ODT B-spline in python and Interactive Plot

Tags: #<Tag:0x00007f3c5da2ccb8> #<Tag:0x00007f3c5da2cb50> #<Tag:0x00007f3c5da2ca10>

(Nick Shaw) #21

I may be wrong, but I seem to remember from commenting them out when testing that the effect of the glow module, red modifier and global desaturation are relatively subtle, and also contribute to problems with invertibility. Since they all run consecutively (with the exception of the AP0 to AP1 matrix happening before the global desaturation, but that can be accounted for) and before any tone curves, could they not be moved to a “default LMT”, to be applied or not as desired, leaving the Output Transform cleaner still? Presumably LMTs can also be made which compensate for the difference between the tone curve in the new Output transform and the combined effect of the current RRT and ODTs, so they can be used by those who need an exact match to the 1.0.3 look.

Also I am sure it was mentioned in the working group meeting, but I forget. How is it envisaged that the dim surround and saturation compensations will be incorporated in the unified Output Transform?

(Alex Forsythe) #22

My 2¢ … the glow and global desat are relatively subtle. The red modifier is more significant, but isn’t invertible.

That’s been proposed …

We’d have to test it, but presumably

This is an open question. The dim surround compensation built into some of the current ODTs is done with a gamma function. This works well for SDR, but probably won’t work for HDR. We need to come up with a new way to do this. Color Appearance Models?