[Mplayer-cvslog] CVS: main/DOCS/tech mpcf.txt,1.30,1.31

[Michael Niedermayer]

Hi

On Sunday 07 September 2003 15:00, Alex Beregszaszi wrote:
> Hi,
>
> > > As shuffle_type is audio only (or am I false?),
> >
> > the synatx allows it for non audio streams, but its forbidden
>
> So what about moving it to audio_stream_header?
advantage: 0.1 byte / sec gain 
disadvatage: reduced generality (we cant cleanly add shuffling to other stream 
types if we want in the future), work to change spec & implementation
IMHO the advantage is not big enough 

> Or have you plans on
> supporting this with video packets too?
not really

> Btw, what's the sense of scrambling the audio data?
error robustness, when damaged packets are mixed with undamged ones then its 
easier to conceal the errors

>
> > btw, IMHO the repeat headers before keyframes rule, should be removed,
> > it could easily mean +100mbyte/hour in some cases (for example some
> > codec which has global headers and are intra only
> > (huffyuv,vorbis,...))
>
> Imho the rule has to be modified to be precise.
yes

> Repeating the headers every 10sec can cause that headers aren't placed
> before keyframes. So if you have a damaged file, and search for the
> headers, eventually you'll skip a lot of non-keyframes in search for a
> keyframe.

possible rules:
the first frame after the headers SHOULD (MUST?) be a keyframe
there MUST be headers before the first frame

-- 
Michael
level[i]= get_vlc(); i+=get_vlc();		(violates patent EP0266049)
median(mv[y-1][x], mv[y][x-1], mv[y+1][x+1]);	(violates patent #5,905,535)
buf[i]= qp - buf[i-1];				(violates patent #?)
for more examples, see http://mplayerhq.hu/~michael/patent.html
stop it, see http://petition.eurolinux.org & http://petition.ffii.org/eubsa/en