Re: [webrtc-encoded-transform] Should we expose the precise timestamp of the frame that was triggered by a generateKeyFrame call? (#128)

One possibility would be to make generateKeyFrame return a Promise<long long> so that applications would get the frameId of the key frame.

With the current spec, we would not be able to distinguish the case of a key frame triggered by generateKeyFrame or a key frame arriving before by chance.
If we happen to need to disambiguate this, we could update the spec without requiring an API change.

Of course, this becomes more difficult if you expect several key frame generations by passing several rids.
One possibility would be a sequence of frameIds, or we would just stick with one frameId (the first one being received), hoping the other frames would be exposed sequentially.

@alvestrand, @aboba, wdyt?

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

Sent via github-notify-ml as configured in

Received on Friday, 28 January 2022 09:03:24 UTC