A solution for gamma-adjustment support in Wayland

[Graeme Gill]

Mattias Andrée wrote:

Hi,

> Ideally I think the display server should support sRGB,
> CIE XYZ, and the monitors' native RGB for input. sRGB is
> useful for programs such as image viewers and video players.

Hmm. It's not a typical approach for a color managed workflow
to have hard wired colorspaces, but instead to allow them
to be configured, typically by using an ICC profile
to specify them.

> Native RGB is useful for utilising the full gamut,

Right, but the reaction so far is that Wayland can't/won't
allow this in any useful way, because the client isn't
allowed to know which output a native pixel will end up
on.

> and
> CIE XYZ is useful because almost all colour models can be
> converted to CIE XYZ in one step, it is device independent
> and can represent any visible colour.

Right, and using very wide gamut spaces like XYZ can also have
drawbacks, namely introducing more processing steps that reduce
fidelity (i.e. add inaccuracy and banding), as well
as potentially loosing control over gamut mapping
and clipping. i.e. they have their place, but they are not
a panacea.

> But of only supporting
> one colour space I think the monitors' native RGB should be
> used and that it is up to the application to find out one
> which monitor each part of the application's windows are
> visible and which colour space each monitor uses.

That's certainly the traditional way of doing it,
but some of the features of Wayland evidently make
such an approach less desirable.

> But supporting matrices is useful for filters such as
> simulating defective colour vision and changing colours
> to make it easier for people with defective colour vision
> to distinguish them.

Sure, and I'm reasonably certain that there are other
mechanisms available too, if a color managed workflow was
available, by suitable construction of input and/or
output device profiles. (This is an area that I'm
currently doing a little research into.)

Cheers,

Graeme Gill.