search

w3c / ColorWeb-CG

repo for the Color on the Web Community Group
47 stars 21 forks source link

Should guidance on gamut mapping be provided? #100

Closed palemieux closed 8 months ago

palemieux commented 1 year ago

The strawman provides examples of algorithms designed to render high dynamic range images to a lower dynamic range display -- the focus of the strawman is on high-dynamic range imagery.

These example algorithms do not tackle the mapping of wide gamut images to a display with a narrower gamut, e.g. from Rec, 2020 to Rec. 709 color gamut.

Should such examples be provided?

See also https://gitlab.freedesktop.org/pq/color-and-hdr/-/issues/18 .

Originally posted by @ppaalanen in https://github.com/w3c/ColorWeb-CG/issues/98#issuecomment-1571844917

ppaalanen commented 1 year ago

I'm afraid there are two unrelated issues being confused here.

https://gitlab.freedesktop.org/pq/color-and-hdr/-/issues/18 and https://github.com/w3c/ColorWeb-CG/pull/98#issuecomment-1571844917 are not about the gamut mapping that I asked about in https://github.com/w3c/ColorWeb-CG/pull/98#discussion_r1213046602 .

The question in Weston issue 18 is how does the display mastering information relate to the signalling color volume.

The algorithm in the W3C document with rec2100PQtoSRGB() seems to be what you are talking about in this issue, and the question with that is what should the reader expect from a function that is named rec2100PQtoSRGB()? As it is, it does tone mapping, but not gamut mapping. To me it is a surprise it does one but not the other, when the name has two standard color spaces in it. That could be simply pointed out in the text.

I believe that we should not assume that a mastering display uses the same tone and gamut mapping algorithms as regular viewing. How are mastering displays driven exactly is something I'd like to find out, but that's not the topic of this issue.

ppaalanen commented 1 year ago

Oh, maybe you meant the third question: how are display mastering parameters used to drive the gamut mapping from signal space to actual monitor space?

That's a good question too, and I expect with no unique answer.

One more question is, are the display mastering parameters a physical description of the mastering display, or have they already been converted into the signal space in production? If so, that would make the question of how a mastering display is driven moot.

palemieux commented 1 year ago

are the display mastering parameters a physical description of the mastering display,

It is the description of the absolute (XYZ) color volume that the mastering display can physically output, i.e. colors outside that volume could not have been monitored.