Academy Software Foundation Technical Advisory Council (TAC) Meeting - September 18, 2024

Join the meeting at https://zoom-lfx.platform.linuxfoundation.org/meeting/97880950229?password=81d2940e-c055-43b9-9b5a-6cd7d7090feb

Voting Representative Attendees

Premier Member Representatives

Project Representatives

Carol Payne - OpenColorIO Representative
Cary Phillips - OpenEXR Representative
Chris Kulla - Open Shading Language Representative
Diego Tavares Da Silva - OpenCue Representative
Jonathan Stone - MaterialX Representative
Ken Museth - OpenVDB Representative

Industry Representatives

Jean-Francois Panisset - Visual Effects Society## Non-Voting Attendees

Non-Voting Project and Working Group Representatives

LF Staff

David Morin - Academy Software Foundation
Emily Olin - Academy Software Foundation
John Mertic - The Linux Foundation
Yarille Ortiz - The Linux Foundation
Andrew Grimberg - LF Release Engineering

Other Attendees

Rob Rowe, CinePaint
Sam Richards, ORI Encoding WG
Deke Kincaid, Digital Domain
Michael Min, Adobe
Youngkonw Lim, Samsung
Lee Kerley, Apple
Lorna Dumba, Framestore
JT Nelson, Pasadena Open Source consortium / SoCal Blender group

Antitrust Policy Notice

Linux Foundation meetings involve participation by industry competitors, and it is the intention of the Linux Foundation to conduct all of its activities in accordance with applicable antitrust and competition laws. It is therefore extremely important that attendees adhere to meeting agendas, and be aware of, and not participate in, any activities that are prohibited under applicable US state, federal or foreign antitrust and competition laws.

Examples of types of actions that are prohibited at Linux Foundation meetings and in connection with Linux Foundation activities are described in the Linux Foundation Antitrust Policy available at linuxfoundation.org/antitrust-policy. If you have questions about these matters, please contact your company counsel, or if you are a member of the Linux Foundation, feel free to contact Andrew Updegrove of the firm of Gesmer Updegrove LLP, which provides legal counsel to the Linux Foundation.

Agenda

General Updates
- OpenQMC #434
- Dev Days 2024 #662
- Project Leads Office Hours #760
DPEL Annual Review #472
OpenAPV #803
Evolving our working groups program #798
Proposal to vote on MaterialX becoming an Adopted member #840

Notes

General Updates
- OpenQMC #434
  - Eric E: no updates at this time
- Dev Days 2024 #662
  - Less than a week away!
  - Larry: held two office hours, first was quite busy, second a couple of people. Not sure what to expect with third one, maybe busy?
  - Middle one may be a bit sparse because people are procrastinating
  - A bunch of poeple signed up, projects ready to go, hope it all goes smoothly
  - Carol: have any project leads seen activity from contributors? We don’t have visibility into that.
  - Jonathan: lots of new folks coming into the channel, but none have asked about specific issues. Maybe share the raw numbers we should expect?
  - Yarille: should have updated numbers - 80
  - Larry: will tell us the number of who signed up, but doesn’t break down by project unless they volunteer the info. Sometimes people jump around at the last minute.
  - Carol: we increased since the last time, after PR posts. Everybody should still repost in your networks the blog post. Would be good to put something in your channel about Dev Days explicitly. Please look at Good First Issues for instance.
  - Larry: would like to know if there are companies with large blocks of people participating
  - JT (chat): I think I mentioned custom github tag for dev days last year, used like the good first issue tag, that seems like a good idea to me for people that will scan the repo for inspiration.
  - Eric R (chat): OTIO has had two people reach out to say they intend to participate
- Project Leads Office Hours #760
  - Next week, we know it’s hard to mix with TAC / USD
  - Working on signed releases, Cary put together great material, a couple of projects have taken this on
  - Identify a project that hasn’t done it and help
Proposal to vote on MaterialX becoming an Adopted member #840
- Jonathan: we presented to the group in February, overview of our projects. At the time we were missing the OSFF requirements. We’ve done a lot of work, and think we now meet all the OSFF requirements to become an Adopted project. Do people want to do that vote with that context.
- Used a lot of the info from Cary / OpenEXR
- Meet the 80% code coverage
- We don’t have quorum (11 present / 12 absent), will do LFX vote
DPEL Annual Review #472
- Slide Deck
- Eric Enderton was previous TSC lead, he is stepping back, Matthew Low, Dreamworks, chair since Feb / March 2024
- Definition: A library of digital assets - 3D scenes, digital cinema footage etc - that demonstrate the scale and complexity of modern feature film production, including computer graphics, VFX and animation. Curated by the ASWF, these assets are available free of charge to researchers and developers of both open source and commercial projects, to test, demonstrate and inspire their ideas.
- Uniform Licensing Infrastructure: ASWF Digital Assets License v1.1
  - Vetted by member companies
  - solely for education, training, research, software and hardware development, performance benchmaring, or software and hardware production demonstrations
  - Now on SPDX, so we have the shorthand alias for these licenses, can embed in assets
  - Interesting model for other projects: Caldera model announced at SIGGRAPH didn’t adopt the same license, but inspired by it.
- Assets Delivery
  - Homepage
  - Hosted on ASWF / S3
  - Web site promotes and advertises these assets
- What’s new
  - GitHub Website Migration
    - Previously managed by LF
    - New GitHub-based version contributed by DreamWorks in October 2020
    - Move management to GitHub
    - Adopt Astro framework
    - Direct contributions via PRs
    - Automated deployment via Actions to Pages
    - Faster turnaround & autonomy: minutes instead of days
    - Local preview
    - 27 PRs, 11 contributors
  - AWS Picchu Edit movie
    - Contributed by AWS & FuzzyPixel in Dec 2023
    - Award-winning short film follows the journay of an Andrean girl named Mayu propelled by the unconditional support of her mother
    - Edited with DaVinci Resolve
    - Resolve Project, source media, rendered movie, and exported OTIO assets
  - Airship Asset
    - Contributed by AWS & Fuzzy Pixel in July 2024
    - From the Spanner short film, alongside Noa character asset
    - Fully rigged blimp
    - High res textures and materials
    - multiple Maya ref files
    - Renderable with Arnold
  - ALab v2.2. Updates
    - Initial contrib by Animal Logic in 2022
    - Updated in July 2024
    - Breaks out TechVar components
    - Defines main OpenUSD asset structure separate from geometry, lights, shaders and rigs
    - Migration to GitHub
  - GitHub for Asset Hosting
    - Previously opaque zip files on AWS
    - Enable greater discoverability and collaboration
    - Nire readable, explorable, linkable
    - Better documentation with GitHub Pages
    - Facilitate easier community experimentation
    - Encourage contributions via forks and PRs, previously was difficult to contribute back to the project
    - Best for assets restructured into smaller text files
    - There are limits in file (100MB) and repository (2GB) sizes (outside of LFS).
    - GitHub does good job for searching and linking with text files
    - Hoping to move more assets to this
  - OpenPBR Shader Playground
    - Contributed by Adobe in September 2024
    - Additional contributions from NVIDIA
    - Novel aspects of OpenPBR Surface
    - OpenPBR nodes within MaterialX documents referenced into OpenUSD scene
    - Imageable within Arnold and Omniverse
      - Requires some future versions, not fully deploying the assets on the web site, but deploying the infrastructure and the asset in “early beta”.
    - Beta hosting on GitHub
    - Good collaboration between DPEL and MaterialX
    - Newer version of USD will support the inter referencing of the asset.
    - Will be available on GitHub shortly
  - Future Assets: StEM v3
    - Standard Evaluation Material v2 (StEM v2) contributed by ASC in 2022
    - Reference material for color and image processing pipelines, display and projector calibration etc
    - Emphasis on HDR, high resolution, wide color gamut
    - Largest asset, over 4TB
    - StEM v3 focuses on Virtual Production / ICVFX
    - From ASC Joint Committee on Virtual Production
    - Mix of 2D & 3D assets
    - Contributions from numerous studios
    - Targeting beta Q4 and v1 @ NAB 2025
    - Movie is already written, waiting for financing to shoot, but keeping NAB 2025 target date
    - Questions around IP/licensing and hosting costs
      - May be much larger than StEM v2 because of 3D assets
      - May not have IP rights to all the contents
- Download Statistics
  - Scripts to generate stats, started in March 2024
  - DPEL Asset Size
    - Largest is ASC StEM v2: 3.8TB
    - ALab: 400GB
    - AWS Noa: 5GB
    - Intel Cloud Library: 105GB
    - ALab Trailer: 1GB
    - 4004 Moore Lane: 16GB
    - Picchu Edit: 7GB
    - AWS Airship: 45GB
  - DPEL Asset Bandwidth
    - StemV2: 40TB
    - ALab: 47TB
    - Noa: 312GB
    - …
  - DPEL Asset Downloads
    - ASC: StEM2: 822 times
    - ALab: 3288 times
    - …
  - Download Analytics
    - S3 Storage Lens
    - CloudFront Metrics
      - Both have now been enabled for the project
    - Want to understand geographic distribution, impact of events like Open Source Days
    - Want to understand cost impact
- Technical Steering Committee
  - Matthew Low, DreamWorks, Chair
  - Ben Fischler, Autodesk
  - Darin Grant, Animal Logic
  - Eric Enderton, NVIDIA
  - Haley Kannall, AWS
  - Joshua Minor, OTIO
  - Michael Johnson, Apple
  - Nick Porcino, Netflix
  - Satish Goda, Netflix
  - Sean MCcDuffee, Intel
- TSC Challenges
  - Participation ebbs and flows, we get 20-30% participation to bi-weekly meetings
  - Not a source code project
    - Different deliverables: web site, hosting / delivery platform, LFX tools may not capture our specific project metrics
  - Contributions are substantial and singular
    - High hurdles to contribution: size, many contributors
    - Contributed later than when created internally, could have been developed for a show and repurposed later. Original authors may not be available anymore
    - Legal, Creative review hurdles before contributing an asset
    - Long runway until an asset can be contributed
    - Have to go back to original author to get update
    - Contributions are segmented, little cross over
    - Copyright remains with the company, contributors to one asset don’t work on other assets
  - Lower engagement, collaboration, and TSC stability
  - Mostly “strategic” goals, few “good first issues”. Little incentive to “stick around” after initial contribution.
  - We’re mostly stewards of the assets
- Opportunities
  - Lower barriers to contribution
  - Create source code components
    - Web site maintenance
    - Astro framework has been huge help, but still barrier to bringing new asset
    - Web site is a bit static / staid, want to make it more dynamic
    - Implement 3D web viewer, OpenUSD / MaterialX. Asset preview hosted by DPEL could be valuable to market DPEL.
  - Grow visibility and stabilize TSC
  - Do pipeline demonstrations with these assets
- USD ALab
  - Talk at Open Source Days by Darin Grant on serial volunteerism
- Future
  - Increase engagement and collaboration
  - Continue GitHub asset migration
  - Website improvements
  - Explore 3D web viewers
  - Grow ASWF engagement and public visibility
  - Enhanced download statistics and analytics
  - Solicit new assets
  - Should we expand outside of just film assets, for instance Game assets (influence of Activision Caldera asset)
Project Adoption Status
- Last time TAC voted on Adopted Project
- Realized we fell short OpenSFF requirements so back to Incubation status
- Now meet all the requirements except one, that was voted as “not required”
- We believe we now meet all the required for Adopted stage
- Request vote to move to Adopted stage
- Questions?
  - Michael: call for participation and better metrics are overlapping a bit? How many projects in ASWF leverage these assets? Can we show that these assets are being used for existing projects, do we have that visibility. Also any modifications to OpenPBR model, I can help with that. Matthew: we got the contribution agreement signed, will stand up GitHub repo tomorrow and get first contribution added. We are in a great spot.
  - John: we will take this to LFX vote!
OpenAPV #803
- Youngkwon: we had internal discussion about Apache via BSD 3-Clause, we see that BSD 3-Clause has benefits over Apache 2, cannot verbally summarize it, but can share what our legal counsel says. Could be similar to OpenEXR? Larry: most of these projects existed before Apache 2.0 existed, and the license switch would have been too difficult. If those projects started today, they might have started with a different license. Youngkwon: we still prefer BSD 3-Clause unless ASWF has real issue with that.
- Carol (chat): Yeah OCIO just can’t switch
- Larry (chat): OIIO did switch when it entered the foundation, after 15 years of BSD (which it chose in 2008 only because that was the common one at the time).
- Matthew (chat): OpenVDB also recently relicensed
- Comparison with other codecs: we don’t have a license for ProRes, our engineers have some tests based on ffmpeg, it shows that OpenAPV has a bit more gain (19%). Eric R: when you say ffmpeg implementation, is that using ffmpeg implementation of ProRes? Youngkown: hard to confirm that, that implementation hasn’t been “blessed” by the owner. Eric: wanted to see an indepth analysis of where it sits in the codec landscape, looking forward to JPEG family codecs. Especially HT J2K for instance, already available inside Resolve, then OpenJPEG implementation of that codec. So maybe a good one to one comparison? Taking on the work of bringing in a new codec into the world is a long path to adopting a codec. There’s all the work around standardizing the container around the codec, what would we advocate. Want to start from metrics to see that the codec is worth putting the effort into. Youngkwon: we don’t have a direct comparison with H2K, but can be compared indirectly through ffmpeg implementations. We also have considered HT J2K, but had issues with patents / IP. Eric R: we haven’t seen IP issues for HT J2K, but can depend on the context / point of view. Youngkown: yes, different between a service provider vs a device manufacturer.
- Youngkwon: about the performance comparison, can’t provide official numbers for now, but we have tested APV on flagship phones, S series, full HD 1920x1080 60P 4:2:2 10 bit video, target bit rate of 400 mbps, phone could encode up to 120 FPS, decoding a bit more complex, around 86 FPS. Target was to encode 4K 60P in real-time, so in that range of complexity.
- Sam: are you going to release an ffmpeg patch? Youngkwon: working on it, will release it source code that way. Simple GitHub under OpenAPV org, can grab it that way and test it. Sam: do you think you can add something to the APV repo talking about licensing? Would be great to have a summary on the web site to point legal council to so we can convince them to add the codec. Youngkown: we do have licensing terms there. Just the standard BSD 3-Clause license. Sam: would be good to have a description that there are no known IP issues. Eric R: the license is the copyright on the code, but is there an attestation that the code is no IP encumbered. Don’t know if there’s legal language that’s typically used for that? Youngkwon: we could mention some of that. We have done extensive search on existing literature and have found nothing newer than 20 years, and haven’t added anything patents held by us. There could always be a surprise, but based on knowledge and process, we don’t believe there are applicable patents.
- Larry (chat): Apache-2.0 says that the contributors of the code attest that any patents they may hold on it are considered licensed for use in the context of using that code base.
- John: we are at the hour, so we can continue the conversation. Youngkown: can continue discussion via email / Slack