Recherche avancée

Sur d’autres sites (9527)

• Extracting a bit broken AC3 audio from a video file

  3 janvier 2023, par WYSIWYG

  I Googled for about 2 hours for this problem and I can't seem to find the answer anywhere.

  


  So I have a 1080p video file which has about 2 seconds within it that's broken (at 3 minutes and 53 seconds). I found a way to extract the non-broken parts of it, also I got a replacement for the broken frames from another version of the same video (in 720p), upscaled them to 1080p and replaced them in the 1080p file (reencoding the whole file in the process). Here comes the problematic part - at the same time (3 min 53 sec), the audio is also broken and it's in Dolby Digital 5.1 AC3 format. When I try to extract the audio track from the original 1080p file and place the result in the new fixed version, which contains the upscaled video frames, the audio obviously is cut at the problematic 2 seconds and it desyncs (starts to play earlier) for the rest of the file.

  


  What I want to do and I'm unable to understand how to do it is to extract the AC3 stream in such a way that the 2 broken seconds are filled with silence and the total length of the audio stay the same as the original, so it won't desync. Is this possible and if it is, what is the correct FFMPEG command for it ?

  


  Thank you in advance.

  


  I've tried :

  


  ffmpeg -i broken.mkv -c:a copy audio.ac3

  


  which presented an output AC3 file, which is missing the 2 seconds from 03:53 to 03:55

  


  so when I combine the resulting AC3 file with the repaired video, by using :

  


  ffmpeg -i fixed.mkv -i audio.ac3 -c copy fixed-with-audio.mkv

  


  it is desynced, so at video time 03:53 I hear the audio that should play at 03:55

  


• ffmpeg combine images with audio and fill gaps with silence or last image (from ffconcat file..)

  27 mai 2020, par inselmensch

  ffmpeg -i image_list.ffconcat -i track.mp3 -y -b:a 256k -ar 48000 -ac 2 -vcodec libx264 -r 25 -filter:v scale=w=1920:h=1080 output.mp4


  



  i'm currently trying to create videos out of an sequence of images and add an audio track to those images/this video.
everything seemed to work fine but now i figured out that the last picture will only be shown within my player since there is no new image for displaying.

  



  my problem is now : i try to combine two of those videos but if there are some gaps (not enough images or not enough audio)... ffmpeg just ignores the gaps and lets start the second video immediately after finishing the first one ... even if the first track is still playing... which gets the audio off..

  



  is there anybody who can might help me with it ? :-)

  



  i think, what i need is :

  



  

  • if the audio track is longer than the image sequence i need to fill the gap with the last image

  


  • if the image sequence is longer than the audio track i need to fill the remaining space with silence

  


  



  or is there some kind of option which tells ffmpeg to only chain both videos after each other without ignoring the possible gaps ?

  



  if found this one but that's just for the audio part i guess and i will try to implement that.

  



  but how do i work around the missing images issue ?!

  


• Using FFMPEG to overlay 2 concat-demux videos

  10 juin 2021, par marco

  I am making a screen recording software that captures gestures/mouse moves as coordinates+timestamps, and would like to overlay those on the screen recording. I have a folder of "Frames" that have a red circle and the rest transparent that correspond to the gestures.

  


  The way I initially got it working was to use concat-demux to generate a video of the gestures, and then I overlay it with ffmpeg by chromakeying out the black background and overlaying it, but that is very slow.

  


  Is there a better/faster way to do it that maybe doesn't make an intermediate video that I have to chromakey ? I have access to the timestamps of the gestures as well as the duration these frames should stay on screen.

  


  my pipeline :

  


  Screen Frames w/ Timestamps+durations --> Concat-demux --> ---------------------------+
                                                                                      |
Gesture Logs --> Frames w/ timestamp+duration --> Concat-demux ---> chromakey --+     |
                    (has transparency)        (loses transparency)              |     |
                                                                                V     V
                                                                               Overlay
                                                                                  |
                                                                                  V
                                                                             Final Video


  


  Example frame from gestures which is the foreground
(The white is transparent)

  


  The background has no transparency as it is a computer background

  


  The command I used to combine the videos I generated is :
ffmpeg -i gestures.mp4 -i screen.mp4 -filter_complex '[0]chromakey=0x000000:.1[bk];[1][bk]overlay' output.mp4