This branch implements support for transforming SDR (sRGB) to a HDR enabled output, while allowing HDR content to stay intact.
Roughly what it does is, when HDR mode is enabled via the experimental property, all the stage views that represents outputs where HDR actually managed to be enabled starts to composite in a linear variant of the target color space (which will be BT.2020) via an intermediate framebuffer. Each thing being painted should then transform its pixels to fit in this color space, using an EOTF, a color space mapping matrix and luminance adaptation.
The upcoming gnome releases are going to be sick
I’d never expected to see HDR support on Linux so soon. At the time I bought my current monitor I specifically didn’t care for HDR support since it was unlikely to work on Linux for the foreseeable future. This was before the Steam Deck was announced.
Tbh, this is just one of the many things that will lead to stable HDR support. But its nice to see these get solved one step at a time.
This is the best summary I could come up with:
Adding to the growing list of changes that is making September’s GNOME 47 desktop release quite a delight, the Mutter compositor has merged another great feature.
The work by Jonas Ådahl to implement color state transforms has now been merged for Mutter 47.
This work is abot supporting the transforming of SDR/sRGB content to an HDR enabled output.
Ådahl explains in the seven month old merge request: "This branch implements support for transforming SDR (sRGB) to a HDR enabled output, while allowing HDR content to stay intact.
Roughly what it does is, when HDR mode is enabled via the experimental property, all the stage views that represents outputs where HDR actually managed to be enabled starts to composite in a linear variant of the target color space (which will be BT.2020) via an intermediate framebuffer.
GNOME 47 is due out 18 September with many HDR/display and Wayland improvements among other enhancements that we have been covering on Phoronix the past several months.
The original article contains 227 words, the summary contains 162 words. Saved 29%. I’m a bot and I’m open source!