ASWF TAC Meeting - December 2, 2020

Video Conference Link

Voting member attendance

Other Attendees

John Mertic, Linux Foundation
Emily Olin, Linux Foundation
Andrew Grimberg, Linux Foundation Release Engineering
JT Nelson, Pasadena Open Source consortium / SoCal Blender group
Alex Meddick, Rising Sun Pictures
Mathieu Mazerolle, Foundry
Deke Kincaid, Digital Domain
Greg Denton, Microsoft Azure
Mitch Prater, LAIKA
Alex Forsythe, AMPAS

Apologies

Agenda

Goals for TAC: Year 2
New member welcome
- Mark Visser, Unity Technologies
Industry Representation Vote results (John)
- John: Call for remaining votes for Daniel and JF as Industry Representatives
TAC chair election process (John)
- John: election is open, email went out to the TAC list on Monday
- Please send a note of interest, short bio, rationale as to why one would be interested in this position.
- You can nominate others as well, but please CC them / reach out to them first to avoid putting them in awkward position
- Close of nominations is next monday at midnight PST
- Daniel: keen to answer any questions as to what the role entails, open and available to answer questions in public or private
- Need to get vote finalized before end of year, so not a lot of time left, please move quickly
- Commitment is a minimum of 2-3 hrs per week to deal with formalities of meeting process, follow up with agenda items. There can be more stewarding / process alignment for working groups, can grow quite a bit. At this stage probably doesn’t require the TAC chair to be involved anymore, the WGs are working well autonomously. But when new projects get started up, that’s when more involvement is required. Governing Board meetings are monthly, with some reporting requirements, activities such as Open Source Days. Lots of work done by David Morin (chair of Governing Board) as well as John and Emily (LF), incoming chair will be able to draw on their experience and support.
- John: try to do as much of the legwork as possible, tend to lean on TAC chair to “have the pulse of the group” and set the driving direction, can “sit in the background” to drive the TAC meetings (Daniel has done that very well).
Project & Technical Community Survey
- Daniel: not a lot of progress due to some health issues. Some effort picked up by John and Emily to get survey published before end of year.
- Can we agree on a timeframe to work towards? OpenCue / Brian are very keen on these results, they were originally on early November timeframe.
- Emily: want to get it out as quickly as possible. Need to cull down list of questions from OpenVDB, and looking for preferentially multiple choice questions, makes it easier to produce results for one thing. Will start putting this into SurveyMonkey when every project has provided info, and can provide preview of SurveyMonkey format.
- Daniel: can we set goal of end of next week? Emily: yes, as long as the projects have provided feedback. Brian: don’t want to run into holidays at end of year. Daniel: will have a bit of time to help with this.
New CLAs (2020 v2.1)
- PR
- John: based on feedback (predominantly from Disney), landed on some tweaks to original CLA documents. Some concerns about being tied to a specific tool. Incorporated some of the language, changes are in the GitHub repo (see PR link). Aim for projects to aim for these revised CLAs. Anyone who had concerns around CLA has approved these new CLAs, so projects should be safe to move to those if they want.
- Each project is able to pick the CLA they want, the decision is up to the TSC, LF is happy to provide feedback / answer questions, and help with transition from previous CLAs. This should have unblocked OSL and OTIO transition to official GitHub ASWF organization.
- Larry: any project that wants contributions from the Disney companies will want to do this? John: largest feedback on previous CLA was from Disney, and this version addresses these concerns.
- Larry: PR was posted into “ramp up” to Thanksgiving, so still making its way through Sony legal. The diffs are useful to get legal departments who had approved the previous version.
- Joshua: putting new CLA through Disney approval process, should hopefully all be fine.
- Daniel: will work with the TSCs to move towards the new CLAs so that all projects use the same CLAs (if that makes sense).
- New branding for the EasyCLA tools, projects don’t have to use the tool, can submit “designated employees” manually if needed
New project proposal (Kimball)
- Proposal (google doc)
- Kimball: 4 years ago raw2aces project was started under AMPAS, WETA has similar tool which processes 400K images per year. Idea is to revive raw2aces project, and merge the WETA technology. A bit strange of a proposal since 2 subprojects, one is a database of spectral sensitivity measurements, spectral light measurements in the future. One project is a dataset, one project is the code that provides the reference implementation, batch capable high speed converter for DSLR images to EXR / ACES.
- Larry: only had a chance to skim the document, are you proposing only a batch tool, or could it be a library that could replace dcraw completely?
- Kimball: there will definitely be a library, would like to have a plugin system to accept contributions from camera manufacturers (ARRI, Sony…), so we can support motion picture cameras as well. But no plans to have a UI, don’t want to make a viewer.
- Daniel: is the project a fork or a move of the AMPAS project? Kimball: existing one is relatively minimal, lots of community interest to have such a tool. So a “move + massive upgrade”. Alex: backstory is that there’s a lot of interest in using DSLRs with ACES, converting raw files to ACES container files, requires creation of an IDT. Based on libraw, a bit of a skunkworks project. Realized there was duplication of effort between AMPAS and WETA, better to build a “bigger / better” version that would serve the needs of the VFX community.
- Eric: who owns libraw? How does it relate to dcraw? Kimball: libraw is a GPL library, not clear who owns / drives it. Alex: there was an intent by the developer of libraw to create a more maintainable version of dcraw (“a big chunk of C code”), so libraw was trying to reproduce the functionality of dcraw in a more maintainable way.
- Larry: libraw is widely used, OIIO uses it, so studios using OIIO to convert raw files are using libraw under the covers. But it is not easy to interface with, an ASWF project would be welcome.
- Larry: what would be the merits of a completely separate project, when it has a foot in ACES, OCIO, EXR worlds… Should it live under one of these projects? Or does it deserve its own TSC as a full ASWF project?
- Kimball: the issue of having the database of reference data may justify having its own project?
- Daniel: if there’s a particular set of expertise that borders on the commercial / proprietary that could encourage camera manufacturers to contribute, that could be useful. Larry: was primary concerned that there may be exact overlap of potential TSC members, but it can be justified if there are separate interests.
- Daniel: logistically, does the project proposal constitute an adoption of the existing AMPAS raw2aces project (transfer of repo, IP), or is it a new, from scratch project? Kimball: there will be a transfer of info, especially the existing spectral measurement tables. Not clear how much of the code would be reusable.
- Alex: don’t have a large development staff at AMPAS, the main developer on raw2aces has moved on to a different position.
- JF: would be useful to figure out where the project lives in the “stack” to avoid circular dependencies between ASWF projects? Kimball: don’t think the project will need to depend on OCIO, but will depend on OpenEXR, and OIIO would likely use it. So would be “somewhere in the middle” of the stack.
- Larry: how large is the code base? Would it be better to just implement it in OIIO as a reader? Kimball: would be larger than a typical library in a plugin folder in OIIO. Will need to produce IDTs for the different cameras.
- Daniel: let’s treat this as a project from the governance / contribution way, we can explore other mechanisms such as submoduling if that makes sense from a technical point of view. We typically have a first proposal for a new project, we take at least a week to digest, formulate questions, and move for a vote at the next meeting. One more meeting until end of the year, should probably move discussion to mailing list to address any other specific questions, and plan to have a vote at the next call.
- From the chat:
  - From Deke Kincaid to Everyone: (1:28 PM) Out of curiosity, why not simply wrap openimageio and improve its raw support if needed?
  - From Henry Vera (DNEG) to Everyone: (1:33 PM)
  - I agree with Daniel, that if it encourages more involvement from the camera manufacturers that would be great!
  - From Larry Gritz to Everyone: (1:35 PM)
  - I’m very enthusiastic about the existence of this project, I would love to replace OIIO’s current use of libraw with this in order to (presumably) improve RAW support.
  - From Henry Vera (DNEG) to Everyone: (1:36 PM)
  - We have some people working in a very similar vein, it would be great to get them more exposure if there are some questions about exactly how to position it. I’m presuming there are other’s with similar experience and tech. Should we arrange a specific call to dive more deeply?
  - From Deke Kincaid to Everyone: (1:36 PM)
  - My biggest worry is film cameras seem to be on our side for vfx but their photographic wings seem to be heavy on the proprietary and encrypting data in the raw formatsWhich is why libraw is a lot of reverse engineering of raw formats
  - From Henry Vera (DNEG) to Everyone: (1:36 PM)
  - Deke: I agree with your concern.
Technical Project updates
- OpenCue: hoping to have user survey data to have graduation proposal
- Larry: Imageworks has finally switched all productions to OpenCue from the internal predecessor system, OpenCue has successfully replaced the internal system and does everything. Largest challenge was switching the database, but that is now done.
Working Groups
- CI
  - We’ve been skirting around the important issue of package management. So far we’ve used Docker, but within the scope of the ASWF, we want to look at the problem of package management, making software available in a more fine grained, dynamic way that the studios require.
  - Next week’s meeting: Ryan Botriell (SPI) presenting SPK dynamic environment package management system. Want feedback on the approach they have been taking, had opportunity to look at the work and it looks very interesting. Encourage anyone interested in this topic to attend next CI meeting.
  - AWS announcement of Intel Mac Minis, availability of M1 Mac Minis in January from AWS and MacStadium.
- Diversity & Inclusion
  - Emily: webinars will start in mid February 2021 after students back from break.
  - Working on blog post around inclusive terminology, looking for ASWF project examples. Looking for tips / easy first steps that companies can make. Would help to provide context.
- Python 3
  - No Python3 WG meeting since last call
- USD
  - Meeting last week, first presentation on the agenda: USD example plugin repository, walks people through how to structure a USD plugin. Some interested parties may be contributing soon. Helps to get people up and running quickly.
  - Discussed possibilities for an open source camera rig in USD, follow up next week.
  - Looking for additional topics of relevance to the group, if you think of anything, please reach out via Slack or email
  - Daniel: nice to see tangible progress / outcomes
- Review & Approval
  - Interesting presentation-oriented meeting to that group on Tuesday. Larry: group is in a phase of getting presentations from a variety of in-house and commercial products in that space. Looking to take inventory of what people have done, what the needs are. Yesterday were Imageworks with itview and frame.io, were supposed to be 10 minutes each, both became quite longer + substantial Q&A, valuable discussion. Sean Looper: Is itview on the tablet for open source? Larry: not yet, just showing off what the different groups think is important in a media and review tool, we’ll see where things lead.
  - Will there be further contributions? Larry: meeting was supposed to make it through 5-6 systems, only made it through 2. Still trying to get a handle of what we want to come out of the WG, so still in exploratory phase. JT: Wiki page is coming along nicely. Daniel: nice to see the Wiki format working well for the WGs.
  - Joshua: a topic that came up in review and approval group and OTIO: how do annotations work, how to do annotations in USD (attaching notes to objects in space), topic will continue to come up. What’s a good way to organize discussion around annotations in the broader sense.
- Asset Repository
  - Wave: trying to “find our way”, had the most contributions from people in latest meeting, one of the things that came out is that the project could be about giving the “CLA equivalent” for donating an asset. Is it for publication / research? Is it “creative commons 0” and can be used, is it for education only? Personally not interested in legal (I am not a lawyer), but it is important. Leverage the Creative Commons effort. Previous discussions about a reference render of 3D assets, the Moana asset from Disney could be re-exported as USD, is being able to render with Renderman sufficient? If people want to bring up topics, please bring those up on the Slack channel. JT reached out to an archive at a University. Is it an archive? A digital backlot? Still trying to figure out.
  - John: can help with the legal definitions, how a CLA fits in. Because of the nature of this, may want to take this to the legal committee, which will likely have some feedback.
  - Wave: would be good to have clarity for the people considering contributions, before you can donate something from a large organization, will need to know under what terms. John: trying to avoid a multitude of licenses. Wave: what’s the ASWF’s opinion on an asset repository? Feel free to discuss this on the Slack channel.
- We should encourage the Rust people to join the call, is there something obvious the TAC can do to help this effort along. Sean Looper: can we pick a TAC sponsor to act as an informal proxy for those conversations? Would be useful to get a summary of progress, even if there isn’t a “formal” WG. Daniel: a WG could be easy, makes it discoverable, will reach out to some of the people in that space and see if they are interested, would someone be interested in leading a WG effort. Sean: having regular visibility would help.
Update on candidate projects
Next meeting: December 16

Action Items (AIs)

Daniel, Emily, Project chairs: Project & Technical Community Survey ready to release by end of next week.
Daniel: Reach out to active Rust group about WG formation or other points of contact with TAC
TAC chair nominations: close Monday 7 Dec
Further discussion of proposal for raw2aces, intend to vote at next TAC meeting

ASWF TAC Meeting - December 2, 2020

Voting member attendance

Other Attendees

Apologies

Agenda

Action Items (AIs)

Notes

Chat