2025-03-31 TSC Meeting notes

Merge the PR to fix the website

• https://github.com/AcademySoftwareFoundation/OpenColorIO/pull/2113

• merging, if it breaks we’ll fix it!

ACES 2 gamut/range limits

• New outputs have a range limit and a gamut limit operation

• Noticed via scopes that there are values that go a little outside the limits

• What is “a bit out”? Within floating point error? More than a few percents would be anomalous

• Kevin: clamp is not necessarily applied in the encoding output space

• Tonescale is implicitly not clamped (rolled off slightly above the limit so you don’t get the linear slope at the top)

• Display output should be clamped - we should check this, Kevin thought it was

• Doug will post some results on the slack channel and we can continue the discussion

What colour space name to store in OpenEXR files?

• proposing a getGenericName() function

• https://docs.google.com/presentation/d/1sY55UHYi2AJF6EPYDYY-mvaJUCzZCkWisrj-a9yhg2w/edit#slide=id.g337d7ca5679_0_24

• In most EXR interop cases, these color spaces are relatively generic/standardized, even if their names are different

• Take a config and a color space name, compare it to the built-in configs name, and if it finds a match, put that in the metadata in the config vs. the specific config name

• Kevin: two times we create files - internal and external

• Important to be able to override and not convert for internal use

• Carol: could we make it so that writing the generic name is the default vs the other way around

• Kevin: but the folks modifying the configs are the most likely to want their internal names to persist. The folks not “paying attention” would most likely get the default names anyway

• Kevin: there should be an attribute in the “write” nodes application where you choose to write

• Mark: would want a field in the config color space definition to add the canonical name vs. having an auto-comparison feature

• Sean: also wary of an auto-detect feature, wants a config author to make these associations

• Will make an issue to track discussion and proposals

Adding CICP and other metadata to display color spaces

• Add a general-purpose key/value metadata attribute?

• https://github.com/AcademySoftwareFoundation/OpenColorIO/issues/1975

• CICP, Enums for display color in the OS / ICC, gamma atom

• Mark: like the CICP example - attach to display vs. colorspace. This would be a really great place to start, as it’s highly used and relatively standardized

• Kevin: CICP not sufficient by itself to write quicktimes

• Doug: if there’s a key/value system, the OCIO API could read those values from the list vs. explicit values

• Mark: I think we should explicitly define this in OCIO. getCICPPrimaries(), etc etc

• We have to decide how much we want to validate/support - is it just you hand us an integer and we pass it along? Group: yes

• Group consensus that the proposal in the issue looks good

• We should separate input from output here - this proposal is targeting output - whether for a viewer or otherwise. We’ll handle input separately, most likely via file rules

Next week's working group slot

• Continuation of CICP/ICC profile discussion

• Addition of Gamma 2.2 display to studio config