On 06/27, Ville Syrjälä wrote: > On Sun, Jun 19, 2022 at 09:31:03PM -0100, Melissa Wen wrote: > > Add 3D LUT for gammar correction using a 3D lookup table. The position > > in the color correction pipeline where 3D LUT is applied depends on hw > > design, being after CTM or gamma. If just after CTM, a shaper lut must > > be set to shape the content for a non-linear space. That details should > > be handled by the driver according to its color capabilities. > > I also cooked up a WIP 3D LUT support some time ago for Intel hw: > https://github.com/vsyrjala/linux/commits/3dlut > But that dried up due to having no userspace for it. > > I also cooked up some basic igts for it: > https://patchwork.freedesktop.org/series/90165/ Yes, I found your work on it, so I based part of my proposal on your previous proposal and also the cubic-LUT by Laurent for rcar-du [1]. They helped me to find a path to expose 3D LUT caps, thanks. > > <snip> > > + * “LUT3D”: > > + * Blob property to set the 3D LUT mapping pixel data after the color > > + * transformation matrix and before gamma 1D lut correction. > > On Intel hw the 3DLUT is after the gamma LUT in the pipeline, which is > where I placed it in my branch. > > There is now some discussion happening about exposing some > kind of color pipeline description/configuration properties: > https://gitlab.freedesktop.org/pq/color-and-hdr/-/issues/11 So, initially I thought we would just map the properties according to hw pipeline, but in fact it isn't a good path to follow and would make a big mess, so Sebastian's proposal for a pipeline description makes a lot of sense to me. I'll join the discussion. Anyway, thanks for all inputs, Melissa [1] https://lore.kernel.org/all/20201221015730.28333-1-laurent.pinchart+renesas@xxxxxxxxxxxxxxxx/ > > -- > Ville Syrjälä > Intel
Attachment:
signature.asc
Description: PGP signature
