Minutes IETF109: avtcore
minutes-109-avtcore-02

Audio/Video Transport Core Maintenance (avtcore) Working Group
===============================================================

CHAIRS:  Jonathan Lennox
         Bernard Aboba

IETC 109 AGENDA

Wednesday, November 18, 2020 21:00 - 23:00 Pacific Time
Thursday, November 19, 2020, 5:00  -  7:00 UTC
Thursday, November 19, 2020 12:00  - 14:00 Banghok Time

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

1. Note Well, Note Takers, Agenda Bashing, Draft status - (Chairs, 12:00, 10
min)

Brian Rosen will be jabber scribe.
Stephan Wenger will be providing notes in Word form.

Bumping Colin to later as his IRTF chair duties keep him elsewhere.

Stephan: We need guidance on ?

Action item: Chairs to follow up on WGLC for VP9 and frame marking draft.

Action: Chairs to follow up with authors on tetra draft and maybe revise the
milestone.

Action: Authors of rtp-evc draft to submit as a WG draft.

Agenda bash to add discussion of evc draft after the VVC draft.

All times in PST

21:03 Start

Meeting tips

Note Well

Change of presenter for JPEG-XS

AI to chairs for payload-vp9 (send to IESG?)

AI for draft-ietf-avtext-framemarking (send to IESG?)

AI chairs: follow up on tetra: chairs to follow w/ authors and remove milestone
as needed.

EVC adoption: AI authors: resubmit as WG draft

Stephan Requests one minute for EVC payload

2.  RTP Control Protocol (RTCP) Feedback for Congestion Control (Colin Perkins,
12:10, 10 min)
https://www.ietf.org/rfcdiff?url1=draft-ietf-avtcore-cc-feedback-message-08&url2=draft-ietf-avtcore-cc-feedback-message-09

21:50 Presentation

Magnus: ECN problem is now solved, DISCUSS cleared.

Jonathan: as an individual, this is all fine.  WG is fine with it (no objection
on the list).

Colin: if people are OK with signaling, then no further involvement of WG needed

Jonathan: WG is ok with the changes.

Barry: post link to diffs on mailing list.  That was done.

Barry: will give final look and approve it.

Action: Colin posted the diffs, no objections. WG OK with the changes.

3. Multi-party real-time text (Gunnar Hellström, 12:20, 15 min)
https://tools.ietf.org/html/draft-ietf-avtcore-multi-party-rtt-mix
https://tools.ietf.org/html/draft-hellstrom-avtcore-multi-party-rtt-solutions

Presentation

21:26: Daniel Havey: How will this lose 3 work with new packet recovery methods
like RACK? Gunnar: what is RACK?

Daniel: RACK is a new recovery mechanism.  Wants to know interaction.

Gunnar: don’t know. Isn't RACK for TCP, not RTP?

Brian Rosen: will look at v10 how his previous comments are addressed; but
likely read to WGLC.  High priority work, emergency services implementations
waiting.

Jonathan: What was the problem with loss detection?

Gunnar: With the new much more efficient packet interleaving method, it is in
most cases not possible to deduce if loss of multiple packets resulted in real
loss of text.

WGLC will be issued shortly after this meeting.  Main RTP text first,
informational text later or never.

Brian wants opportunity to review 10.  WGLC after that review

Action: Chairs to follow up with WGLC after review of -10 by Brian Rosen.

4. RTP Header Extension Encryption (Justin Uberti, 12:35, 15 min)
https://tools.ietf.org/html/draft-uberti-avtcore-cryptex

21:35 Presentation

21:44 Q: should we adopt this as WG item?

21:45 Harald : Can we lose 1-byte header extensions while we're at it?

Justin: perhaps not

Harald: would not like to see people turning off extensions just to fulfill
conditions of 1 byte header

Jonathan: Let's discuss this on the list.

21:46: Mo: mechanism expects all middleboxes to be aware of this?

Justin: Not everything, packet capture tools and such would work, but if you
try to understand the packet SFUs need to decrypt.  CSRC concerns not that big
a deal.  Can be used in a Point to point encryption system as well. Cryptex is
not end-to end like PERC, its hop-to-hop.

Cullen: The 1 byte header issue is orthogonal to this draft. Let's handle that
separately.

Action: Chairs to issue a call for adoption.

5. SFrame RTP Encapsulation (Sergio Garcia Murillo, 12:50, 15 min)
https://tools.ietf.org/html/draft-omara-sframe

22:00 Presentation

22:06: Magnus: clarification sframe on decoders smallest independent unit.
Relation on RTP depends on things like scalability.  Need to be careful in
terminology.

Justin: notion of “IDU” “independently decodable unit”.

Action: Sergio, et al. to submit a draft on Sframe generic RTP packetization,
as well as SDP negotiation.  (No guarantee that this would be accepted.)

22:16 Colin: payload formats are explicitly designed not to be payload
agnostic.  Declined to standardized codec agnostic payload formats in number of
cases.  If that is to be changed, needs to be done deliberately. Are we going
to have to define Sframe RTP payload formats for each codec?

Sergio: that would be an implementation and interoperability nightmare.

Mo: (not recorded)

Magnus: echo Colin.  Need to support “smallest independent decodable units”.

Sergio: The Insertable Streams API provides access to frames, but application
can parse the bitstream, so as to divide it into smaller units (e.g. NALUs,
tiles, etc.)

Justin: stacking up with cryptex, most header extensions are stuff SFUs want to
see.  But there may be exceptions.

Stephan: .

Jonathan: Most issues need to be addressed in the Sframe WG.

Harald: need an RTP payload format because decoders will try to look into the
RTP payload and fail.  “No user-serviceable parts inside”.  SDP codec
interaction is tricky.

Jonathan: expect something here, work probably in sframe.

Cullen: sframe not chartered for RTP payload formats.  We are changing
architecture here.

Bernard: Will sframe WG define IDUs?

Cullen: maybe, but need hard discussions

6. RTP Payload for VVC (Shuai Zhao, 13:05, 10 min)
https://tools.ietf.org/html/draft-ietf-avtcore-rtp-vvc

22:32: Presentation

Jonathan: SDP payload review process ok, will do the same thing in EVC.

Action: Authors to request WG guidance on outstanding items.

7. RTP payload for JPEG XS (T. Bruylants, 13:15, 10 mins)
https://tools.ietf.org/html/draft-ietf-payload-rtp-jpegxs

One open issue.
Action: after issue resolved and new draft submitted, Chairs to issue WGLC.

This was presented by Tim Bruylants (not S. Lugan)

8. QUIC RTP Tunneling (Samuel Hurst, 13:25, 10 min)
https://tools.ietf.org/html/draft-hurst-quic-rtp-tunnelling

Did not get to this document (slides not submitted, and we ran out of time)

9. Multiplexing Scheme Updates for QUIC (Bernard Aboba, 13:35, 10 min)
https://tools.ietf.org/html/draft-aboba-avtcore-rfc7983bis

Action: Conclusion was reached for WG to adopt the draft. Once adopted, address
Martin's comments.

10. Wrapup and Next Steps (Chairs, 13:45, 15 min)

23:03 Meeting adjorned.