JVET-AP0217 AHG9: On FGC SEI message payloadType code points in AVC [R. Skupin, Y. Sanchez, C. Hellge, T. Schierl (HHI)]
The contribution asserts that adding support for the VSEI FGC SEI message in AVC leads to issues with interaction of persistence signalling between the then two payloadType code points for FGC SEI messages. Further an issue with contradicting information in applicable FGC SEI messages is identified.
It is proposed to
- add clarifying language to the persistence related semantics of the FGC SEI message in AVC and VSEI
add bitstream constraints related to the PON-nesting and content of VSEI FGC SEI messages to Annex D.2.1 of AVC. the language needs to be fixed for because payload type is not defined in VSEI.
Agreed to add AVC items to AVC spec, and to add VSEI items to errata as reflected in a -v2 version of the document.
AHG9: Study JVET-AO2032 TuC for VSEI (96)
There was discussion about a table of SEI messages listing those in the TuC.
A Google sheet was created. Proponents were requested to fill it in for their SEI messages, to be concise, and to follow the same order as in the TuC document. [Ed. Where is it now?]
It was planned to discuss this Saturday just after lunch, and try to finalize it by end of Sunday, so is available for JVET plenary on Monday.
Use bug tracking for simple bug fixes.
VSEI: https://vcgit.hhi.fraunhofer.de/jvet/publications/VSEI
TuC: https://vcgit.hhi.fraunhofer.de/jvet/publications/VSEI-TuC
Need to be efficient.
Set a timer for initial presentation for 10 minutes. Multi-part contributions may require more time.
Move to offline discussion when relevant
Only have long discussion when other participants support having more time.
Can defer contributions to later meeting.
Discussion about table describing SEI messages during 1400-1505 on Saturday 25 April 2026.
It was suggested that software availability column should indicate whether processing is implemented or just syntax parsing. Some SEI messages that are descriptive may not need processing.
In terms of timeline for a new standard, it is noted that new versions of the video coding standards that reference it would also be needed.
A participant suggested that a first WD or PWD be output at next meeting rather than this meeting so there is time to make SW available.
At JVET plenary yesterday, it was noted that an ITU-T SG21 meeting is planned for April 2028, and that April 2028 may be an appropriate time for consent of VSEI v5.
HLS Track recommends that we plan to output a WD or preliminary WD for VSEI v5 at the July 2026 meeting
The current spreadsheet table will be included in a contribution document by Monday for discussion in JVET plenary.
It was suggested that proponents bring contributions to the next meeting with proposals of including particular SEI messages in the WD.