[FFmpeg-devel] [RFC][PATCH] Move H.264 intra prediction functions into their own context

Kostya kostya.shishkov
Wed Sep 5 09:31:39 CEST 2007


On Wed, Sep 05, 2007 at 09:16:11AM +0200, Reimar D?ffinger wrote:
> Hello,
> On Wed, Sep 05, 2007 at 02:09:41AM +0200, Diego Biurrun wrote:
> > On Wed, Sep 05, 2007 at 12:22:14AM +0200, Diego Biurrun wrote:
> > > On Tue, Sep 04, 2007 at 11:39:34AM -0700, Mike Melanson wrote:
> > > > Aurelien Jacobs wrote:
> > > > > As much as I appreciate splitting, I don't want it to delay RV40 at all !
> > > > > Anyway, I will probably have a look at this issue during my next
> > > > > splitting round.
> > > > > So if splitting is not trivial, don't care about it, and concentrate on
> > > > > RV40 support instead ;-)
> > > > 
> > > > This might be this first I have heard of someone actually *caring* about
> > > > the coming RV40 support. :) Personally, I only care about it for the
> > > > sake of completeness.
> > > 
> > > Huh?
> > > 
> > > This is the greatest achievement in codec support ever since we got
> > > working H.264 ...
> > > 
> > > Maybe your world is all flash nowadays, but lots of stuff gets streamed
> > > as RealVideo.  Of course there are binary codecs, but hey ...
> > 
> > Put another way: This largely obsoletes our binary codec packages.
> 
> Well, I lately have seen more people with problems with WMV2 due to
> J/X8-frames than people missing Real. Might of course be because there
> are supported 64 bit codecs for RealVideo (and they are less buggy than
> the 32 bit ones...)

Maybe it will be implemented too. BTW, X8-frames employ something similar to
H.264 intra prediction.  
 
> Greetings,
> Reimar D?ffinger




More information about the ffmpeg-devel mailing list