A solution for gamma-adjustment support in Wayland

[Chris Murphy]

On Tue, Dec 27, 2016 at 11:35 AM, Mattias Andrée <maandree at kth.se> wrote:

>
> > get away with 8 bit per channel CIE L*a*b* but even
> > that's questionable these days.
>
> I would reject CIE L*a*b* because it requires more
> expensive conversion and more than two conversion
> steps.
>

The conversion is lookup and interpolation in any case - the expense is at
the front end when the display profile is built. There are XYZ and Lab
display profiles, and in practice table based ICC display profile will be
CIE L*a*b* based, and matrix+TRC display profile will be XYZ based.

The general idea is that a very well behaved display can be both accurately
and precisely described with a matrix+TRC profile. And that leads to having
to evaluate whether more measurements in a table based display profile
provides better accuracy while not excessively causing discontinuities
(less smooth, noticeable posteriziation); or if the display is such crap
that you just have to suck it up and go with matrix+TRC to get smooth
results that lacks uniform accuracy.



-- 
Chris Murphy
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.freedesktop.org/archives/wayland-devel/attachments/20161227/11234c82/attachment.html>