Meeting minutes
TPAC agenda planning
CN: We are planning 2 meetings during TPAC. The media group in meeting Monday. Then we are joint a group meeting with WebRTC Friday.
CN: going to focus on the Media WG meeting here. The link above is the issues we 'd like to focus on on the meeting. We have a lot of specs at WD state, and when can move some of those to CR etc. Of course, I don't want to focus the administrivia during the meeting, so want to make sure we have time to cover technical aspects
CN: So trying to figure out who is traveling there physically.
Marcos: We can also check registrations to know who's coming
MC: I think we can check the registration page to see who is actually attending.
https://
CN: in addition to issue 40 above, apart from allocating equal time to each spec, I've started putting together a slide deck about the various important issues for the group. (link above)
CN: it would be helpful for folks folks to have a look, and see if I've captured the right issues. Suggestions welcome
PY: it would be important to talk about web codecs. WebRTC has more codecs support. But WebRTC has a lot of codecs that seem to have limited use. It would be great to have common ground across the two groups
BA: that's a pretty major topic, but it's definitely a topic for the joint meeting
CN: are there anything there that are not WebRTC specific, as those are joint meeting specific
BA: we need to talk about the privacy issues, that affect both specs.
CN: We need to also talk about media source extensions. Are there any interop issues that we want to include. I have a list from Matt that should review and see what is unresolved.
CN: There are actually a lot of media capabilities issues. We need to triage those.
CN: Things I've pulled out are things like transitions and how it interacts with EME, there are some privacy related parts, alignment on HDR with web codes, and capabilities for WebRTC .
CN: Alistor, thanks for sending suggestions Autoplay policy detection
CN: those things might not need a lot of time, and might be resolved via GitHub
CN: we would like to some input with regards Media Session
CN: apart from Marcos, I don't know if anyone else is involved with media session
MC: I just started looking at it yesterday, I'm hoping to have gone through the issues today or tomorrow
CN: I'll get in contact with the relevant Chrome folks who have been working on that
CN: Andy, I put some time in for the media session coordinator
CN: Andy, if you are gong to join remotely. Does that work?
AE: It should work
CN: are there any other things from folks on the call that we should cover?
CN: otherwise, Marcos and Chris (Chairs) will drive the agenda
BA: would like to discuss the impact of machine learning on codecs
BA: I don't know if we can make any definitive statements about them
CN: particularly on the encoding side
BA: There is concern about how compute intensive they are. There is a lot of pre/post process. There are some fairly extreme hardware requirements. This is all very new.
CN: it could be something interesting to discuss this
BA: I know some people... I can ping them and see if they might show up
EZ: Browsers have capabilities to convert between color spaces, and web codecs doesn't really support that, so people might end up using Web Assembly to achieve this. What we could discuss what's the best way to do the conversion between different pixel formats.
CN: depending on how this pans out, we don't need to cover all the specs.
CN: Marcos, anything else from you
MC: no
Managed Media Source
CN: the thing that I put in the agenda was to discuss the privacy concern around the to quality change attribute + event
CN: It would be good to understand what the privacy concerns.
JN: some background on why we put this spec together, we didn't try to do bandwidth detection, if the device is under strain, then how does the device choose. The OS knows better than page what the state of the phone is (e.g., how hot is the phone). That was the intent, but we didn't try to implement . In terms of privacy, yes, it does add entropy.
JN: It would be hard to specify the quality change in a way that is is useful, but also not increase the fingerprinting. So we understand the privacy concern. Given the current experimental experience, we could demote this (quality indicator) out of the spec... and treat it as a separate issue?
DC: You would keep onstartstreaming/onstopstreaming?
CN: I think breaking up in this way sounds like a good approach, so we can figure out what we agree or disagree on. As the thread is getting long, it might be good to spin up new issues. But still send a PR for the parts we agree.
JN: is Paul part of the group?
CN: Yes, and hopefully he will one at TPAC
JN: it is his intent to do a PR
CN: there are bits of feedback... like on bufferchange event, like could it be applied to media source. There is also about append buffer rejections. There are concerns about the complexity that introduced. So to the extent that we can separate these things out.
CN: it would be great if we could land some parts.
JN: I will ask JN to split the PR into pull requests on things we agree on.
CN: Anything else regarding managed media source generally?
CN: that feels like a good next step
CN: that was all I had for the agenda for today. Any other topics we want to discuss today?
CN: I was just rereading the comments from Paul...
JN: would like to explore media capabilities more. For audio, people have hacked their way around spatial audio capabilities.... we've hit a wall where we had assumed that media capabilities don't change for the life of the page... but that's not nessearily true. We should have a larger discussion around media capabilities have changed.
BA: this has come up in other contexts.
JN: there are privacy issues that come up... like a site detecting that video is being rendered in other window etc.
CN: it feels like something we should make time to discuss
CN: I don't have anything else for today. I suggest we wrap up here.
BA: some of these topics require significant preparation. A lot of folks are on vacation. We should discuss who is signed up to do what.
CN: should we try to do that now?
BA: we could do it over email probably
Media WG and WebRTC joint meeting
CN: the comparison of functionality
BA: I have someone in mind, but can take some other volunteers. There is a lot to cover over two hours.
BA: we can coordinate over email, figure out who is signed up for what. Some might require some testing across UAs, which will require a few weeks worth of work
CN: happy to pick this up over email with you and others
CN: We will do the same with the Media WG meeting, not just for WebRTC joint meeting
CN: Meeting adjourned.
CN: See some of you at TPAC! Will be in touch to prepare