<scribe> scribenick: kaz
https://docs.google.com/presentation/d/1P1wcomxHlq1jmuU72GqETuqITm6EOgNiOEAsqxP18Q8/edit Chris's slides
Chris: For this call, I want to
present a summary of TPAC from a Media & Entertainment
perspective,
... not just our IG, but also other media related groups.
... I also want to use some of our time today to plan our next
steps for potential Task Forces,
... and topics for future monthly calls.
Chris: (Slide 2) The IG meeting
itself on was on the Monday during TPAC.
... 37 participants were there, a similar number as last
year.
... The morning session was for the IG's own activities, so we
heard updates from the various TV standardization bodies,
... as well as a report from our Media Timed Events TF.
... The afternoon session was for work happening in groups outside
of the IG, where the IG is interested.
... media-related incubations
Chris: (Slide 3) At the IG meeting,
we agreed to publish our use case and requirements doc as an IG
note.
... We heard update on progress from Giri and myself, who are
leading the TF.
... We received approval from WICG to start incubation, i.e., draft
a specification in WICG.
... Giri and myself will be involved, and Eric from Apple was also
interested.
... We decided publish our use case and requirements as an IG Note,
it still needs some work to complete the document.
... So I propose that we continue the TF to complete it, and
continue monthly calls for the Media Timed Events TF.
... suggest you join the call if interested
... Our idea is to use Webkit's DataCue interface as the starting
point, rather than W3C HTML5 DataCue.
... The WebKit DataCue is more generic than HTML5 DataCue, more
flexible as it can take strings or JavaScript objects or arrays,
whereas HTML5 DataCue is only an ArrayBuffer.
... Mark and I discussed offline having a registry based approach
for in-band metadata definitions, for the kinds of events we expect
to obtain from a media stream or media container.
... We also want to bring more interest to implementers, so
interested to talk to people from Google, Microsoft and Mozilla
about this.
... Regarding synchronization of events, in the IG meeting we were
advised to simply raise an issue on the WHATWG HTML spec.
... But, we should write the timing requirements into our use case
and requirements document.
Mark: That's a great summary.
... In terms of timing, the advice was to file pull requests with
spec changes for the HTML spec by WHATWG in this case.
... This is an interesting and positive change.
Chris: I agree, things seem much more open to changes now. I suggest that the TF works on writing that PR.
<Zakim> nigel, you wanted to ask if we know that the requirement is to programmatically specify timing accuracy or have a general lower threshold
Nigel: We need to check the
requirements for what we submit to the WHATWG repo.
... Will we need to submit test cases? Testing synchronization
isn't straightforward.
... Is there a requirement timing accuracy or general lower timing
threshold that we should specify, e.g., 20 milliseconds, for frame
accuracy?
... Or do we want a way for applications to programmatically
specify their required timing accuracy?
... This came up during the WebVMT breakout session.
... What's our preferred option?
Chris: I would suggest we pick this topic up on the next TF call, is that OK?
Nigel: Sure
Chris: There may be different timing requirements for the different use cases that we have.
Glenn: I recommend that we consider
two cases: timing relative to media timeline, and timing relative
to the wall clock for live broadcasts.
... Please take a look at application signalling in ATSC 3.0,
document A/331 available on atsc.org.
... As we make progress, appropriate liaisons expected.
Mark: We can add reference to that to the TF report.
Chris: Certainly.
Chris: (Slide 4) Mark Vickers has been involved, with Louay from Fraunhofer FOKUS.
Mark: CTA WAVE did an API spec, which
identifies the APIs to be implemented for media devices.
... On the basis of that, a test suite has been made
available.
... https://webapitests2017.ctawave.org
... To run the test suite on smart TV and embedded devices, we had
to add improvements.
... Fraunhofer FOKUS did this work, I discussed with Louay and
people from Web Platform Test.
<MarkVickers> sorry. just lost audio. reconnecting.
Kaz: I have also talked with Boaz
Sender and Simon Pieters from Web Platform Tests.
... They were very interested in this effort.
... so I let them know about the URL.
... can introduce them to you (the IG Chairs).
Mark: Our next step for this is to open a pull request, to start the conversation.
Chris: Thank you, Kaz, an introduction is very welcome.
Chris: (Slide 5) One of the major
outcomes from the IG meeting was the proposal to create a new media
Working Group.
... There are a number of specs in incubation at WICG, and many of
the topics are considered ready for standardization.
... These include codec switching in MSE, Media Capabilities, Media
Playback Quality, Media Session, Picture in Picture.
... Also EME related specs: Persistent Usage Record, HDCP
Detection, and Encryption Mode.
... Now, we know that EME is a sensitive issue. So how should we
handle this?
... The outcome from the discussion at TPAC I think was to separate
the EME specs from the other media related specs,
... so that the Working Group could proceed without EME, and then
to discuss EME separately, in particular with potential
objectors.
... W3C Team, and Francois in particular are working on the draft
Charter.
Francois: I should be able to circulate the draft charter next week.
Chris: Thanks, Francois. Any comments?
Mark: As we can continue work on the
specs in WICG, the timing is not really such a worry.
... WICG allows us to write complete specs, and we can have
implementations as well.
... It would take longer to get approval for the WG, but I see no
harm to include EME, I think.
Chris: So, a discussion we may want to reopen is about the scheduling of this possible Working Group.
Francois: Right now I'm preparing a
draft charter without EME.
... We can include EME, but it would require more time.
... It's something we should discuss, I want to get feedback from
IG members.
... Before we move to the AC review, we need to understand possible
objections,
... and engage with previous objectors to find a middle ground that everyone would find acceptable.
... I'm not saying it's not doable but we should be careful.
... I would want to see the response from the IG review, as
discussed at the IG F2F.
Pierre: Does anybody object to include EME for the first draft?
Igarashi: I don't object, we can include it in the first draft.
Mark: I would agree with Igarashi-san.
Pierre: In that case, we should include EME in the initial draft.
[No objections from the group]
Francois: Regarding engaging with previous objectors, my understanding from TPAC meeting is that you were ok to take the lead on that, Mark?
Mark: I am happy to join any
preliminary discussion before the AC review.
... Either way, happy to have a draft before discussion,
... and also happy to have discussion before the initial draft.
Chris: Yes, at TPAC we did discuss having a call with objectors, could be as an IG call or not.
Pierre: Having discussion before the
actual AC review is fine,
... but we need some concrete draft proposal, otherwise the
discussion would be abstract.
Kaz: I would propose a 3-step
approach:
... 1. continuing discussion between the IG Chairs and the W3C Team
to generate an initial draft Charter,
... 2. bringing the initial Charter to the IG and start discussion
within the IG Members,
... and then 3. have public discussion on GitHub and/or mailing
list.
Scott: I met with Google guys to talk
about moving the WICG specs to a Working Group, on the Rec
track.
... We have 3 concrete proposed changes for EME, so we can limit
the scope of the WG to those just those changes, rather than the WG
charter be open ended.
... The other specs in WICG haven't got so much attention, so we
want to create a WG to bring more input, with regular calls to
track progress.
Mark: EME has been implemented for years. I think we can limit the scope to those 3 points.
Chris: Having a proposal to communicate that would be in line with Pierre's proposal.
Kaz: The 3 points are Persistent Usage Record, HDCP Detection, and Encryption Mode?
Chris: That's correct.
... Scott makes a good point about the visibility of the
incubations. We reviewed Media Capabilities on an IG call,
... but haven't really looked in detail at some of the others,
e.g., Media Session or Picture in Picture.
Chris: (Slide 6) Nigel, can I ask you
specifically about the possible new Task Forces we discussed at the
IG meeting,
... i.e., IMSC native browser support, and captions for 360
video.
Andreas: We had some interesting
conversations during TPAC, related to the TextTrackCue
interface,
... and general requirements to for DataCue and text cues.
... I think we should schedule a call to discuss the proposal,
that's my view for IMSC support.
Pierre: I suggest adding a this as a recurring agenda item for the next couple of months.
Chris: We talked about starting up a Task Force in the IG. We're happy to assist where we can, or could be done within Timed Text WG?
Pierre: Subtitle and captions are important for media & entertainment in general.
Chris: Let's keep it as an open item, and we can possibly have a dedicated call on this in a few months time.
Will: Regarding DataCue work for
caption support, it would be strange to have a separate
mechanism,
... so I recommend we bring this together, to have a harmonized
mechanism.
... Also, IMSC native support is attractive.
Andreas: We need to discuss together,
not sure how this relates to the proposed Media WG charter.
... We could come up with some proposal during an IG call.
Chris: We're almost out of time, I have a few topics remaining to cover.
Andreas: Anything needing our input? Suggest taking it to the mailing list.
Chris: It was mostly informational updates, not necessarily needing input. I'll share my slides on the list, so you have all the details.
Chris: The next Media Timed Events TF
call is on Monday, November 19.
... And the next IG call is on Tueday, December 4.
... Thanks everyone.
[adjourned]