Re: [webrtc-stats] Add insertedSamplesForDeceleration and removedSamplesForAcceleration (#408)

> I have another question, are the insertions and removals happening at playback clockrate or at packet arrival, the reason for asking is -- "would make sense to count the number of times the decision to accelerate/decelerate is made?"
> to calculate something like average number of insertions or removals?

I would discuss this separately, but I don't think so. I think the number of times a decision is made to accelerate or decelerate would be implementation specific and if acceleration/deceleration is tiny I think the number of times it happens doesn't tell you anything, versus it happening not very often but happening at a great magnitude.

I think what is interesting to calculate is the average number of accelerated/decelerated samples compared to the number of samples received, and that does not require any new counters, we already have samplesReceived.

GitHub Notification of comment by henbos
Please view or discuss this issue at using your GitHub account

Received on Tuesday, 2 April 2019 07:42:46 UTC