On Fri, 29 Mar 2002, Marc Aurele La France wrote:
> On Fri, 29 Mar 2002, Marc Aurele La France wrote:
>
> > > > > While you are at it, you might wish to change the default colo[u]rkey
> > > > > from 0. Black is a common enough text colour; with Keith's patch
> > > > > highlighted text shows the video layer beneath :-)
>
> > > > I stole a few lines from the R128 driver and stuck them into the radeon
> > > > driver. The video key is now configurable (option "videokey" "42") and
> > > > has a non-zero default value (30).
>
> > > > New diff is at:
>
> > > > http://keithp.com/~keithp/download/radeon_xv-2002-03-29.diff
>
> > > > I note that the config file uses the phrase "videokey" instead of
> > > > "colo[u]rkey". That conveniently avoids the spelling ambiguity...
>
> > > I just want to point out that GATOS code already supports this.
> > > As for 8bbp - I will insert the code to disable hardware Xv support in
> > > this case and print a meaningful message.
>
> > Just my 2 cents' worth. I don't think this is a good idea. It's not
> > necessary to do so in the Mach64 and R128 cases.
>
> > > As for Radeon colorkeying the reason we use RGB values is because Radeon
> > > no longer supports keying on the value of graphics pixel, rather it allows
> > > you to specify ranges of alpha and R, G, B so all graphics pixels that
> > > fall within them are replaced by video. The comparison is done _after_
> > > Radeon has converted framebuffer value into internal 32bpp format.
>
> > Oh? An alpha channel for 8bpp? If so, where does it come from? The LUT
> > is only 3 * 768 long. Perhaps a specific alpha value can be used as the
> > key.
>
> Ooops. Synapse cross. I meant 3 * 256.
Ehh.. What is LUT ? The way I read in the docs is that there are 32 bit
range registers which have 8 bits for Alpha, R, G and B. And if you set
Alpha to a constant value (i.e. bound it with, say, 8 both from above and
below) nothing will display. The documentation says that in this case
the lower range for alpha should be set to 0 and the upper range to 0xff -
which makes sense. So it is compared with something in the hardware - but
with what we don't really know. (And it is possible that for 8bpp
framebuffer the alpha of the result is always zero..)
Vladimir Dergachev
>
> Marc.
>
> +----------------------------------+-----------------------------------+
> | Marc Aurele La France | work: 1-780-492-9310 |
> | Computing and Network Services | fax: 1-780-492-1729 |
> | 352 General Services Building | email: [EMAIL PROTECTED] |
> | University of Alberta +-----------------------------------+
> | Edmonton, Alberta | |
> | T6G 2H1 | Standard disclaimers apply |
> | CANADA | |
> +----------------------------------+-----------------------------------+
> XFree86 Core Team member. ATI driver and X server internals.
>
> _______________________________________________
> Xpert mailing list
> [EMAIL PROTECTED]
> http://XFree86.Org/mailman/listinfo/xpert
>
_______________________________________________
Xpert mailing list
[EMAIL PROTECTED]
http://XFree86.Org/mailman/listinfo/xpert
