[Mplayer-cvslog] CVS: main/vidix/drivers mach64_vid.c,1.8,1.9

[Michael Niedermayer]

Hi

On Wednesday 20 February 2002 20:24, Nick Kurshev wrote:
[...]
> >  // Bits 16-19 seem to select the format
> > -// Bit 28 seems to toggle something affects luma & chroma
> > +// 0x0  dunno behaves strange
> > +// 0x1  dunno behaves strange
> > +// 0x2  dunno behaves strange
> > +// 0x3  BGR15
> > +// 0x4  BGR16
> > +// 0x5  BGR16 (hmm, that need investigation, 2 BGR16 formats, i guess 1
> > will have only 5bits for green) +// 0x6  BGR32
> > +// 0x7  BGR32 with somehow mixed even / odd pixels ?
> > +// 0x8	YYYYUVUV
>
> mga g200 compatibility? ;)
dunno

>
> > +// 0x9	YVU9
>
> I've amazed !!!
>
> > +// 0xA	YV12
> > +// 0xB	YUY2
> > +// 0xC	UYVY
> > +// 0xD  UYVY (not again ... dont ask me, i dunno the difference)
>
> Maybe radeon.h can help you? There is still unsupported (untested by me)
> YVYU!
not really :(
0xC==0xD appearently there is no visible difference
and the remaining modes are not BGR or YUV like (setting a single bit affects 
several pixels ... and such weird things) so it seems that 0x0-0x2 & 0xE-0xF 
are not legal values at all, at least not on my card

>
> > +// 0xE  dunno behaves strange
> > +// 0xF  dunno behaves strange
> > +// Bit 28 all values are assumed to be 7 bit with chroma=64 for black
> > (tested with YV12 & YUY2) // the remaining bits seem to have no effect
[...]

Michael