2023-04-03

Owned by Carol Payne
Last updated: Apr 04, 2023 by Kevin Wheatley
2 min read

April 3, 2023

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
John Mertic - Academy Software Foundation / Linux Foundation
Carol Payne (TSC Chair) - Netflix
Mark Titchener (TSC) - Foundry
Carl Rand (TSC) - Weta Digital
Doug Walker (TSC Chief Architect) - Autodesk
Kevin Wheatley (TSC) - Framestore
Zach Lewis - Method Studios / CO3

Apologies:

  • Michael Dolan

OCIO TSC Meeting Notes

  • Issue #1781

    • RGB→RGBA should start with alpha = 1 rather than 0

    • Current behavior is inconvenient because 0 = transparent, applications might want to default to 1 so it's visible

    • Kevin: abstractly, makes sense that it should be 1 - it's pretty new - anyone know of anyone relying on it being 0?

    • Doug - change would wait for 2.3

    • Remi, Zach: seems like a bug fix, no big issue 

  • Issue #1668

    • Opt-in or opt-out for Windows Python 3.8 path fix?

    • Used to be that Windows Python would look at path var to load dynamic libraries, but they've changed it

    • Anders made a PR to set the path to restore previous behavior, which is what OIIO and USD did

    • Current merged PR is to make this fix opt-out, but have gotten some feedback that this is potentially dangerous, and it should be the other way around

    • Kevin: gut reaction is to limit the search path due to security concerns - but who is the default user? PIPy user? Application vendor? The approaches might be different.

    • Could also move to statically link - Remi - that's the way the python packages are already he thinks (will double check)

    • With static linking - how far do we go for things like system libraries?

    • Default for the standalone build should be the same as the wheels, and then this fix is not really necessary?

    • Maybe look at how others are doing this beyond OIIO - NumPy, etc? Could be more complex than needed for our case.

    • Possibly something to ask the CI working group for an opinion? Should the ASWF be in sync on this for all projects or is it not worth it?

  • Continued config merging feature discussion

    • Doug: updated the slides a bit more with context from the last meetings feedback

    • Doug: working on adding specific use cases for the three problem spaces presented

    • Remi: we actually had a use case this week for merging with LED wall calibrations

    • Kevin: is it truly merging we need, or is it merely more like "including" needed?

    • Doug: proposal would include options to only add, and not merge

    • Kevin - helper method to do smarts around different references spaces in config merging - different reference spaces, etc

    • Doug - hueristics added in 2.2 do some of this already - difference in reference spaces, etc

    • Could possibly help with instances with different grading spaces for cdls, etc

    • Will add a section to the slide deck for concrete use case examples. Please help us to add to them!

  • GSoD

    • OCIO was not selected this year.

    • Will try to pursue other avenues for funding, but nothing to report at this time.