Hi! > >> I have been thinking about the last Multicolor Framework proposal by > >> Dan. > > > > We talked on the phone... and I promissed some links. > > > > I'd really prefer drivers to present output power to the of the > > kernel/userland, and deal with the RGB conversion else where. So if > > driver was asked for (20, 50, 100), red output would go to (20/1000) W, > > green (50/1000) W, ... > > > > RGB is described here: > > > > https://en.wikipedia.org/wiki/RGB_color_model (see section > > "nonlinearity") > > > > https://en.wikipedia.org/wiki/Gamma_correction > > > > And concrete values are here: > > > > https://en.wikipedia.org/wiki/SRGB#Specification_of_the_transformation > > > > While playing with colors, I came up with this: > > > > https://github.com/pavelmachek/unicsy_demo/blob/master/monitor/notint.py > > > > See change_color for exact transformation (it could be improvede, > > linear part is not there). It can be also used to display same color > > on monitor and LED. It get somehow reasonable results for me... > > > > Does this use the multi color framework interfaces defined here? > > https://lore.kernel.org/patchwork/patch/1060683/ > > Minus the sync, sync_enable and brightness_model interfaces. It does not look like that after looking > at the notint script. > > Or is this your own multicolor framework implementation? > > We could meld these ideas together. This uses legacy kernel interfaces. It is meant to demonstrate that arbitrary color setting is feasible. Pavel -- (english) http://www.livejournal.com/~pavelmachek (cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html
Attachment:
signature.asc
Description: Digital signature
