Open Bug 1477548 Opened Last year Updated Last year

getStats: only a single outbound-rtp for simulcast

Categories

(Core :: WebRTC, defect, P2)

defect

Tracking

()

Tracking Status
firefox63 --- affected

People

(Reporter: fippo, Assigned: ng)

References

Details

steps to reproduce:
1) go to https://fippo.github.io/simulcast-playground/firefox
2) observe the local bitrate graph at the top

The local bitrate is the sum of all outgoing bitrates and pc1.getStats() only has a single report with type=outbound-rtp. I would expect three outbound-rtp stats with the proper ssrcs each.

This is probably from the underlying webrtc.org engine but unlike https://bugs.chromium.org/p/webrtc/issues/detail?id=9547 the framerate is correct ~30fps.
Nils, who should have a look at this one?
Flags: needinfo?(drno)
Nico, can I ask you to have a look at this please?
Rank: 15
Flags: needinfo?(drno) → needinfo?(na-g)
Priority: -- → P2
https://github.com/w3c/webrtc-stats/issues/318 spec issue here. I suspect what hbos says about sender stats vs rtp stream stats makes sense.
Nils, certainly.
Assignee: nobody → na-g
Flags: needinfo?(na-g)
Flags: needinfo?(na-g)
Apologies for the late response, I may have accidentally deferred the NI. We only report stats for the "primary" stream. It looks like the substream stats will be available in the Call API after the webrtc.org 64 merge lands. However there are still some open spec issues [0][1], and I suspect that we will need to create some upstream fixes. (Also see the chromium bug in Comment 0.)

[0] https://github.com/w3c/webrtc-stats/issues/318
[1] https://github.com/w3c/webrtc-stats/issues/378
Depends on: 1376873
Flags: needinfo?(na-g)
You need to log in before you can comment on or make changes to this bug.