RE: General Xvidcore questions
"Harsha Viswanath" <[email protected]> Tue, 16 Mar 2004 17:37:47 +0530
| Newsgroups | gmane.comp.video.xvid.user |
|---|---|
| Message-ID | <[email protected]> |
Michael, Thank you very much for your answers, this cleared some of my doubts. Thanks, -Harsha > -----Original Message----- > From: Michael Militzer [mailto:[email protected]] > Sent: Monday, March 15, 2004 9:00 PM > Subject: Re: [XviD-users] General Xvidcore questions > > Hi, > > Quoting Harsha Viswanath <[email protected]>: > > > I would appreciate answers for a couple of general questions - > > > > Can the Xvidcore code be used as a reference decoder for the profiles > > and levels that it supports? What is the level of testing that it has > > undergone? > > I find this implementation a much cleaner implementation compared to > > Momusys. > > yes, the code is more easily understandable and we're proud about this ;-) > Nonetheless, xvidcore is not meant to replace a reference implementation, > so I wouldn't recommend such a use (even though I know people do so). The > decoder has been tested with some of the MPEG conformance bitstreams, > however there is no *full* compliance to any profile - more below: > > > Right now Xvidcore does not support ASP profile right? If no, then when > > can we expect support for this profile. > > The latest versions of xvidcore (1.0 RC3 source code or dev-api-4 branch > on xvid cvs) support the advanced simple profile - at least the most > common > features. This means: Quarterpixel MC, BVOPs and Global MC are supported, > but the following is missing for full compliance: data partitioning, > reversible vlc and field prediction. So you see, XviD can't serve as a > complete replacement for a reference implementation. > > Hope this helps! > > regards, > Michael > > > _______________________________________________ > XviD-users mailing list > [email protected] > http://list.xvid.org/mailman/listinfo/xvid-users