[mythtv] Keyframe adjust table (posibel bug(s))

Torbjörn Jansson torbjorn.jansson at mbox200.swipnet.se
Mon Jan 5 13:23:04 EST 2004


> On Mon, 05 Jan 2004 11:14:36 -0600, "Geoffrey Hausheer" said:
> > 
> > On Mon, 5 Jan 2004 17:23:20 +0100, "Torbjörn Jansson" said:
> > > It will not be as easy as if first thougt, because if I 
> can't use the
> > > timecode it will be hard to keep the audio and video in sync.
> > > 
> > That is why I said you need to use both.
> > 
> > The timestamps are correct in a relative sense (you can 
> rely on them to
> > match for a given audio/video pair).  The problem is that 
> you can't rely
> > on a timestamp that is marked as 5:30.00 to actually occur 
> 5 minutes in. 
> > It could actually occur 2 minutes into the recording.
> > 
> > so when you do a seek, you should do so using frame 
> numbers, then read
> > the first timestamps and reinitialize your A/V sync based on that.
> > 
> Let me add that this is exactly how myth does it today (and why the
> videos play fine in myth)
> 

That sounds like a good way to solve it. That will fix the the problem when
seeking to a new position.
But I will still have problem when I encounter a position where the
timecodes make a jump.

Maybe I can adjust the timestamping when a sync frame is encounterd, that
woud probably solve all the problems.
After a file is transcoded all cuts will be keyframes and have sync headers
before, right?



More information about the mythtv-dev mailing list