2021-10-12
- Michael Dolan
October 12, 2021
Host: Michael Dolan
Attendees:
OCIO Config Working Group Meeting Notes
IDT implementation working group:
Carol: Went well. Receptive to ideas around using CLF. Still early in group, looking at how to create IDTs, and how to improve that experience for vendors. Haven't thought as much about how to implement the IDTs yet. Open to us doing this, and working with us for naming, storage, process, etc. But not ready for the discussion in their group. We explained we are wanting to work on this soon. Showed the doc and suggested we keep working on that and fleshing out the CLF specification. We will talk to camera vendors about how this could fit into their workflows, and then go back to IDT group to discuss again. As a starting point can use the existing ACES IDTs in repo.
Doug: Agree it went well. Just wanted to make them aware of what we're doing and collaborate. Make sure there are no objections to working on this in parallel. Not sure it speeds up our work, but makes it coordinated with their group.
Michael: Any thoughts from camera vendors?
Carol: Joseph said he would look at doc. Wondered about handling different IDTs from vendors. Different versions, etc.
TODO: Michael will propose a simpler version scheme than current URN approach for next time.
CLF document updates:
See document: https://docs.google.com/document/d/1uYNnq1IlKqP8fRXnPviZHrAAu37ctvVsjJZeajOFF2A/edit?usp=sharing
Carol: If vendor adds new IDT to ACES, do we add that to OCIO as builtin, or as CLF for referencing into config?
Doug: Builtins makes it easier to use, so users don't need to implement them directly. Popular color spaces will probably become builtins over time.
Carol: So these CLFs are placeholders for transforms that don't have builtins (yet).
Doug: The main goal with these is to build the ACES studio config. There are IDTs needed that aren't builtins, so will need CLFs. Not sure we should make a PR for CLFs that do have builtins. Maybe PR should be examples of transforms needed in the studio config.
Kevin: Would need to use a published transform, from a paper. These CLFs are a means of bridging when the transforms exist vs when they are made as builtins. Some will be added to OCIO and other less used transforms won't. ACES would also appreciate the portability of these CLFs.
Carol: Think we should pull out ALEXA LogC IDT as an example for a starting point.
Doug: In the doc right now is the ACEScct example. LogC will look similar, but with different numbers. Can put it in the document.
TODO: Doug will add LogC EI 800 example to doc.
Doug: Will need to clarify for vendors what we're looking for. If we already have builtin transform, maybe we need to clarify this is for new IDTs. Or do we ask them to take over maintenance? Want to make sure we don't ask for stuff we already have.
Carol: For vendors where we already have builtins, can let them know what we already have, ask for review, and ask for CLF for any transforms to add to that. For those without builtins, ask for CLF or offer to help.
Doug: Vendors don't always publish this data. Main cinema vendors have a vendor outreach program and provide SDKs and whitepapers through NDA. We can't sign an NDA since this will be open. They may provide paper to implement. Do we need an argument for why they should take ownership of this?
Carol: If they don't want to provide the needed info, what do we do? What can we do? We can make our case for why things need to be added to OCIO config. Make it clear in the end that we need to follow open source rules, and we may not be able to include in config otherwise.
Doug: Do we need to expand the introduction of this document to explain what OCIO is, and what it's used for?
Michael: We could add a brief introduction, and provide links to existing explanations.
Carol: Might be good to work with a vendor like Black Magic first to test the process and get feedback.
Michael: Great idea.
Items for next meeting agenda:
Discuss CLF document, with the goal to finalize it.