[MPlayer-dev-eng] Re: [RFC] change lacv's MPEG-4 encoding options defaults

Rich Felker dalias at aerifal.cx
Fri Dec 30 02:28:18 CET 2005


On Thu, Dec 29, 2005 at 08:00:53PM +0100, Michael Niedermayer wrote:
> Hi
> 
> On Thu, Dec 29, 2005 at 07:23:21PM +0100, Alexander Strasser wrote:
> > [...]
> > > >>>The default number of b-frames for x264 is currently 0 while most
> > > >>>people use 2/3 as standard. Fair enough b-frames stored in avi aren't
> > > >>>perfect, but standalones can play the produced files, most if not all
> > > >>>software can play the produced files, so is this a real problem?
> > > >>>     
> > > >>Like I said: mencoder is broken so B frames cause a/v desync!!
> > > >
> > > >??
> > > >That's only true under some circumstances. Normally mencoder + bframes 
> > > >works very well. Additionally B-frames have been recommended for the 
> > > >doom9 codec comparison. I don't think Corey made this to make mencoder 
> > > >look bad :-)
> > > 
> > > depends on how many b-frames you use: in mpeg4 tipically only up to 2 
> > > are used,
> > > but in h264 people seem to even 5 and more.
> > > The more there are the more audio and video will desync.
> > 
> >   Also note that there probably wasn't much need for a/v-sync at
> > doom9's video codec comparison...
> 
> do i missunderstand something or wouldnt simply delaying the audio by
> (1+b_frames)/fps or so fix the whole AV-sync issue with mencoder?
> and yeah sure iam aware that b frames in avi are fundamentally wrong
> and in mpeg-ps/ts which doesnt have pts per frame either they are less wrong

IMO they're not wrong in AVI; mencoder just handles them incorrectly
with regard to 0byte frames. Situation is far worse in formats that
need pts to mux, since mencoder can only provide dts (offset by decode
delay) to the muxer.

Rich




More information about the MPlayer-dev-eng mailing list