VP9: Framemarking not usable

[aboba]

The following post indicates that framemarking was not usable with VP9:
https://mailarchive.ietf.org/arch/msg/avt/aOufzlzicXOXpP-1s48IrGUIraU/

Section 5,4 Framemarking does not mention these issues:
https://tools.ietf.org/html/draft-ietf-payload-vp9-10#page-16

This section should be updated to do one of the following:

1. Suggest a solution to the implementation problems.
2. Suggest a subset of scalability modes that are usable with framemarking.
3. State that framemarking is not usable with VP9.

[JonathanLennox]

In my opinion this is an issue with framemarking, that it only supports certain fairly restrictive scalability modes (not including K-SVC), not an issue with the VP9 format. Does more than this need to be said?

[juberti]

I agree with Jonathan, although that suggests a dependency for the vp9 document. Is this on the agenda for the upcoming interim?

[aboba]

@juberti Framemarking (implementation experience and WGLC resolution) and VP9 (WGLC comments) are both on the agenda.
As I understand it, one of the reasons framemarking was removed from Chrome was lack of utility for VP8 & VP9, which raises the question as to why a framemarking section is needed in VP9. Could you contribute a slide or two on framemarking implementation experience with VP8 and VP9 (or convince Philip Eliasson to do so)?