ASWF TAC Meeting - February 24, 2021

Video Conference Link

Voting member attendance

Other Attendees

John Mertic, Linux Foundation
Rachel Romoff, Linux Foundation
Andrew Grimberg, Linux Foundation
JT Nelson, Pasadena Open Source consortium / SoCal Blender group
Mathieu Mazerolle, Foundry
Carol Payne, Netflix, ASWF D&I WG
Scott Wilson, Rust WG
Nick Porcino, ILM
Rod Bogart, Epic Games
Tiffany Fung, Pixar
Van Bedient, Adobe
Greg Denton, Microsoft Azure

Apologies

Agenda

Upcoming Events
- EasyCLA webinar after this meeting
  - For CLA managers, or interest in how this tool works
  - Requires pre-registration
  - Will be pre-recorded, if you register but can’t attend you will be sent the recording
- Open Source Forum
  - Everyone on initial invite list should have gotten an invite email
  - Anyone know execs who should be invited in their organization?
- Others?
Welcome to Rachel Romoff
- Taking over from Emily at Linux Foundation for marketing role
- Already part of other LF foundations
- Attended ASWF meeting last week
Governing Board Meeting Yesterday
- Lots of time spent on project updates (presented by Kimball)
- Excited by progress in OCIO
- Lots of discussions about Asset Repo
- As a group we need to keep in mind how to demonstrate values for all members, “why you want to be part of ASWF”.
Technical Project updates
- OSL
  - Discussing a new logo, crowd sourcing
  - Usual technical updates
- OpenCue
  - Recently had a chance to go over survey responses, got a lot of good data from that.
  - Wrapped up some of the last bits on the CII badge, need to update form, all the work should be done before preparing graduation proposal.
  - TSC meeting next week to review that, so will send proposal before next TAC meeting
  - Kimball: CII badge status of various projects was a question at Board Meeting. Brian: “the last 10% took a long time!”
- OCIO
  - Michael: talking about ASWF survey at last meeting, how to follow up with people who wanted to contribute to the project
  - Moving meeting notes out of the repo and into the ASWF wiki, lower noise in repo.
  - Carol Payne: applied to GSoC as an independent project, waiting for updates
- OpenEXR
  - Close to ready for a release, but still some cleanup issues. One of the Linux distro package managers has been finding some issues here and there, glad that got done before a release. Otherwise would have been scrambling to fix issues after public announcement.
  - Originally said next week for the release, but that’s probably not going to happen. But should have beta release before Open Source Forum.
  - Larry: all of our CI tests with the ASWF docker containers are careful to have only one version of dependencies installed, so downstream projects can fail to notice that they have OpenEXR 2.0 installed while wanting to test with EXR 3.0 (order of include files…). Can be tricky, and will be good to have a beta.
  - Cary: complexity about generating bindings for Python 2 and 3 at the same time. Driving the desire for simplicity.
  - Kimball: time to deprecate the Python 2 bindings…
  - Larry: just deleting the old versions is not a good general purpose approach. Still working out the details. Challenging to get the downstream CMake build systems to get the include in the right / desired order.
  - Kimball: is there general knowledge to be derived, for instance the transition to OCIO 2.0
  - Cary: our CI is testing against an artificially simplistic environment.
- OpenVDB
  - No representative present
- OTIO
  - Josh: meeting tomorrow to look at TSC charger and move forward, have all the right approvals to move
  - Continuing work on CII badge
  - Just turned on CI across Linux/Windows/macOS, uncovered some issues, but largely pretty smooth. Gives us confidence that we’ll get more testing across Python flavors and platforms.
  - Lots of feedback about “why does this not install on Windows”, hopefully the CI will help with that.
  - Hoping to use that to draw in more contributors who use Windows primarily
  - Splitting some pieces of OTIO repo into separate repos, separating language bindings and adapters / plugins into separate repos, to make it clear what the core project is responsible for, whereas other components can trail a bit, and that you can chose which components of OTIO you want to install.
  - Other studios are starting to create specific plugins (for instance for Hiero) that doesn’t necessarily fit into the core repo. So there’s now a “plugin template” that helps with that.
- RAW2aces
  - Kimball: pre-TSC meeting, trying to identify other people who can be involved. Discussions about re-licensing the code, splitting the repo to make clear what is data and what is code (helps with licensing)
Working Group updates
- CI
  - No meeting since last TAC meeting
  - Next meet agenda: moving to a project format to have more access/ visibility in CI efforts
  - CI group could look into the issues reported by Larry about dependencies
- D&I
  - Carol: first webinar happening tomorrow, second one the week after, then every 2 weeks. 7 in total
  - Although not target audience, everyone is welcome to attend to give feedback
  - Next effort will be scholarship and ambassador programs
- Python 3
- USD
  - Cory: Setting up a series of demos, including Pixar Backlot
  - Mike M: USD Camera Workgroup discussing schemas, creating Requirements Doc
- Asset Group
  - https://docs.google.com/document/d/1HuxNSlV6fS_3km6QlwA5Z9oBhNgx_NlhZjV2jg2DrFg/edit#heading=h.4ybh7vyxdywk
  - Wave: Trying to communicate with the TAC what we have been working on for the last few months
  - Not going for “quantity”, going for quality. Going for production level assets, ideally from a production setting, not artificial, although have talked about one way of having an asset repo could be code that generates a sequence of OpenVDB / OpenEXR files (for instance generate EXRs with a bunch of AOVs from a 3D scene).
  - An organizing principle is that there should be strong alignment with the actual ASWF projects, and should either be consumed by or generated by ASWF projects. Bias towards assets that can be used by as many parts of the ASWF as possible, as well as ASWF members such as software vendors who can use these assets for their own use.
  - Not targeting “educational” assets for now, instead going for complex, production grade assets.
  - Not a “digital backlot”, not a repository of digital assets to be used in production, that already exists (both public and private repositories)
  - Kimball: was there a discussion about a goal to help how to get contributions to this repo, such as listing the steps that were required for licensing / legal buyoff? Wave: yes, strawman proposal by Eric Enderton. “Make it easy for organizations to submit assets”, could be an issue if too many proposals come. Talked to Nick Cannon at WDI about the process of the Moana Island asset. The long time to produce a USD version of this asset demonstrates how complex the process can be.
  - There were suggestions such as camera manufacturers shooting HDR footage, talking to the ASC (suggestion at board meeting yesterday), they have a library of digital assets. But are we “a list of pointers to assets hosted somewhere else”, “a clone of existing libraries”, or “the primary hosting location”?
  - TAC Project representatives should suggest what assets would be useful to their project.
  - Making it useful: will require people and resources. Orca is a library that NVIDIA has and makes available to researchers. Presentation from NVIDIA as to how much work is required to clean up and make assets usable.
  - Need a “pipeline TD” to do this work, could be a GSoC-like position? Looking at what the D&I team is putting together in terms of outreach programs, could we get funding to get some kind of internship, minimal code writing (maybe some Python), but still useful project.
  - Trying to put together a curated set of production grade assets that would be directly useful to the projects and ASWF membership. Not targeting educational assets. Do whatever can be done to make easy for licensing and hosting. Make it clear what is desired as contributions. Asking for TAC members about opinions of what they want as assets, a wish list of “assets you with you had”.
  - Gordon: being able to test DCCs against these assets would be very useful. One of the things vendors find difficult is finding assets they can demo in their videos / promotional material.
  - Blender is working on a new short film, OTIO would like to get cuts and media, they already have pretty permissive license.
  - RV / Shotgun was using Big Buck Bunny as demo material for a long time.
  - Kimball: did we learn anything from Blender’s tutorial assets? Wave: it’s more interesting to get a dirty / crufty production asset at the end of a movie with all sorts of typical leftover “detritus” in the asset, as opposed to a carefully crafted asset for tutorial purposes. DCC vendors probably want to test against “crufty” production assets.
  - Josh: having assets that are clear to include in demos and presentations is really important, as opposed to just “for testing only” assets. It’s a real pain to get approval to use assets for demos. It would be very beneficial.
  - Josh: we also need to recognize that large studios may not want to release assets that have valuable IP associated to them
  - Kimball: if ASC is working on a new STeM, we should provide that feedback to them
  - Wave: multiple ways to achieve that, we could have a website that points where to find this, we could have a badging system to indicate “production level assets”, that could be the answer if we can’t actually host ourselves. But if Disney made Moana available, it is possible!
  - Michael Min: early discussions motivated by Autodesk for an Asset Model, would like another 2 weeks to consolidate. Gordon: will have people working on this proposal present at the next meeting. Not meant to be an asset / library repository, but more a schema to structure asset above the raw geometry. An interesting question: what is the opportunity / overlap / relationship between that work and the asset WG work.
  - Wave: Intel and Amazon have a number of assets they are interested in contributing.
  - Wave: this is a WG, not a project, so it has a limited scope. Suggestion would be to wait until the Autodesk presentation, people will have an opportunity to comment. Can present at next TAC meeting to vote, not super clear on what the outcome should be. Need to hear from TAC members.
  - Eric: if the TAC approves the proposal, can present to the board, and request resources. Do we want to launch an asset repo project, derived from the work of the WG.
  - Open Asset Model (OAM) schema hasn’t been presented yet.
  - Gordon: is there some intermediate goals that can be carved out before going to the governing board. Are there some questions that can be answered before that? Wave: organize the assets by ASWF project, a clear way to organize it. Drifted a bit from that towards assets like Moana that don’t neatly fit into a single project, but could work across them. “What do we want in an ASWF asset repository”, come up with a wishlist, and “Are there any assets in ASWF member companies that could be contributed”. Also are there “recipes” that can be contributed (for instance Intel has code to generate OpenVDB test assets), doesn’t have legal / licensing issues since the assets weren’t used in a movie.
  - Gordon: something to be said about real production assets. We also don’t need permission to create a static site with links pointing to existing assets out there. Could we get to a real agreement that studios could sign?
  - Wave: yes, that’s the first deliverable.
  - Eric: first thing that got agreed to in the discussions. Wave: goes back to “go back to your facilities and ask what you might agree to”.
  - Kimball: is there something we should vote on today? Wave: let’s let people to read the proposal, come back in 2 weeks to vote on something.
  - Eric: also on the Slack channel.
  - Gordon: OAM presentation to TAC, Kimball: will add to next agenda
- Rust
  - First meeting as WG last Friday
  - Creating Rust crates, cargo index, looking at transferring ownership to ASWF
  - Cppmm should automatically C and Rust bindings
  - Next meeting in two weeks: goal is to have OpenEXR have as complete bindings as possible (meeting cadence once a month)
  - Larry: where does the tooling status lies? Scott: creating bindings for OpenEXR / imath is the immediate goal, tooling is mostly done. Should have automatic tooling in place.
  - Kimball: have a pending PR for a C library for OpenEXR
  - Scott: won’t generate the C bindings for OpenEXR to avoid duplication of effort, but perhaps the tooling might help
  - Kimball: C bindings will be the underlying infrastructure. C++ will be on top.
  - Scott: Trying to figure out how to organize ourselves
Open Asset Model (mmin)
- Note: We’d like to delay discussing this to the next TAC meeting
Next meeting
- 10 March 2021

ASWF TAC Meeting - February 24, 2021

Voting member attendance

Other Attendees

Apologies

Agenda

Action Items (AIs)

Notes

Chat