Re: [avtext] Frame marking & VP9 SVC

[Jonathan Lennox <[e-mail]>]

Sergio — the updated VP9 payload draft I submitted Monday includes a section describing its usage with frame marking.

Let me know what you think?  It changes the S and T fields to reflect the latest VP9 and frame marking drafts, but it does not include the P or U bits.

I think the intention is that frame marking’s B bit will provide a coarser-grain version of what P and U do (and the “don’t use it with weird structures” requirement will make that sufficient, though that may need to be made more specific).  What do you think?

On Jun 28, 2017, at 10:23 AM, Sergio Garcia Murillo <[e-mail]<mailto:[e-mail]>> wrote:


Yes, at least in VP9 SVC you can define how many temporal and how many spatial layers do you want simultaneously in the encoding. Note also, that VP9 supports quality layers as spatial layers without any resolution change, so the term "spatial layer" is used to represent both spatial and quality layers.

BR
Sergio
On 28/06/2017 16:18, Bernard Aboba wrote:
Sergio said:  "So, shouldn't the LID reference to the spatial layer ID only and omit the quality layer id completely? Also, the spatial layer id is 3 bits on that draft."

[BA] Question:  In VP9 (or AV1, for that matter) is it possible to use both spatial and quality scalability in the same encoding?

On Wed, Jun 28, 2017 at 5:35 AM, Sergio Garcia Murillo <[e-mail]<mailto:[e-mail]>> wrote:
Hi all again,

Did you had a chance to review the changes to add the VP9 SVC spatial layer information into the LID of the frame marking extension? (at https://github.com/murillo128/draft-berger-avtext-framemarking/commit/35cdb229108251f12bc1cdf77909df83a51d3994)

Also, FYI, I have submitted a CL to libwebrtc to support framemarking extension. Currently it supports VP8, H264 and non-SVC VP9, but I would love to tackle VP9 SVC as soon as we get an agreement on the changes to the LID:

https://codereview.webrtc.org/2954503002/

Best regards
Sergio


On 24/04/2017 11:28, Sergio Garcia Murillo wrote:
On 18/04/2017 17:40, Jonathan Lennox wrote:
One of the goals of Frame Marking is that an SFU can be agnostic to the details of the layering, and just treat layer IDs as opaque layer identifiers.  Putting non-layer bits into the LID field would break this goal.

At the IETF meeting in Chicago, both Bernard Aboba and I thought that frame marking might need enhancements in order for spatial scalability to be fully supported. I’m adding him to the Cc, as well as my co-authors on the VP9 payload.

Sergio, if you want to send a pull request to contribute text for the VP9 spec (which would be welcome), it’s in Github at https://github.com/juberti/draughts .
Currently the VP9 LID mapping is described on the frame marking draft, and not in the VP9 draft. We could move the full description from one text to another, but at least for me, it was easier to provide a small patch for current draft to address my comments:

https://github.com/murillo128/draft-berger-avtext-framemarking/commit/35cdb229108251f12bc1cdf77909df83a51d3994

I have not found the .xml in any github repo, so just downloaded the latest draft and applied the changes on top of it.

Please, let me know what are your views about it.