[Openicc] XICC specification draft (Xinerama vs. composite).

Kai-Uwe Behrmann ku.b at gmx.de
Tue Jun 28 08:28:23 EST 2005


Am 27.06.05, 15:46 -0600 schrieb Keith Packard:

> On Mon, 2005-06-27 at 23:09 +0200, Kai-Uwe Behrmann wrote:
> > Exporting the screen geometries to applications would allow them to use 
> > whatever conversion they like and could work more soon.
> 
> In any multi-head configuration, applications don't generally know what
> screen their window will be presented on, so some default mechanism for
> doing color correction seems required.

And this information can be obtained somehow from the appropriate API?
 
> > The stage to convert colours with plugable CMMs controlled by the 
> > composite manager and hide this stuff from applications will need a bit of 
> > preparation. This is in my opinion the way for the long run.
> 
> I'm already starting to figure out how to have applications cooperate in
> the compositing process. This is currently being driven by needs for
> accessibility, both color adjustment and magnification.  Perhaps this
> kind of mechanism could be used for more accurate color correction.

The good news would be CLUTs support for CG shader colour conversions. 
A CMM usually provides a way to store colour transformations in a CLUT.
This is a possible input with ICC convention.

Uploading a shader programm to the compositor would be like a small 
plugable CMM. Of course current CPU deploying CMMs would not fit into.

FYI,
Colour adjustments where discussed on this list as a special kind of ICC 
profiles called abstract profiles. Such profiles contain CLUTs and curves 
to transform colours within the CIE*Lab colour space. Such profiles can 
be used very flexible.

Possibly bad news for very sophisticated applications is, they are forced 
to bring theyre data first from n-channels to a 3 channel colour space and 
then leave the remaining conversion to the a second composite internal 
conversion. Maybe with the higher precission of shaders this becomes a non 
issue. But I am not shure about CG lattice math.

But these ideas may go far beyond the current intention regarding support 
for accessibility.

regards
Kai-Uwe Behrmann
                                + development for color management 
                                + imaging / panoramas
                                + email: ku.b at gmx.de
                                + http://www.behrmann.name




More information about the openicc mailing list