[XviD-devel] OBMC decode support patch for dev-api-4

Christoph Lampert chl at math.uni-bonn.de
Tue May 6 18:39:10 CEST 2003


On Tue, 6 May 2003, Luca Piccarreta wrote:

> Out there there are VFW codecs that support OBMC (ours for instance).
> But not in the free domain, as far as I know.
> Weirdly enough, there is no profile that supports OBMC.
> DIVX does not.... but OBMC really improves quality, even though
> implementation is not that easy (motion compensation must be
> performed one MB after MV decoding).

Do you have any numbers on that? I thought OBMC is good for low bitrate
(video telephones), but not in our bitrates. 

> But does XVID+AVI really need to be compatible with any profile?

Yes.

> Does XVID need to be compatible with DIVX?

No, not really. But on the other hand, thinking of users, having two
mainstream codecs which are 98% compatible is ugly.

The problem with OBMC would also be old XviD/DivX files: How to detect if
OBMC is really wanted or if just the bit is falsely set.

gruel


> Compliments for the excellent job.
> Luca Piccarreta (TILAB consultant)
> 
> ----- Original Message ----- 
> From: "Christoph Lampert" <chl at math.uni-bonn.de>
> To: <xvid-devel at xvid.org>
> Sent: Tuesday, May 06, 2003 9:47 AM
> Subject: Re: [XviD-devel] OBMC decode support patch for dev-api-4
> 
> 
> > On Mon, 5 May 2003, Edouard Gomez wrote:
> >
> > > Christoph Lampert (chl at math.uni-bonn.de) wrote:
> > > > Hm, are you sure?
> > > >
> > > > OBMC is in the ISO Standard, but it isn't part of any MPEG-4 profile,
> so
> > > > using OBMC breaks MPEG-4 compatability.  There are no codecs outside
> that
> > > > use it, as far as I know, it could be removed from MPEG-4 at any time.
> > > >
> > > > Why add a feature that won't be used for decoding and shouldn't be use
> for
> > > > encoding?
> > >
> > > Ok, let's  add this patch to  my never ever growing  "don't comit" patch
> > > queue.
> >
> > I didn't say that. Or maybe I did? Anyway, we should find a policy for
> > MPEG-4 stuff that isn't in any profile supported by us. At the moment,
> > that's at least
> >
> > * reduced resolution
> > * OBMC
> >
> > What do you think? Should we include those and add a "compatability
> > mode" so they aren't used for encoding. Or just throw them out, because we
> > are not compliant and will just bloat up code?
> >
> > How about OBMC/RRn just for decoding? I won't be needed, but it wouldn't
> > hurt, either...
> >
> > gruel
> >
> >
> > _______________________________________________
> > XviD-devel mailing list
> > XviD-devel at xvid.org
> > http://list.xvid.org/mailman/listinfo/xvid-devel
> 
> _______________________________________________
> XviD-devel mailing list
> XviD-devel at xvid.org
> http://list.xvid.org/mailman/listinfo/xvid-devel
> 



More information about the XviD-devel mailing list