[payload] Payload WG notes - Please read

["Ali C. Begen (abegen)" <[e-mail]>]

 
Please have a look and send any corrections to the chairs by July 28th. Presenters, please fill in the gaps denoted by a question mark.

Thanks.
-acbegen

IETF 93 - Payload Notes
Chairs: Roni Even, Ali Begen
Note takers: Stephan Wenger, Mo Zanaty
Jabber scribe: Jonathan Lennox
 

Payload Status Update Chairs
https://www.ietf.org/proceedings/93/slides/slides-93-payload-0.pdf
 
H265: SDP directorate review by Flemming Andreasen noted no overall
grammar and no formal grammar for fmtp parameters. No interoperability issues,
so no change needed, according to Stephan, Magnus Westerlund, others. Magnus:
recommendation in how-to: when you have complex parameters, then you should
include ABNF.  Otherwise not.
Stephan: there is an ABNF in the draft for one complex parameter.
 

VP9 RTP Payload Format (Magnus Flodman)
draft-ietf-payload-vp9-00 <http://tools.ietf.org/id/draft-ietf-payload-vp9-00>
https://www.ietf.org/proceedings/93/slides/slides-93-payload-3.pdf
 
Flexible and Non-Flexible GOF (Group of Frames) structure: Mo thinks
this may be useful for other codecs, authors have no issues with reusing this.
Stephan noted that other codecs (e.g. H.264/5) have other means to convey this
within the payload itself not in payload headers.
 

Flexible FEC RTP Payload Format (Varun Singh)
draft-ietf-payload-flexible-fec-scheme-00 <http://tools.ietf.org/id/draft-ietf-payload-flexible-fec-scheme-00>
https://www.ietf.org/proceedings/93/slides/slides-93-payload-1.pdf
 
Bernard Aboba: How can a sender signal specific FEC patterns, like only
protect the base layer of SRST RTP streams? Authors clarified that no static
signaling (e.g. SDP) is necessary since the sender is free to pick dynamic
patterns and all receivers must handle this.
 
Slide 8: open issue: 
Jonathan: putting SSRC into RTP payload is scary (layer violation)
Mo: we are already in layer violation due to SN; could go Fecframe model
Jonathan: ??? something about Perc, 
fec over encrypted payload
 
Magnus: Dislikes no support for jumbo packets > 4KB. Better to
support 64KB packets and expand the FEC header. Stephan agrees with expansion.
 
Different opinions expressed on whether source flow SSRC(s) should be
included in the FEC header, or other identifiers like MID or the newly proposed
ESID/RSID. Also different views on signaling multiple source flows.
Result: Need a side meeting to decide what source flow identifier is
used to associate with a FEC repair flow and how to handle multiple source
flows. Varun will lead and copy the list.
 

MELPe Codec (Roni Even for Victor Demjanenko)
draft-demjanenko-payload-melpe-03 <http://tools.ietf.org/id/draft-demjanenko-payload-melpe-03>
https://www.ietf.org/proceedings/93/slides/slides-93-payload-4.pdf
 
Stephan has read it and thinks it is a well written RTP payload format
draft. No other reviewers, so chairs will ask the list to review.
 

Interleaved RTP Payload Format (Rachel Huang)
draft-huang-payload-rtp-interleave-00 <http://tools.ietf.org/id/draft-huang-payload-rtp-interleave-00>
https://www.ietf.org/proceedings/93/slides/slides-93-payload-2.pdf
 
Jonathan: Note that PT must be the same.
Rachel: yes
Jonathan: assumption aggregation never changes number of packets.  
Rachel: yes
Jonathan: with more flexible sequence number mechanism, you could do better
Rachel: yes
Imed: Interleave an IDR, what’s the gain?
Rachel: ???