[XviD-devel] Changes in Source <-> Xvid 1.0

Christoph Lampert chl at math.uni-bonn.de
Sat Nov 29 10:49:01 CET 2003


On Fri, 28 Nov 2003, Edouard Gomez wrote:

> I think you minimize too much  the matrix quant problem. I showed how it
> can  impact both  image quality  and the  stream syntax.  It's  not only
> theory, it happened  to me using transcode, it happened  to a doom9 user
> too.

Good! Honestly, that was the answer I had hoped for. So let's agree that 
MPEG-quant was a severe bug as well. Since you cannot disable the ability
of XViD to be run in two sessions at the same time, it had to be
fixed. Thank you that you spending the time!

Michael's point is not void by this: You replaced crucial code which was
well working for many people (single-thread users?) by new code, which has
to be tested thoroughly. PSNR difference for a one sequence is of
course not enough for this (I know, you didn't say it was.). Let's hope
that by the Beta we will find any possible problems, before they turn up
in a new codec comparisson. 

gruel




More information about the XviD-devel mailing list