On Wed, 21 Dec 2016 21:05:13 +0100 Kai-Uwe <[email protected]> wrote: > Hello Mattias, > > Am 21.12.2016 um 12:31 schrieb Mattias Andrée: > > If there is no API for these thinks, functionality > > such as inverting the colours, lowering the temperature, > > and (requires matrices) simulating defective colour > > vision must be done in the compositor. This is not > > a sustainable solution, an API is necessary. It's bad > > enough that colour management is done by the > > compositor. > > The compositor is a good place for doing color > correction, as X11 and Wayland core developers had > repeatedly suggested. I mentioned to you on the XDG list, > that your API will be bypassed by any proper ICC based > CMS and both might start fighting about > resources/settings.
It will not. Each filter is given a priority that determines the order the filters are applied. An ICC based colour correction filter is given the priority 0. If the correction is uses two filters, one for gamma and one for brightness and contrast, the gamma has the priority 0 and the brightness and contrast has the priority -0x4000 0000 0000 000. Filters with higher priority are applied before filters with lower priority. So their is a system in place to ensure that it does not happen. > That's no good I think. A ICC monitor > profile has to set a calibration curve into the display > pipeline. It does currently on X11 systems and Weston has > some code to do as well CLUT setup. Without knowing each > other that's a bug filler, which no one want (at least I > hope). > > For getting more specialised support into the compositor, > the suggested surface device links for each output are > very expressive [as discussed by Niels+Graeme+me in the > color management thread]. A CMS can work with device > links and add features without involving the application. > So it can add device links on its discretion to surfaces, > which do not have them already and support new features > like the redshift curves or more advanced effects. That > would be a big benefit for both communities I guess. For > that to work seamless your curves, need to be abstracted > and instead of curves are expressed as ICC profiles of > type abstract. So a CMS can integrate them easily, > without to know the bells and whistles redshift offers > its users. > > Kai-Uwe
pgpjHzHKy5KOJ.pgp
Description: OpenPGP digital signature
_______________________________________________ wayland-devel mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/wayland-devel
