a webpage that lets users select a start and endpoint on this video - and create a clip. Here's an example of the ffmpeg command that I use for that :

  -ss 577.920 -i source.mp4 -t 011.980 -codec:v copy -codec:a copy -vf scale="960:-1" clip1.mp4

  The problem is the clip is not always time-accurate enough. Usually the audio is accurate enough but the video stops half a second early or something.

  Is there any way to make this accurate and in-sync to say, 0.2 of a second ?

  EDIT :
  Adding -force_key_frames 00:00:00.2 didn't help.

• vdpau : add av_vdpau_bind_context()

  4 octobre 2014, par Rémi Denis-Courmont

  vdpau : add av_vdpau_bind_context()
  This function provides an explicit VDPAU device and VDPAU driver to
  
  libavcodec, so that the application is relieved from codec specifics
  
  and VdpDevice life cycle management.
  A stub flags parameter is added for future extension. For instance, it
  
  could be used to ignore codec level capabilities (if someone feels
  
  dangerous).
  Signed-off-by : Anton Khirnov <anton@khirnov.net>
  

  • [DBH] doc/APIchanges
  • [DBH] libavcodec/vdpau.c
  • [DBH] libavcodec/vdpau.h
  • [DBH] libavcodec/vdpau_internal.h
  • [DBH] libavcodec/version.h

• How to acurately trim audio and video with ffmpeg ? [closed]

  6 janvier 2024, par ws90

  I'm trying to automate the trimming and concatenation of video clips that also contain audio using ffmpeg.
The following commands are being used to trim clips and then concatenate the trimmed clips.

  


  .\ffmpeg -ss $startInSeconds -i $inputFile -t $partDurationInSeconds $outputFile


  


  This is done once per input file, the values of $startInSeconds and $partDurationInSeconds are different for different clips.

  


  .\ffmpeg -f concat -safe 0 -i .\list.txt -c copy $concatOutputFile


  


  list.txt is a list of $outputFile from the first trim command.

  


  The audio of the concat video file gradually de-syncs over time (slowing down compared to the trimmed clips), which is the problem I'm looking to fix.
It seems to slow down by about half a frame at each concat join.

  


  Why is the concat command causing this de-sync and what can I do about it ?

  


  I thought this was due to a mismatch in duration between the audio and video tracks of a trimmed clip (the audio would often be shorter than the video after trimming). I then tried padding the audio to match the video before concatenating but the problem persists.
I also found an example where the tracks were identical in length and the problem persists, so I think the concat command is the culprit, not the trim command.