Recherche avancée

Sur d’autres sites (13477)

• How to delay audio after a specific position with ffmpeg ?

  11 septembre 2020, par C.K.

  I have a 10 seconds a.mp4 with two streams : Stream #0 is a video stream and Stream #1 is a audio stream.

  


  Now, I want to delay the audio stream by 4 seconds after the time position 00:03. It is to say, in the output file, I want that : 00:00-00:03 is the original audio, 00:03-00:07 has no sound, 00:07-00:14 is the original 00:03-00:10 audio.

  


  I've tried this :

  


  ffmpeg -i a.mp4 -t 00:00:03 -i a.map4 -itsoffset 4 -ss 00:00:03 -i a.mp4 -map 0:v -map 1:a -map 2:a -codec copy output.mp4


  


  But it seems that there are two audio streams in the output.mp4 and only one of them can be played once. Then I tried amix filter :

  


  ffmpeg -i a.mp4 -t 00:00:03 -i a.mp4 -itsoffset 4 -ss 00:00:03 -i a.mp4 -filter_complex "[1:a][2:a] amix=inputs=2" -map 0:v output.mp4


  


  But it also doesn't work. I'm new to ffmpeg so I have no idea what should I do now ? Any idea for me ? Very much thanks !

  


• How can I concatenate MP4 videos and exactly match this combined video to an audio track in ffmpeg ?

  17 juin 2020, par mattador

  Basically I want to combine 4 videos into a single video, then swap out the concatenated audio for another track. The audio track needs to be unaltered and sync exactly to the video.

  



  Specifically, I'm creating multipaged dynamic audio spectrograms (visualizations of a sound file) using ggplot2 and gganimate in R, which I want to combine using ffmpeg. So the original WAV is chopped up into 3 sec segments, a video is created to visualize frequency changes over time for each segment, which I want to combine into 1 video with the original audio track matched up.

  



  All of my relevant files are at github here in the vids folder : DL link here

  



  The individual dynamic spectrograms export fine (1-4.mp4 in the archive). They are each 20ms too long ; they should be exactly 3s, but otherwise great.

  



  I then use ffmpeg to concatenate them :

  



  ffmpeg -f concat -safe 0 -i "mp4Segments.txt" -codec copy -t 12 -y "(1-4)vid.mp4"


  



  but you can see the end of (1-4)vid.mp4 is getting a little off (the cursor is not all the way to the end, so I guess the video is getting cut off before the end. Not a deal-breaker, since the audio still aligns very well to the video cursor.

  



  Now, since the concatenated video introduces glitchy audio between clips, I want to take the uncut original audio and add that to the assembled video.

  



  ffmpeg -ss 0 -i "(1-4)vid.mp4" -i "origWAV.wav"  -c:v libx264 -map 0:v:0 -map 1:a:0 -c:a aac -ac 1 -b:a 192k -y -vsync 1 -t 12 "(1-4)vid+origWAV.mp4"


  



  but the audio is slightly off now, and ends even further before the end of the video. Since it's really important to align audio and video precisely for this, how can I speed up the video to fit the exact duration of the audio or otherwise fix this issue ?

  


• lavfi/vf_vpp_qsv : allow special values for the output video dimensions

  9 janvier 2023, par Haihao Xiang

  lavfi/vf_vpp_qsv : allow special values for the output video dimensions
  Special values are :
  
  0 = original width/height
  
   1 = keep original aspect
  This is in preparation for reusing the code for other QSV filters.
  Signed-off-by : Haihao Xiang <haihao.xiang@intel.com>
  

  • [DH] libavfilter/vf_vpp_qsv.c