1330587 - framerate in webrtc getStats report does not change timely

Status: RESOLVED DUPLICATE of bug 1348657

(Core :: WebRTC: Networking, defect)

Description

[shuquan825]

User Agent: Mozilla/5.0 (Windows NT 6.1; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/55.0.2883.87 Safari/537.36

Steps to reproduce:

1. make a video call
2. limit bandwidth using tools like Netbalancer, monitor video quality and frameRate change in getStats report
3. recover the bandwidth to good condition, monitor again


Actual results:

when video becomes very slack, the framerate still keep big value for a long time sometimes several minutes.
when network recovers from bad to good situation, the video is very fluent, but the framerate still keep low value like 3~5 fps for a long time.

Compare with Chrome, it will change timely.


Expected results:

framerate should reflect video fluency.

Comment 1

[Byron Campen [:bwc]]

I guess this is because we're reporting the average for the entirety of the call, but as far as I can tell this statistic isn't part of the specification anyway.

jib?

Comment 2

[Jan-Ivar Bruaroey [:jib] (needinfo? me)]

Yes, hence the stat names frameRateMean and frameRateStdDev. - https://jsfiddle.net/ec9ossmu/

These were never in the spec, and we never prefixed them. See bug 951496 comment 32.

Comment 3

[Jan-Ivar Bruaroey [:jib] (needinfo? me)]

It's a rolling average as I recall.

Comment 4

[Jan-Ivar Bruaroey [:jib] (needinfo? me)]

What stat are you looking at in Chrome?

Comment 5

[Randell Jesup [:jesup] (needinfo me)]

aha. Looking at that bug I think we wanted to report them for our own use in Telemetry and about:webrtc.  Then after Olli's comment, we un-prefixed them.

Comment 6

[Randell Jesup [:jesup] (needinfo me)]

And Steven Lee from comment 5 there: 
> After discussing with CJ, I added some more statistics data, standard deviation, minimum and maximum values.

Comment 7

[Jan-Ivar Bruaroey [:jib] (needinfo? me)]

Please use new standard framesEncoded available in Firefox 55 instead. See bug 1348657 comment 23 for a demo.

I've filed bug 1367562 to remove these two stats.