Re: [webrtc-pc] Clarify the definition of "playout" for `RTCRtpContributingSource`. (#2172)

@henbos, play-out time = reception time + jitter buffer time depth.

As @chxg pointed out, the playout time will always be an estimate. For the use case of detecting when someone is speaking, which is how I understand this data is to be used, I think that recording at any point in @chxg's flow at or deeper than 2 is sufficient. The further down the flow the data are recorded, the greater accuracy becomes.  If the goal is to achieve that high level of accuracy, then I would think that this would be a callback and not a polling interface. I think the question of original intent is important here. I would to like to float phrasing the playoutTime as an estimate, and allowing implementations to leverage the knowledge of their internals to make the **best possible** estimate.

-- 
GitHub Notification of comment by na-g
Please view or discuss this issue at https://github.com/w3c/webrtc-pc/issues/2172#issuecomment-486752771 using your GitHub account

Received on Thursday, 25 April 2019 16:53:21 UTC