[MPlayer-dev-eng] container format

D Richard Felker III dalias at aerifal.cx
Thu Feb 6 21:42:45 CET 2003


On Thu, Feb 06, 2003 at 09:03:04PM +0100, Michael Niedermayer wrote:
> Hi
> 
> On Thursday 06 February 2003 19:59, D Richard Felker III wrote:
> > On Thu, Feb 06, 2003 at 06:06:01PM +0100, Michael Niedermayer wrote:
> > > > 2) Why does sz always have to be at least 1 character long - why cant
> > > > it just be a null?
> > >
> > > mpeg startcode emulation (if we allow 0 then there could be 0,0,1 which
> > > is a startcode prefix in mpeg) perhaps we should drop the startcode emu
> > > thing, dunno
> >
> > What's the purpose of mpeg startcode emulation? Does this somehow
> > allow a raw mpeg audio or video decoder to play our files without any
> > particular demuxing? If so, this sounds like a worthwhile benefit --
> > mp3 inside mpcf would be playable by most legacy players, plus it
> > would be seekable by mpcf-aware players. But perhaps I
> > misunderstand...
> yes, that was the idea, but there are problems
> 1. i dunno if it will work at all ..
> 2. there is a video & audio stream normally so one would confuse the playback 
> of the other very likely, but IMHO, its worth a try to see if a mp3 or 
> mpeg-es player would correctly play a stream from a mpcf file ...

Well I still think it's worthwhile. If we could put mp3 files in mpcf
wrapper (and still name them .mp3 even) and have them be playable by
"normal" mp3 players, it would be a nice boost in getting people to
adopt the format. Especially since it would facilitate seeking in
mpcf-aware players and provide a new format for tags that doesn't suck
(id3v1 is rather limited and id3v2 is horrible).

BTW, does mpcf allow the file to end with the data in a packet? If so,
this would be good, since an info-type packet with id3v1 could be put
at the end for legacy players.

Rich



More information about the MPlayer-dev-eng mailing list