- From: Harald Alvestrand via GitHub <sysbot+gh@w3.org>
- Date: Fri, 09 Sep 2022 17:46:16 +0000
- To: public-webrtc-logs@w3.org
Contemplating the timestamp ... if the keyframe is generated much later than the expected time, having the timestamp would immediately seem to be more useful. The transform (which observes the resulting keyframe directly) only needs it to distinguish a keyframe generated by the request from a keyframe generated by a random timer expiring - but the sender has no other way of knowing when the keyframe goes out. However, the issue of multiple RIDs, which potentially may result in multiple keyframes with differing timestamps due to congestion budgets, makes the API be awfully messy if we want timestamps and want them to be truthful. My current guesstimate of what's right for a sender would therefore be multiple RIDs, no timestamp return. -- GitHub Notification of comment by alvestrand Please view or discuss this issue at https://github.com/w3c/webrtc-encoded-transform/pull/146#issuecomment-1242283391 using your GitHub account -- Sent via github-notify-ml as configured in https://github.com/w3c/github-notify-ml-config
Received on Friday, 9 September 2022 17:46:18 UTC