[XviD-devel] [BUG]

[Michael Militzer]

Hi,

> Michael Militzer (michael@xvid.org) wrote:
>
> > one of the assembler functions. - Have you checked if the frames still
> > become  pink when  you use  the c-version  of the  interpolation code?
> > Which optimizations did you normally use? (MMX, XMM or 3DNOW?)

> I tested  with the C-version  and it's the  same. I would  like people
> checking this with windows systems because perhaps it has something to
> do with  my transcode code (but i  doubt, that's not the  first time i
> modify the export_xvidcvs).

Next questions: Is this pink clearly visible or is it just a hardly
noticeably pinkish shade or something like that? Do you have the pink
problem only with some special clips or do all transcode encoded clips
become pink?

>From my side I can definitively say that there's no clear pink in a fixed
quant encoded test clip using halfpel intgerpolation on windows. I haven't
analysed the clip frame by frame now, so if it's just a slight pink I could
have missed it, but I doubt it. - Which decoder did you use btw?

bye
Michael