WebRTC WG Virtual Interim

<dom> Slides

[ recording now in progress ]

Introduction

Welcome from the chairs. The group is reminded that we are targetting to request transition to CR before the end of April. To that end we would need to identify features that are at risk and have input from implementors on what feature they are (or not) planning to implement.

Up to the group to decide what criteria to set up for removing stuff due to not being implemented, and present to the director. Will probably be different criteria for "optional" and "non-optional" features.

W3C Process document: Section 6 W3C Technical Report Development Process

WebRTC-PC Pull Requests

Issue 526 / PR 1011: NetworkError is not defined and might not be needed (Bernard)

Decision to go with what the PR proposes (Keep the CloseEvent simple use RTCError for data channels errors).

Issue 1021 / PR 1109: Configurable packetizationinterval (Taylor)

Everyone happy with ptime text/PR.

Decision to accept this PR.

Issue 1085 / PR 1098: RTCRtpContributingSource.audioLevel not guaranteed to be in sync with audio playout (Taylor)

Need to keep the old definition in stats (want time on wire rather than after jitter buffer there). Sync method - be careful. Seems a bit expensive. Otherwise, we're OK with delaying to after playout.

Result: PR can be merged, but J-I to file a new Issue for making get contributing source async.

Issue 1091 / PR 1099: When exactly is an SSRC RTCRtpContributingSource object updated? (Taylor)

Should we always update SSRC? Yes.

WebRTC-PC Issues

Issue 1101: RTCRtpContributingSource naming (Taylor)

Consensus for there always being an SSRC value, but we need a way to differentiate SSRCs and CSRCs values. Consensus to add getSynchronizationSources to go along with getContributingSources.

Issue 1073: Which members in “sendEncodings” are used (Taylor)

Need to identify what the app can change in sendEcodings. Consensus for sendEncodings where all of the non-read-only things can be set.
Need to sort error handling.

Issue 763: Handling of Simulcast Errors (Bernard)

No resolution taken. Bernard will continue to try to come up with a complete proposal.

Issue 1086: Make legacy API optional to implement (Stefan)

Group agrees to make the legacy APIs optional.

Issue 1092: DTLS failures (Bernard)

Seems to be consensus around that we need something for recoverable DTLS errors. We should provide actionable info to script.
Bernard said he would come up with something.

Media Capture Main Pull Requests

Issue 434 / PR 440: Disable user media by default in cross-origin iframes (Stefan)

In current spec, iframes only get to use getUserMedia if the parent allows it. Proposal to use fine grained permissions as defined by the new feature-policy specification.

Dom reported that the feaure policy approach was liked by people that reviewed that spec. The group thought that this spec was quite recent and that we should wait for momentum before changing our current spec.

Stefan will write a 2-pager to describe how we could do the "let audio and video come into an iframe" thing.

WebRTC Stats Pull Requests

Issue 189 / PR 191: Remove isRemote (jib)

Jan-ivar proposal was liked. Harald, who wrote the original text, liked this new version better.
Resolution to accept the PR.