W3C

– DRAFT –
MEIG monthly meeting

01 February 2022

Attendees

Present
Andreas_Tai, Barbara_Hochgesang, Chris_Lorenzo, Chris_Needham, Francois_Daoust, Kaz_Ashimura, Kazuhiro_Hoya, Nigel_Megitt, Piers_O'Hanlon, Takio_Yamaoka, Tatsuya_Igarashi, Xabier_Rodriguez_Calvar
Regrets
-
Chair
Chris_Lorenzo, Chris_Needham, Tatsuya_Igarashi
Scribe
cpn, tidoust

Meeting minutes

Agenda bashing

<kaz> Agenda for today

cpn: Today is about reviewing progress on a number of topics. Any other topics that you're interested in, please raise, we can probably fit that in.
… Current status of the Media Timed Events activity. Some updates happening in Chrome, e.g. Picture-in-Picture.
… Also media areas where feedback is sought for
… I'd like to come back to application performance that we discussed during TPAC. I'd like to report about chairs discussions and ask for feedback.

DataCue and video SEI event messages

cpn: Under the Media Timed Events TF umbrella, we have a couple of activities:
… 1) the DataCue proposal, which is a way to allow web applications to trigger timed metadata events along with audio/video tracks, specificially looking at emsg boxes in collaboration with DASH-IF.
… 2) Recently, we heard from ByteDance about SEI metadata exposure.
… Questions: Commonality between the two? Could they become part of the same proposal? Should they be addressed separately?
… Separate meeting happens on the third Monday of each month. Next call on Monday 21 February at 4pm UTC.
… The broader point that I wanted to make with this media timed events activity in general is that more participation would be helpful.
… If we want to move these proposals forward, we would need someone who could edit the documents.
… At the moment, for the DASH event work, there is some level of interest in the industry, although that is hard to evaluate. In order to move forward, we need companies to show support. This is a call for input, starting with expressions of support.
… Beyond that, someone actively involved who can write some of the explainer documents would be very welcome.
… Thank you to those of you who already participate.

Picture-in-Picture API for abitrary content

cpn: I saw an intent to prototype announcement from Chrome on a new API or an extention of the Picture-in-Picture API to work with any kind of element.
… The API that the Media WG develops allows apps to create an always-on-top overlay from a video.
… Useful feature, we use it internally.
… What this new API proposes is to apply that to other elements to create pop-out windows that can contain arbitrary elements.
… Very early stage proposal. Some opportunity for us to give some input.
… Is this a feature that we'll want to use? Do we see some additional services beyond the ones envisoned.
… An example use case that the BBC is considering is to show our custom video player with our own styling and custom controls.
… Right now, the controls in the Picture-in-Picture API are up to the browsers.
… I've raised our use case in the discussion. I'm wondering about other use cases that are relevant for each of you that we could put forward.

<Barbara_H> Use Case - Impact to Education? Education market is web based

cpn: Also, do we foresee any problem with the proposal?

Barbara_H: I don't know if there's an IG that monitors education, but Intel focuses a lot on the future of education. I can exchange with an architect and see if they have some feedback.

cpn: Please do.
… The idea that you just gave me is that you could have some tutorial video but you may also want to answer questions in the page, moving the video around.

MSE v2 - Chrome origin trials

cpn: Developed in the Media WG. A couple of features where developer feedback would be useful to confirm that the API shape is going in the right direction.
… First one is MSE for WebCodecs. There's an origin trial that is ongoing (so far until March). One of the use cases is if you're using HLS video and MPEG-TS, currently, you may have to extract video from the transport stream and mux it to an MP4 container and pass the result to MSE.
… The idea here is that you could take the result of MSE directly and feed it into MSE, where you get the benefit of buffered playback that MSE provides.
… Now is a good time to provide feedback. If you've tried this, we'd be very interested in hearing back from you!
… The second origin trial is enabling MSE to operate in a worker context. This is a feature that should give some performance benefits to sites
… You could leave the user interface responsive by doing the heavy MSE work in a worker.
… It would be interesting to know if any of the common playback libraries (dash.js, shaka player, video.js, etc.) have experimented the feature and seen benefits.

Web Application UI performance on TVs

cpn: One of the hurdles that we heard during TPAC are application UI performances on TVs. In my slide, I link to 3 blog posts that talk about that.
… There's really a trend towards looking at how we apply different technologies to make a much more performant TV User Experience.
… The question for us is: in the IG, what should we look at next in relation to all of this?
… Not everybody is moving to canvas-based rendering. There's a lot of applications that are HTML-based for TV. Perhaps something useful to look at is around gathering some practical developer experience for developing TV applications.
… Are there particular browser features that are problematic?
… That might be something targeted at TV application developers
… Is there something that could be done around testing of applications to see where performance issues arise and give application developers some better metric about the performance of their application?
… Some advanced memory hooks could perhaps be a useful thing to have as well.

Piers: On the performance, it seems that some of these things are going to Rust-compiled code which they say is faster. It would be good to understand what makes things faster. It could be memory management.

cpn: If we're looking at application implementations, the question is what is the role of standardization groups?
… We'll need application developers to come and talk to us about their experience.

kaz: First next step should be looking at what kind of problems users have, and which part of the delivery mechanism have performance issues.
… Then think about how technologies such as WebAssembly could be useful to address performance issues.

cpn: It may be that some of these groups that are looking at alternate approaches, in a sense have success with WebAssembly, solving their own pain points and, from a standardization point of view, W3C has done its work by standardizing WebAssembly.
… The question here for me is: what can the IG be doing? How do we bring the right participation to move issues forward?
… The final thing that I'll say on this topic is that there is a conversation between the chairs of this group and our counterparts in the CTA WAVE Project.
… The discussion looks at how our groups can effectively collaborate.

nigel: I've had people express concern about accessibility of WebAssembly-based approaches. There does not need to be any negative impact for accessibility just because the app is written using WebAssembly.
… But the Web platform now has a bunch of technologies that can be used to re-implement features that have some accessibility baked-in. It's worth noting that accessibility need not be dropped.

ChrisLorenzo: As a web developer, there is a low bar of entry to build an app with HTML/CSS/JS. But then performance hits. LightningJS is canvas-based rendering. And then there's WebAssembly, which requires a larger investment to start with.
… WebAssembly might be the future and perhaps we're just waiting for some companies to show the way.
… And understand how accessibility can be supported, similar to how it is using more regular Web technologies.

cpn: We're at the stage where we're recognizing that this is a topic, and evaluating what next steps we should pursue.
… We need to figure out who we can approach and invite to participate.

W3C/SMPTE Media Production Workshop

cpn: We ran the workshop end of last year.
… We were quite please with the number of talks that were submitted.
… Focus was on professional content authoring applications. It's worth looking at the recordings if you haven't watched the talks already.
… People are really using some of the latest emerging technologies and getting into the details and limitations around, for example, processor instruction set support in WebAssembly.
… There is a draft report available.

Draft workshop report

cpn: We're now wondering what useful next steps we could pursue in the Interest Group.
… The IG would seem like a good hub for that work, to coordinate feedback on whatever specification, be it WebCodecs, Web Audio, WebAssembly, etc.
… We may want to send a number of requirements to WGs.
… I don't think that we've identified major gaps, but always something that we're happy to talk about.
… Once we've got more concrete feedback through the survey and spoken with more participants, I think we'll have a better idea about the level of support for continued activity around this.
… If there is interest, we can have a discussion about whether work should take place in the IG, a dedicated CG or elsewhere.
… There was significant interest during the workshop. That suggests that may be some appetite to continue the exploration.

Workshop survey (meant for workshop participants, but input from anyone is welcome!)

<cpn> Francois: It closes at the end of the week, but we can extend it

<nigel> +1 to extending please

kaz: Originally, the scope of this workshop was casual kind of production, e.g. using iPhones to create content.
… However, we organized the workshop as "Professional Media Production" in the end. So we should think about hardware collaboration as well, how to use devices such as cameras in the broadcasting studios

<Barbara_H> Thanks on the workshop survey link. Intel's attendees were from the PRC. I will ping them to submit their feedback.

kaz: That implies we may want to think about time synchronization among web content, devices, metadata.

<Barbara_H> +1 on not limiting as well the line will blur over time

cpn: I don't think that we want to restrict the discussion to "professional" vs. "casual". Some companies are looking at how they port their applications from desktop to the Web and have lots of practical experience.
… We're still gathering input from workshop participants. Topics can touch different technologies and spaces, including WebRTC, media capture.
… At this stage, I don't have a specific activity that we know we're going to launch around this. We're really looking for feedback and expression of support.

Barbara_H: One potential future topic: metaverse and entertainment.
… More an emerging topic.

cpn: Good point. We had discussions a couple of years ago with the WebXR group.

Next meeting

<Barbara_H> Future Topic Metaverse - https://www.xrtoday.com/virtual-reality/media-and-entertainment-in-the-metaverse-the-future-is-already-here/

cpn: Tuesday 1 March 2022 at 3pm UTC. Agenda is pretty much open for the time being. Please get in touch with ideas!

<kaz> [adjourned]

Minutes manually created (not a transcript), formatted by scribe.perl version 185 (Thu Dec 2 18:51:55 2021 UTC).

Diagnostics

Succeeded: i|Today|-> https://lists.w3.org/Archives/Public/public-web-and-tv/2022Jan/0006.html Agenda for today|

Succeeded: s/baked-in./baked-in. It's worth noting that accessibility need not be dropped.

Succeeded: s/We should/However, we organized the workshop as "Professional Media Production" in the end. So we should/

Succeeded: s/hardware collaboration/hardware collaboration as well/

Succeeded: s/... That/kaz: That/

Succeeded: s/synchronization between/synchronization among/

Maybe present: Barbara_H, ChrisLorenzo, cpn, kaz, nigel, Piers