I have this video file that at one point introduces an async of ~1 second. So I tried to make it work by duplicating a short segment at that point as it's silent there:
$ ffmpeg -i DFW.mp4.mkv -filter_complex ' [0:a] aselect=n=3:e=lt(t\,18*60+42.100)+gt(t\,18*60+41)*2 [as] [ae] [am]; [am] asplit [am1] [am2]; [as] asetpts=PTS-STARTPTS [as]; [am1] asetpts=PTS-STARTPTS [am1]; [am2] asetpts=PTS-STARTPTS [am2]; [ae] asetpts=PTS-STARTPTS [ae]; [as] [am1] [am2] [ae] concat=n=4:a=1:v=0 [a]' -map 0:v -map '[a]' -c:v copy DFW-1.mp4
In other words I have
asthe start until 18:41.000
aethe end from 18:42.100
amthe middle section between 18:41.000 and 18:42.100, which should be repeated once
The segmentation works, the concatenated audible part sounds good. However,
ae turns out to only be ~5 seconds short and
ffmpeg starts to print spammy "Buffer queue overflow, dropping"-lines. The entire video is 1h30 long leaving me with plenty of silence until the end.
In other words the video and audio are fine until about 18:48.000 and the remaining 1h11 of the video has no audio.
ffmpeg doing that and how can I fix it?
I also tried to include a silence audio segment instead as plan B, but failed.