radeon: different colors on different heads?

[Alex Deucher]

On Wed, Nov 11, 2009 at 9:29 PM, Csillag Kristof
<csillag.kristof at united-consult.hu> wrote:
> Alex Deucher wrote:
>> Does adjusting the gamma on the problematic head help?
>
> If I understand your question correctly, you suggest that the gamma
> correction function might be off.
>
> I don't think this is the case, because on my monitor, the relation
> between (at least) the blue part of the pixel data and the actually
> displayed color seems to be a non-monotonic function.
>
> See the photo I made about the color selection dialog from GIMP.
> (Before taking the picture, I stretched it out horizontally for best
> visibility.)
>
> On proper head:
> http://picasaweb.google.com/lh/photo/OB7B3DgO3LKLmL4NUxx5Jw?authkey=Gv1sRgCM60oc7kuoW24gE&feat=directlink
>
> On bogus head:
> http://picasaweb.google.com/lh/photo/ZRIy7uJmmPmVWxEcSmVTKg?authkey=Gv1sRgCM60oc7kuoW24gE&feat=directlink
>
> The current setting is 0000FF (RGB), and as you can see, the gradient
> from 000000 to 0000FF (the 6th line, besides letter B), and
> the one from 0000FF to 00FFFF (the 5th line, besides letter G)
> are messed up in creative ways. (They should be monotonic, but they are
> not.) As far as I can tell, wrong gamma does not do that; it does not
> introduce non-monotonic transformations.
>
> It looks almost as if some bits of the blue data were sometime lost in
> transition. (But where?)

Is the problematic screen the one driven by DVO or internal tmds?
Post your Xorg log and specify which output is the problematic one
(DVI-0 or DVI-1).

Alex