W3C

– DRAFT –
MEIG monthly meeting

04 June 2024

Attendees

Present
Chris_Lorenzo, Chris_Needham, Ewan_Roycroft, Francois_Daoust, igarashi, John_Riviello, Kaz_Ashimura, Tatsuya_Igarashi
Regrets
-
Chair
ChrisL, ChrisN, Igarashi
Scribe
cpn, kaz

Meeting minutes

Agenda

Slideset: https://www.w3.org/2011/webtv/wiki/images/c/cf/2024-06-04-meig.pdf

[Slide 2]

cpn: launching web applications on TV devices (follow-up)
… Issue #107 Channels 2.0
… Issue #4 Frame accurate seeking of HTML5 MediaElement
… Volumetric media
… TPAC 2024 Plans
… anything else?

Launching web applications on TV devices

[Slide 3]

cpn: What do we do for the next step?
… It seems we need conversations with manufacturers, but have not had much response yet.
… We got response from one manufacturer, suggesting to have some more comprehensive documentation
… A lot of this have been done already in the MEIG repo.

Developing web applications on consumer products

cpn: Would hear from you about suggestions, and contacts you may have at manufacturers.

cl: We have the video too.
… I'm happy to join any meetings, also happy to work through the discussion.

kaz: Just to make sure, the "charter.md" is a charter for a possible TF, right?

cpn: Right

kaz: In that case, the title of the document should say "some TF" explicitly
… also we should make an official call for participation to the whole IG to see who will work on this topic.

cpn: Chris Lorenzo, let's continue to talk with the manufacturers directly about web apps on consumer devices.

Channels 2.0

[Slide 5]

Issue 107 - Channels 2.0

cpn: This is a proposal on GitHub, covers a lot of use cases, e.g, content related to TV channels, interactive experiences, homapages
… having sub-channels in addition to channels
… and then proposes an XML format to describe

kaz: It's kind of big and need some more clarifications

cpn: Yes, would need analysing for different components. Channels aren't really in scope at W3C, it's more for TV industry specific groups

<Zakim> fd, you wanted to react to kaz

fd: We previously reviewed schema.org metadata, seems related to this.

fd: Also, as an idea, could look at what data is surfaced through Media Session API in the UI

johnriv: I did skim this, it comes from an end user perspective, but to move forward needs to be a pain point for manufacturers or streaming services

Frame accurate seeking of HTML5 MediaElement

[Slide 6]

Issue 4 - Frame accurate seeking of HTML5 MediaElement

cpn: This is a long discussion thread. It covers a number of topics and use cases:
… frame accurate rendering, seeking, single-stepping of individual video frames
… essentially needed for non-linear video editing applications. Can be either editing directly in the browser, or using as a proxy to cloud-stored media.
… I've had some conversations recently about browser based video editing applications, with BBC colleagues but also other companies
… BBC R&D created a frame-accurate seeking video player using WebCodecs. Some complexity, particularly with managing buffers of encoded video and decoded VideoFrames.
… Francois presented frame accurate seeking gaps at TPAC 2019,
… and quite a lot of things have been changed since then.

[Slide 7]

cpn: We now have HTMLVideoElement.requestVideoFrameCallback(), and there's an example showing how to use it to seek to specific frames, using timestamps. But not always accurate.
… DOM Event timeStamp is now DOMHighResTimeStamp, although this is a "coarse" timestamp, so I don't know if there are still issues.
… The interest in high resolution event timestamps came from multi-device synchronization, also enabling synchronisation within the same client.
… Improvements to TextTrackCue event timing in Chromium, resulting from our work here on Media Timed Events.
… WebCodecs is the major new feature, it enables decoding to VideoFrame objects, lets you build a player that renders to a canvas.
… WASM enebles cross-compilation of libavformat/libavcodec. Allows use of professional codecs that WebCodecs doesn't support.
… Improvements to TexTrackCue event synchronisation in Chromium. It's not frame-accurate but much more accurate now.
… With WebCodecs, how to manage the memory of all the decoded data, it's still challenging to build a player for non-linear video editing.
… Web Audio AudioWorklet is good for audio rendering.
… Overall it's all possible but quite difficult.
… If there is sufficient interest, happy to deal with any of these issues again.

[Slide 8]

cpn: What still remains?
… HTML issue whatwg/html#3754 is about negative playbackRate not supported across all browsers. The use case is from NRK of a web page including video and images,
… and responds to the user scrolling the page. The video should play backwards if the user scrolls up the page.
… But because that is not well-supported, the page needs some complex code to update the current time.
… Anybody interested should follow the link.
… Something Nigel pointed out before is audio seeking in MP3 is not accurate. I also found this with my open source waveform visualiser library.
… Needs more investigation for the possible resolution?
… Then, on 360/VR video rendering, there's this Chromium bug on synchronising the video element state with WebGL,
… I don't know the details, but some need for identifying the frame in the GL frame buffer.
… Finally, there is SEI timed metadata events. Is it enough for web apps to handle these, or should we expect WebCodecs to parse them and attach them to VideoFrame objects?
… There is an Explainer from @leonardoFu that describes their needs.

kaz: So you're reviewing some interesting topics for this year in MEIG, potentially.

cpn: Yes, if there's interest we can start discussions about those topics.

Francois: You mentioned rVFC, it's still an incubation, but supported everywhere except Firefox. It should have moved to a WG, but hasn't yet.
… It shouldn't be in incubation any more, so worth investigating that

kaz: which WG do you mean?

cpn: Potentially it goes to WHATWG, or maybe Media WG if it's in the charter as a potential deliverable. Media WG will need to discuss.

Volumetric media

[Slide 9]

cpn: Metaverse Standards Forum (MSF) has a WG on Volumetric Media Interoperability
… discussion between multiple SDOs and industry forums
… There's an initial meeting next week, 14:00 UTC on Thu. 13 June 2024
… and an invitation to join. MSF membership required to participate, but there's no cost.
… I will send an invitation with more details to the mailing list.

kaz: What is the Membership mechanism?

cpn: You need an MSF membership but it's free. Also W3C is one of their Members.

Kaz: If people are interested in that discussion, they're encouraged to join the forum?

cpn: Yes

TPAC 2024 plans

[Slide 10]

cpn: TPAC is 23-27 Sep in Anaheim, California, US
… Here is the proposed schedule.
… MEIG: Monday 9-10:30, 11-12:30; Thursday 16:30-18:00
… Media WG: Monday 14-16; THursday 11-12:30, 14-16, 16:30-18; Friday 9-10:30
… Timed Text, Audio and Second Screen will also meet
… if you want to present anything, please let me know and we'll be preparing within the next few months.
… During the morning session, I expect we'll talk about next generation audio (also in previous TPACs), but there should be time for other agenda items.

kaz: On Monday, 23 Sep during the MEIG meeting, right?

cpn: Right
… Please let me know if you have any suggestions, and we'll see how to include them.
… that's it
… any comments, questions?

(none)

[adjourned]

Minutes manually created (not a transcript), formatted by scribe.perl version 196 (Thu Oct 27 17:06:44 2022 UTC).