[MEncoder-users] Fixing A-V desync and -mc 0

[Vladimir Mosgalin]

Hi D Richard Felker III!

 On 2005.02.09 at 17:35:30 -0500, D Richard Felker III wrote next:

> > I already asked this in mplayer-users, but got no answer that time...
> > Since it's more about mencoder, maybe someone here knows the solution?
> > 
> > I have several files with A-V sync problem. On playback, ct: value in
> > status line constantly grows, and A-V sync goes off. -mc 0 solves this.
> > 
> > But remuxing file with mencoder -oac copy -ovc copy -mc 0 doesn't fix
> > the problem, the resulting avi behaves exactly the same. I need to make
> > correct sync permanent. Is there a way?
> 
> my guess is that one of the headers is radically incorrect, probably
> something related to samplerate. the best way i know to fix the file
> is to figure out what's wrong and hexedit the header. reencoding audio
> might also solve it, at the expense of quality. and if the audio is
> just mp3, you could try dumping it with -dumpaudio and then remuxing
> from the external .mp3 file using -audiofile.

Well. Let's split cases then. I created one of the file myself; I
changed video fps and audio sample rate. I had to do some external
processing of audio track and change video fps (it was incorrect for
some reason), so I splitted audio track off, did some stuff to it,
including resampling, and tried mux it again to video with new fps.
Of course, the length of audio and video tracks is the same. But.. this
is what I get. The file looks correct - 15 fps, 48khz audio, both are
encoded by mencoder...

The other file is has 23,976 fps and 44100 khz audio; all headers look
correct. I don't know what happend to it, but it behaves the same.


If there are separate audio and video tracks, both look normal, how come
that A-V desyncs after muxing? Where that broken sync information comes
from? And the original question - if -mc 0 fixes this, why can't it
permanently fix it?

-- 

Vladimir