2026-02-02 TSC Meeting notes

Update on Kevin’s work

• Kevin notes the infra on the GitHub actions CI is pretty fragile - getting lots of failures randomly

• We should get clang-format and clang-tidy going before more of this work, it will help a lot, revive Remi’s old PR

• Remi - remembers running clang-tidy, a lot of noise in the output. Are other projects using it?

• We should agree on a minimal set of options, and then add specific things

• Should we run on CI and/or pre-commit hooks??

• Wasn’t able to find big wins for build parallelization - it's 1 or 2 big files that take the time. Only real thing we can do there is to break them into multiple files

• The single header file is a good target - to split into functional pieces which might bring savings elsewhere

Display color spaces document is published

https://github.com/AcademySoftwareFoundation/ColorInterop/blob/main/Recommendations/02_DisplayColorSpaces/DisplayColorSpaces.md

PRs needing review

• Vulkan test harness (needs 1 review)

• Alpha channel should be 1 (needs 1 review)

• AMF python script update (needs two reviews)

OCIOView status / ideas

• Carol will get the repo created

• Thomas/Michael can help move stuff and get things going

• Carol proposes ocioview as a focus for next Dev Days on May 14, to help us along to target a first release with OCIO 2.6.0

• More soon!

OCIO Processor inspection

• Debug tool for OCIO Processor

• One way is to recursively go through the config

• Other way is to use the processor metadata, but that’s limited

• Next way is to create a group transform from the processors

  • Has some limits, need to add metadata to a couple places (ex: LUT1D and LUT3D)

  • Also might need changes for CDLs, etc

• Kevin - definitely had cases where this would be super useful. Especially in baking inverse LUTs

• Doug - right now formatMetadata is using when reading files, if you want to write out a new file, it uses that information.

• Example output Remi was thinking of:

  • <MatrixTransform direction=forward, fileindepth=unknown, fileoutdepth=unknown, matrix=[0.6953833103179932, 0.1406652629375458, 0.1639514267444611, 0, 0.2722287178039551, 0.6740817427635193, 0.05368951708078384, 0, -0.005525882821530104, 0.004025210160762072, 1.001500725746155, 0, 0, 0, 0, 1], offset=[0, 0, 0, 0]>

    <ROOT src="/Users/remi/ColorConfig/ocio-tcam/transforms/AP1_2_XYZ.spimtx"></ROOT>

    <MatrixTransform direction=forward, fileindepth=unknown, fileoutdepth=unknown, matrix=[1.449494910454111, -0.3159127624306129, -0.1335818131351159, 0, -0.4839271831021616, 1.333326616848397, 0.1505992071282109, 0, 0.09097355512640085, 0.05089709966893959, 0.8581292687352601, 0, 0, 0, 0, 1], offset=[0, 0, 0, 0]>

    <ROOT src="/Users/remi/ColorConfig/ocio-tcam/transforms/EGamut_2_XYZ.spimtx"></ROOT>

    <Lut1DTransform direction=inverse, fileoutdepth=32f, interpolation=linear, inputhalf=0, outputrawhalf=0, hueadjust=0, length=4096, minrgb=[-0.004631, -0.004631, -0.004631], maxrgb=[127.938, 127.938, 127.938]>

    <ROOT src="/Users/remi/ColorConfig/ocio-tcam/transforms/TLog_2_Linear.spi1d"></ROOT>

    <Lut1DTransform direction=forward, fileoutdepth=32f, interpolation=default, inputhalf=0, outputrawhalf=0, hueadjust=0, length=101, minrgb=[0, 0, 0], maxrgb=[1, 1, 1]>

    <ROOT src="/Users/remi/ColorConfig/ocio-tcam/transforms/TLog_2_sRGB_TCAMv2.cub"></ROOT>

    <Lut3DTransform direction=forward, fileoutdepth=32f, interpolation=tetrahedral, gridSize=64, minrgb=[-5.96046e-08, 0, 0], maxrgb=[1, 1, 1]>

    <ROOT src="/Users/remi/ColorConfig/ocio-tcam/transforms/TLog_2_sRGB_TCAMv2.cub"></ROOT>

• Might be useful for optimizations, too, ex. in color space comparisons, etc

• Doug: the Processor seems like the correct place to put this

• Remi will create an issue

Project Roadmap review & discussion

https://github.com/orgs/AcademySoftwareFoundation/projects/21/views/1

• OS Color management - app needs to be able to tag the buffer going to the OS from OCIO

• Will also require further work in the Color Interop Forum space - linking OS level enums to OCIO color spaces

• Filmlight released a video around this space and why it’s important: https://youtu.be/_6TQ66cEiLw?si=G-g-mQ4B1if5F9oH

• Will add issue around clang-format and clang-tidy

• Please review issues & align with your company’s priorities