Minutes IETF105: avtcore
minutes-105-avtcore-00

Audio/Video Transport Core Maintenance (avtcore) Working Group

======================================================

CHAIRS:  Jonathan Lennox

         Rachel Huang

Note Taker: James Gruessing

Jabber Scribe: Magnus Westerlund

AGENDA

Tuesday,  23 July at 17:10 - 18:10

-------------------------------------------------

1. Note Well, Note Takers, Agenda Bashing - (Chairs, 17:10, 10 min)

   Status of working group drafts:

      draft-ietf-avtcore-multi-media-rtp-session-13

      draft-ietf-avtcore-rtp-multi-stream-optimisation-12

      draft-ietf-avtext-rid-09

      draft-ietf-avtext-lrr-07

            [RFC Editor Queue: MISSREF]

1. draft-ietf-avtcore-multiplex-guidelines-08 [AD follow-up]

Magnus: Not quite done, would like others to look at questions on the mailing
list

Colin: I looked at the questions on the mailing list, and don't have anything
to add.

Magnus: So... it's over to the AD.

2. draft-ietf-avtext-framemarking-09 [WGLC]

Jonathan: I had asked the author to answer some of the outstanding questions,

which they did not. So, I spoke to the author, and he said he would respond to
it.

Mo: We have some edits to make, a lot of comments need clarifying and explaining

Jonathan: When do expect it to be done by?

Mo: Hopefully by the end of this week

Jonathan: There's been no technical changes?

No: No, there won't been no technical changes.

Bernard: Which scalability modes this is compatible with?

Mo: Do you mean the restrictition of temporally nested several layers?

Bernard: There are some scalability issues with the AV1 codec

Mo: AV1 don't appear to be submitting a payload draft, nor use the framemarking
draft

Bernard: The question about enabling frame marking in AV1 has come up

Roni: I heard there will be no RTP payload for AV1

Jonathan: It's complicated...

Mo: Nobody has been working on a public draft in the IETF for AV1

Bernard: The same question arised for VP9, even if AV1 is out of scope the
question remains

Jonathan: Most payloads already support non-temporal nested layers. This draft
says "if you're encoding in a way that is not compatible with this spec, don't
put frame markings on this".

Mo: In order to be able to understand what is coming in the streams, you have
to have a full dependency tree of all prior frames up until the last I frame,
which is unreasonable to signal

Jonathan: Framemarking just needs to say that it is not capable of expressing
every scalability mode that codecs support, if you are encoding a mode that
isn't supported, you must not put framemarking on it

Mo: I think the question from Bernard is "is that okay? Is that too much of a
constraint?"

Bernard: No, I'm asking for clarification over what supported modes are okay.

3. RMCAT CC Feedback (Colin Perkins, 17:20, 20 min)

draft-ietf-avtcore--cc-feedback-message-03

Colin: We received lots of feedback at Prague, and submitted a -04, where
nothing in the packet has changed

Colin: We added clarification that if FEC and retransmission has been
negotiated, then congestion control feedback is sent for those packets

Colin: We clarified when you're using the SDP bundle extension, the feedback
signalling matches the RTCP FP attributes

Colin: Long discussion if SDP offer provides several different weighting
options, and where possible the receiver should pick its preferred and use it
consistently

Colin: In terms of RTCP, nothing changed but we put in some clarifications, if
you configure your bandwidth fraction weirdly, you may have to spread your
messages over multiple packets and you shouldn't do that.

Roni: Is this a different behavior if don't receive other RTCP packets than if
I don't receive any congestion control?

Colin: This is specific guidance about congestion control

Jonathan: If you receive ??? then you probably have a funny middle-box

Jonathan: Do you expect to have a draft out in the summer before the students
get back?

Colin: Yes

4. Status of Payload WG (Roni, 17:40, 20min)

Roni: Since last meeting, we've had RFC 8627 which was published, thank you Mo

Roni: There is a TSVCIS payload that was sent to publication in February, with
an AD change in March

Roni: Working Group documents includes TETRA, which was strange as it was
submitted by ETSI in the middle of our development cycle, the author claims
they want to proceed, but never did any update

Roni: The JPEG-XS document has comments on the list, and authors will submit
new revision

Roni: The TTML, the media subtype was registered by W3C, we had a WGLC but no
reviews.

Jonathan: I meant to do a refresh of the VP9, but I talked to Justin who
suggested to convert TODOs into Github issues, which is done

Jonathan: Does anyone care about SLI?

Stephan Wenger: I don't care anymore about SLI

Justin: A few on vacation at the moment, aiming for before 106

5. Next steps and open mic - (Chairs, 18:10, 10 min)

Colin: I wanted to finish the meeting with some good news

Colin: Steve Casner and Eve Schuler - will be awarded the IEEE award in 2020
for technical leadership in development of internet and multimedia protocols,
including RTP in this working group