--- Begin Message ---
On Tue, 2003-09-30 at 02:48, Vincent Lefevre wrote:
> On 2003-09-29 13:06:59 +0200, Michel Dänzer wrote:
>
> > > This is strange because with xgamma, only the image (displayed by
> > > Mozilla) seems to be affected, whereas with the -gamma option of
> > > XFree86, the whole screen was affected (but the results on the image
> > > were similar to what xgamma gives).
> >
> > Weird, they should produce the same result given the same value,
> > otherwise it's probably a bug.
>
> This may be a feature: pixels that have a known gamma values are
> not affected. But this shows that the gamma correction is not a
> hardware one, thus giving bad results.
This does indeed sound like software, as the hardware affects the whole
screen. The X gamma correction uses the hardware though.
> In other words, here what seems to happen: Each RGB color is indexed by
> 256 values. Instead of doing a continuous gamma correction on these 256
> values (as output of the graphic card), it seems that the X server does
> a mapping 256 values -> 256 values for some pixels, i.e. as if the
> pixel values were different.
Right. The mapping is calculated using gamma curves with xgamma and the
X server -gamma option, and they should both give the same result with
the same values, otherwise there's a bug somewhere.
--
Earthling Michel Dänzer \ Debian (powerpc), XFree86 and DRI developer
Software libre enthusiast \ http://svcs.affero.net/rm.php?r=daenzer
--- End Message ---