Audio/Video Transport Core Maintenance (avtcore) Working Group
CHAIRS: Jonathan Lennox
Marius Kleidl

IETF 122 Agenda
Location: Bangkok, Thailand
Session: III
Room: Boromphimarn 3

Date: Thursday, March 20, 2025
Time: 15:00 - 16:30 Bangkok time

IETF 122 info: https://www.ietf.org/how/meetings/122
Meeting link: https://meetings.conf.meetecho.com/ietf122/?session=33886

Notes: https://notes.ietf.org/notes-ietf-122-avtcore
Slides:
https://docs.google.com/presentation/d/1E_XfsZbcSMyNl-p3L8jUk2HkD1hfqY35TPzr0aXcV0w/

Notetakers: Youngkwon Lim, Mo Zanaty

Preliminaries (Chairs, 15 min)

Note Well, Note Takers, Agenda Bashing, Draft status

Jonathan has presented Note Well, Participant Obligation highlighing IPR
disclosure, Code of Conducts, Meeting tips for In-person participants
and remote participants, Meeting resources, new co-chair Marius Kleidl
and honoring the former co-chair the late Bernard Aboba.

Jonathan has explained the status of the drafts.

Magnus has commented about ongoing work on RTP paylaod format registry
with mediaman.

Richard has expressed interest on the draft on sframe. Jonathan has
clarified that the draft will be parked for the moments and unparked
whenever the folks are ready to work on it.
Chair action: Park sframe.

Gorry noted that meeting slides are missing in the materials.

RTP Payload for V3C (Lauri Ilola, 5 min)

https://datatracker.ietf.org/doc/html/draft-ietf-avtcore-rtp-v3c

Lauri has presented the status of the draft.

• 2nd WGLC is until 3/27.
• All issues has been closed since the first WGLC.
• One new issue regarding IANA feedback has been opened.

RTP over QUIC (Mathis Engelbart, 10 min)

https://datatracker.ietf.org/doc/html/draft-ietf-avtcore-rtp-over-quic

Mathis has presented the status of the draft

• one issue regarding IANA review will be merged after the meeting
• interop tests are ongoing

Mathis has questioned about going to WGLC. Jonathan has clarified that
it was the previous agreement that SDP work needs to be finished before
WGLC but he now thinks it can be parallel.

SDP Offer/Answer for RTP over QUIC (Spencer Dawkins, 15 min)

https://datatracker.ietf.org/doc/html/draft-dawkins-avtcore-sdp-roq

Spencer has presented the progress of the draft.

• updates on use of the SDP parameters
• Non-Normative QUIC/RTP/AVPF example
• updates on security considerations
• Open Questions Identified in the Editor's Version
• Feeback from Harald (consider possibility this is out of scope) and
  Jonathan (unsure how to express UDP/RTP vs QUIC/RTP in SDP without
  capneg which is dead)

Harald has commented that flow-ID and BUNDLE are at completely different
levels

Jonathan has commented that use of single flow-ID for a BUNDLE group
would be one way to go.

Harald questioned if anyone is not multiplexing RTP and RTCP. No
responses.

Spencer has explained that the authors are planning to produce -01
version before the AVTCORE interm with the request for WG adoption poll

Spencer asked if anyone else was interested in providing early review
comments, and Gurtej Chandok raised his hand.

RTP Payload for Haptics (Hyunsik Yang, 5 min)

https://datatracker.ietf.org/doc/html/draft-ietf-avtcore-rtp-haptics

Hyunsik has presented the updates of the draft

• updates in SDP sections
• the authors believes that the draft is ready for WGLC

RTP Payload for V-DMC (Hyunsik Yang, 10 min)

https://datatracker.ietf.org/doc/html/draft-hsyang-avtcore-rtp-vdmc

Hyunsik has presented the draft

• updates addressing comments
• updates regarding MPEG specification
• inviting people to review and participate

Hyunsik does not think it is reday for WG adoption yet

RTP Payload for APV (Youngkwon Lim, 10 min)

https://datatracker.ietf.org/doc/html/draft-lim-rtp-apv

Youngkwon presented goals of the draft and project. ISE is starting pub
review of version -03. New bitstream structure with PBUs and access
units, similar to other standards.

Mo asked if the marker bit applies to the entire AU or each layer.
Youngkwon clarified it applies to the entire AU.

Mo asked about ISE submission vs AVT WG RFC. Youngkwon clarified the
codec bitstream will be an ISE submission while the RTP payload format
will request AVT WG adoption.

Gurtej asked how to know when a tile is complete to decode it. Jonathan
answered when the fragment counter reaches 0.

Mo asked if the fragment counter results in high buffering and latency
at the encoder to know the number of fragments. Youngkwon clarified this
is often known up front before encoding, using filler to hit the known
target.

Automatic Corruption Detection (Erik Språng, 10 min)

https://datatracker.ietf.org/doc/draft-sprang-avtcore-corruption-detection

Erik has presented the draft

• It would be useful if we could actually detect the error at the
  decoder immediately
• RTP header extension to indicate the allowed error thresholds and
  std deviation
• encryption of the header extension is vital
• currently implemented in Chromium M132+

It has been clarified that the intention is to detect the bad decoder
implementation instead of error in the received bistream

Stephan asked if a new payload header or separate payload would be
better than a header extension. The author stated a header extension was
easiest to implement.

The authors considers WG adoption at the coming interim.

LTR feedback (Erik Språng, 5 min)

https://datatracker.ietf.org/doc/draft-sprang-avtcore-frame-acknowledgement

Erik has presented the draft

• Header extension asking to send RTCP feedback about the decoding
  status of a particular video frame
• existing message formats are either packet level or codec speicific.

Magnus has question if this can be extended to tile level indication

Gurtej has commented that using frame ID is a good approach as timestamp
cannot specifically identify a frame if there are multiple layers.

Mo has commented that there are some vendor specific implementations but
no standard way.

If time permits: RTP Payload Format for Geometry-based Point Cloud Compression (Mathis Engelbart)

https://datatracker.ietf.org/doc/draft-engelbart-avtcore-rtp-gpcc/

Mathis has presented the draft

• G-PCC compression is standardized by ISO/IEC23090-9
• The authors welcome review and feedback

Wrapup and Next Steps (Chairs, 5 min)

Jonathan has noted that there might be an interim in half way between
the IETF123.

Marius has listed the action items:

• Park draft-ietf-avtcore-rtp-sframe for now. Richard Barnes indicated
  potential interest in a few months. Document this on the mailing
  list as well
• Initial WGLC for draft-ietf-avtcore-rtp-over-quic. We still want to
  wait for more interop results and maybe the SDP work before a final
  WGLC
• No call for adoption for draft-dawkins-avtcore-sdp-roq or
  draft-lim-rtp-apv yet
• Start WGLC for draft-ietf-avtcore-rtp-haptics