On Fri, 29 Mar 2002, Keith Packard wrote:
>
> Around 18 o'clock on Mar 29, Vladimir Dergachev wrote:
>
> > As for Radeon colorkeying the reason we use RGB values is because Radeon
> > no longer supports keying on the value of graphics pixel,
>
> Well, I just experimentally determined that (at least on my M6 LY), this
> is not true. For the component video modes (DirectColor/TrueColor), it
> expects you to decompose the RGB pre-LUT values and store them in the
> "color key" registers, but for pseudo color modes, it wants the pixel
> value replicated across all three bytes stored in the "color key"
> registers. So much for believing the manufacturers documentation.
Thanks, I'll update GATOS CVS. This is not the first time there is
discrepancy in the docs..
Vladimir Dergachev
>
> I tested this at 16bpp by using xgamma to frob the LUT -- pixel values 0
> and 0xffff always worked for "color key", but all other pixel values
> worked only with a gamma of 1.0. Switching to pre-LUT values made it work
> correctly with an arbitrary gamma value.
>
> (The gamma value has no effect on the brightness of the overlay, the
> overlay has separate controls to adjust the brightness curve).
>
> I tested this at 8bpp to note that the post-LUT values didn't work, so I
> tried replicating the pixel value across all three components of the
> "color key" registers; that works just fine. (I didn't try using just the
> pixel value, I have a feeling that won't work though).
>
> Video plays just fine now at 8bpp, but the rest of my UI is really ugly.
>
> Diffs at
>
> http://keithp.com/~keithp/download/radeon_xv-2002-03-29-19.diff
>
> Keith Packard XFree86 Core Team Compaq Cambridge Research Lab
>
>
_______________________________________________
Xpert mailing list
[EMAIL PROTECTED]
http://XFree86.Org/mailman/listinfo/xpert
