2024-08-19

Owned by Carol Payne
Aug 19, 2024
3 min read

August 19th, 2024

Host: Carol Payne

Secretary: Carol Payne

Attendees:

Rémi Achard (TSC) - DNEG
Mark Boorer (TSC) - Industrial Light & Magic
Mei Chu (TSC) - Sony Pictures Imageworks
Sean Cooper (TSC ACES TAC Rep) - ARRI
Michael Dolan (TSC) - Epic Games
Patrick Hodoul (TSC) - Autodesk
Zach Lewis (TSC) - Method
Thomas Mansencal (TSC) - Weta FX
Cuneyt Ozdas (TSC) - Autodesk
Carol Payne (TSC Chair) 
Mark Reid (TSC) - Animal Logic
Mark Titchener (TSC) - Foundry
Carl Rand (TSC) - Weta Digital
Doug Walker (TSC Chief Architect) - Autodesk
Kevin Wheatley (TSC) - Framestore
Name - Organization

Apologies:

  • Thomas Mansencal, Mei Chu

OCIO TSC Meeting Notes

  • ACES 2.0 planning (library and configs)

    • Good progress from both Remi and Thomas

    • Library: Remi has implemented built-in stubs for all 50 transforms

    • Thomas is working on the config portion now - starting with the spreadsheet to drive the config generation. This is a completely new spreadsheet for ACES 2.0.

    • We should likely have a configs WG meeting to talk through these things separately

    • Doug will take a stab at linking Remi's work to Thomas' spreadsheet

    • Remi: the HLSL issues are interesting. Not sure if it's easy to fix... it's ok in GLSL. Likely shouldn't use binary search.

    • Kevin: Maybe try just doing the hues as constants? And then use the texture lookups on the rest?

    • Remi - that could be a good workaround for now, until we have a more optimized solution.

    • Kevin: managed to reduce the amount of binary searching needed, but couldn't eliminate it. 

    • Kevin: also worked on optimizing the tables to reduce the amount to search. Code is very different from CTL, slightly visually different. Needs a lot of testing. 

    • Doug: release is approaching quickly, how are we feeling about the API/ABI state? Remi, how can we help out?

    • Remi: any help with testing, and will make a list of TODOs so folks can help.

    • Doug - working on putting together a list of tests, images, error metrics. Waiting on what those tests specifics and also images, though we theoretically can proceed without them.

    • Kevin: image data set creation would be with Scott. Working on that.

    • Kevin: as far as test strategy, we have a base core list (it's not been tested), but could be a start.

    • Doug - can you share this list?

    • Kevin: will share when it's public.

    • Kevin: all of the API ABI decisions need to be set in stone, so we should be looking at that to make sure we can do the updates as a point release. What are all the fixed, named, builtins, etc. 

    • Remi: the only reason I could see to have a different breakdown of functions would be optimizations...

    • Carol: should we get a meeting scheduled with Remi, Kevin, Eric, Doug, Thomas, to talk details?

    • Remi: already have a script to test things using ocioconvert, can share that soon. 

  • ACES D65/D60 

    • Mark: has the ship sailed on including the D60 sim versions? Can we not?

    • Kevin: on the ACES side, the ship has sailed. 

    • On the OCIO side, we have options. We can make the builtins, but decide how to expose them - two different configs, defaults, etc... we can talk.

    • We could choose to not include top level Built-in Transforms for the D60 versions. And just rely on building things for fixed functions for the D60 versions.

    • Remi: we currently have the built-ins in the PR. Would be harder to remove once added

    • Doug: it's really not a lot of code to add the library to have these built-ins. No real reason not to add it.

    • Doug: we could also as OCIO choose to not "build in" the D60 version. 

    • Mark: I just don't want a huge list in the API. It gets quite complicated to maintain for operations that might be happening outside 

  • Dev Days

    • Volunteers?

    • Create/Tag more issues? (current issues here)

    • Dev Environment Containers - let us know if folks have ideas for this / want to help

  • CIF - please write down your opinions on Color Space Naming before the next meeting.