- From: François Daoust <fd@w3.org>
- Date: Wed, 15 Feb 2023 07:06:52 +0000
- To: "public-media-wg@w3.org" <public-media-wg@w3.org>
Hi all,
Minutes of yesterday's Media WG call, focused on rechartering plans, are
available at:
https://www.w3.org/2023/02/14-mediawg-minutes.html
... and copied as raw text below.
I will update the draft charter and circulate a new draft. Feedback is
also welcome any time (but as soon as possible) on
https://github.com/w3c/charter-media-wg/issues
Thanks,
Francois.
-----
Media WG meeting
14 February 2023
[2]Agenda. [3]IRC log.
[2]
https://github.com/w3c/media-wg/blob/main/meetings/2023-02-14-Media_Working_Group_Teleconference-agenda.md
[3] https://www.w3.org/2023/02/14-mediawg-irc
Attendees
Present
Alastor Wu, Bernard Aboba, Chris Needham, Dale Curtis,
Eric Carlson, Francois Daoust, Frank Liberato, Gary
Katsevman, Greg Freedman, Jer Noble, Joey Parrish, Mark
Watson, Sushanth Rajasankar, Tommy Steimel, Youenn
Fablet
Regrets
-
Chair
Chris
Scribe
cpn, tidoust
Contents
1. [4]Welcome and AOB
2. [5]Media WG rechartering
3. [6]WebCodecs #270 Support per-frame QP configuration by
VideoEncoder
Meeting minutes
Welcome and AOB
cpn: Main item today is to talk about re-chartering. Group is
chartered on a 2-year basis. Time to review the scope and
milestones.
… Any technical issue that we should also look at?
dalecurtis: An issue on WebCodecs, fine if we don't get to it
this time.
Media WG rechartering
cpn: [7]Current charter is set to expire at the end of May. A
[8]draft charter is available for next period.
[7] https://www.w3.org/2021/07/media-wg-charter.html
[8] https://w3c.github.io/charter-media-wg/
cpn: A number of questions are outlined in the agenda.
… Starting with [9]issue #22 to look at milestones and
understand where we are.
[9] https://github.com/w3c/media-wg/issues/22
cpn: Goal is to review deliverables in turn.
… Starting with Media Capabilties, we are perhaps looking for
someone who can join as editor.
dalecurtis: Will departed Google. Johannes who worked on WebRTC
might be interested. I'll get back to you.
cpn: I'm wondering how far away we are from moving the spec
from draft to candidate recommendation status. That means that
we have some tests in place and interoperability demonstrated.
bernard: I don't think we're far from it. The functionality is
pretty much done, some tests may be needed.
cpn: I'm really looking at milestones estimates in the charter.
Some rough estimation of where we will be in the next
chartering period.
… CR in second half of this year for instance? And then
Recommendation status mid next year? Is that reasonable?
bernard: I note that you don't have to dump all requirements to
move to CR. They can be addressed later on.
dalecurtis: There's still some questions to answer on Media
Capabilities integration with WebCodecs. We can probably answer
these questions in the timeline you suggested, and then it
depends on what we decide.
cpn: For the purpose of rechartering, that's as deep as we need
to go.
… Picture-in-Picture, we have an issue open to [10]track
progress to CR.
[10] https://github.com/w3c/picture-in-picture/issues/184
Frank: I can check with Francois (Beaufort) and report to the
group on what may still be blocking.
cpn: OK, similar kind of timescale.
… Media Session.
Youenn: Some parts are more mature. Camera and microphones are
less mature. If we want to go fast, we could go with more
mature features.
… The WebRTC bits may be more time consuming.
… If we want to split the work, we can move to Candidate
Recommendation pretty quickly.
tidoust: We used to be stricter on CR, so once published a CR
you can publish other drafts in the meantime
… you can change the scope later, so OK to start with initial
features and add others later on, both ways would work
youenn: Maybe we should schedule a quick discussion with Tommy,
and see where we want to go.
cpn: In terms of timescale, it's probably slightly longer than
the first two deliverables, given that some features are still
further along.
… Next on the list is Autoplay Policy Detection.
… We've just gone through First Public Working Draft, and
completed horizontal reviews.
alwu: We filled horizontal reviews, still expecting some of
them.
… TAG review is done.
cpn: In terms of implementation timeline, what do you expect?
alwu: Current plan in Firefox is to release in next version
112. I also saw positive feedback from Apple. I don't know
about others.
cpn: Perhaps a similar timeline as Media Session.
cpn: Next one is Media Source Extensions (MSE).
… We have an updated draft. Some work in progress with regards
to worker support. Not sure it fully landed.
… Question is: are there changes that we might want to see in
scope for the next charter period?
markw: I cannot tell whether there are things that still need
to integrated relative worker support. I'm happy to help but
cannot take sole editing role on the spec though.
dalecurtis: Matt left. My understanding is that there is not
much that remained spec-wise for ongoing features. One open
question was around the definition of a stricter MSE for use in
embedded devices.
cpn: I'd like to see progress there.
markw: Do you know whether there is a PR for that already?
dalecurtis: I don't think there is one yet.
cpn: In terms of editing, if Matt is not able to continue, do
you have thoughts of who could step in?
dalecurtis: Not at this stage.
cpn: The other thing that we talked about was integration with
WebCodecs.
dalecurtis: We built a prototype in Chrome and no one used it,
so we sort of abandoned that path for now.
cpn: We can probably leave it in the charter.
dalecurtis: I still think it's a good idea, but maybe we're not
there yet.
cpn: Yes, it might be a post-v2 feature. We can be flexible on
how we manage that.
… It makes the timeline discussion a bit more tricky to answer.
dalecurtis: I'm happy to find time to help Mark sorting issues
out.
markw: There is only one PR opened, [11]related to WebCodecs.
Question would be: are there issues that would require
specification work?
… I haven't been following the issues closely for a while.
… It may be worth doing another pass through them.
… 30 issues labeled as v2 issues.
[11] https://github.com/w3c/media-source/pull/302
cpn: We can do that separately from this call.
markw: We could estimate remaining spec work to better evaluate
whether we need another editor recrue.
cpn: OK, we may organize a separate call then.
cpn: Moving on to Encrypted Media Extensions (EME).
… I emailed editors this morning to suggest an update.
… When we met at TPAC, we suggested we would aim at
[12]publishing a FPWD with the main updates. We're coming to
the end of the charter period.
… Is doing something fairly soon feasible for everybody?
… The concern with the EME spec is that it has been in our
charter since the creation of the Media WG but we've never
published the spec, so Advisory Committee may look into this
and wonder why the spec is still in our list of deliverables.
[12] https://github.com/w3c/encrypted-media/issues/501
joey: It's been my intent to get something done but my
availability has been low.
gregfreedman: I can do HDCP detection in the next few weeks.
Can we do the easy part and work on the API to discover
existing sessions later?
joey: I'm willing to drop that. I don't think that's critical.
… That was useful but no concrete agreement and can do without
it.
cpn: FPWD triggers a call for patent exclusion, so ideally
features we want in should be in already.
… But if there's overall consensus that we can drop that
feature.
joey: The feature has not been implemented, and I don't know of
any concrete plan to have it implemented.
cpn: The charter expires end of May. I we can get a FPWD
published by then, that would send the right signal.
… Otherwise, we'll have to say FPWD will be published Q2 or Q3.
joey: I agree.
gregfreedman: I agree as well.
cpn: Happy to organize meetings as needed!
cpn: Moving on to WebCodecs
… A lot of activity. Main spec, a set of registries. What do we
think in terms of timescale?
… Chromium is furthest ahead in terms of implementation.
… It would be good to know about other plans. It feels to me as
longer timescale.
Youenn: Form WPT side, WebCodecs video is the interop scope
this year. Broader interop next year. Nothing we can share from
an implementation perspective.
tidoust: I note that you mentioned having implementation plans
before going to CR, CR in theory is a call for implementation,
so nothing prevents moving forward without that, so long as the
group agrees on features.
… Publishing as CR triggers horizontal review and call for
patent exclusion, so useful for determining scope
cpn: and the PR stage is where the interoperability step comes
in
bernard: I find it useful to label issues as CR-blocking.
… That gives you a good list of priority and estimation of work
that remains.
cpn: That may be something we could do in a future call,
indeed.
cpn: Audio Focus is our newest spec. In the current charter,
it's listed as potential normative deliverable. In the new one,
it will be a proper deliverable.
… There was some discussion on naming, I'd like to review that
now.
… Youenn suggested to rename into "Audio Session".
… I didn't hear strong concerns regarding that name.
tommy: personally, I'm confortable with Audio Session.
cpn: OK, that sounds good.
… Time to FPWD: later this year.
youenn: We discussed with Alastor already about the explainer.
It would be good to get feedback on it from Google as well.
tommy: I'll take a look.
cpn: Next is Media Playback Quality. We've agreed to integrate
into HTML, but needs someone to pick up the editorial work.
tidoust: WHATWG would also need to see tests, and we may not
yet have good test coverage in WPT. A couple of tests should be
enough for that spec.
cpn: Don't need to decide who does it now, but we'll leave it
in the charter until the work is done.
cpn: Finally DataCue and Sourcing In band Media Resource Tracks
specs. Should we leave these in scope as potential
deliverables, is there implementer interest for DataCue?
ericc: We're interested in DataCue, yes. Sourcing In band
tracks is not so useful.
cpn: DataCue is in WICG, we welcome help and effort to move
that along
… Sourcing In band tracks, it's referenced from HTML but has
things never implemented, so could use tidying up
… But nobody is really pushing to update it, so propose we
don't include it in the charter. There was a CG that created
it, no longer active. Anyone who does want to work on it is
welcome to do that in future.
WebCodecs #270 Support per-frame QP configuration by VideoEncoder
dale: I wanted to flag this issue for offline discussion:
[13]w3c/webcodecs#270
[13] https://github.com/w3c/webcodecs/issues/270
cpn: so it could go in the main WebCodecs spec, for all codecs,
and interpretation depends on the codec-specific registration,
or could be specced independently for each codec
bernard: [missed comment related to how this relates to
encoding optimisation when background blurring is in use]
markw: All modern codecs use quantisation, so might have a
parameter, but its meaning might vary, or in future there might
be multiple parameters, so would want to design to be future
proof
dale: We have VBR and CQ modes where it's automatically
adjusted
… No need to conclude here, please weigh in offline
[adjourned]
Received on Wednesday, 15 February 2023 07:06:57 UTC