webrtc-stats
webrtc-stats copied to clipboard
w3c
WebRTC Statistics
We create an awful lot of `codec` stats objects. Every PT is listed for every m= section. And they are listed for both inbound and outbound direction, so you get...
The spec for [RTCRemoteOutboundRtpStreamStats](https://www.w3.org/TR/webrtc-stats/#dom-rtcremoteoutboundrtpstreamstats).remoteTimestamp says: > _remoteTimestamp_ of type DOMHighResTimeStamp > [remoteTimestamp](https://www.w3.org/TR/webrtc-stats/#dom-rtcremoteoutboundrtpstreamstats-remotetimestamp), of type [DOMHighResTimeStamp](https://www.w3.org/TR/hr-time-3/#dom-domhighrestimestamp) [[HIGHRES-TIME](https://www.w3.org/TR/webrtc-stats/#bib-highres-time)], represents the remote [timestamp](https://www.w3.org/TR/webrtc-stats/#dom-rtcstats-timestamp) at which these statistics were sent by the remote endpoint....
The definition for `roundTripTime`, `totalRoundTripTime`, and `roundTripTimeMeasurements` in `remote-outbound-rtp` are all, seemingly, incorrect. They refer to the `DLRR` report block of Sender Reports(SR) but reference extended RTCP reports(defined in [RFC3611](https://www.rfc-editor.org/rfc/rfc3611))....
best reviewed ignoring whitespace diff fixes #658 *** Preview | Diff
Probably easiest to remove tests for unimplemented provisional metrics, in case they get implemented in the future we should add tests as part of those efforts
https://w3c.github.io/webrtc-stats/#dom-rtcoutboundrtpstreamstats-totalencodedbytestarget is implemented by libwebrtc but returns 0. So far nobody has complained (but me but I complain about *everything*) This doesn't offer much value over comparing targetBitrate to 8*bytesSent/second...
Good source of reference: https://webrtc-stats.callstats.io/verify/ @vr000m @alvestrand @dontcallmedom
It's only implemented in inbound-rtp, [not remote-inbound-rtp](https://wpt.fyi/results/webrtc-stats/supported-stats.html?label=experimental&label=master&aligned). Let's just move it?
Clarify which metrics are only applicable to a particular kind according to implementation interest.
Some metrics are defined for both audio and video but have only been implemented for one of the kinds. As a follow-up to "move unimplemented metrics to provisional spec" (#621...
Getting rid of `"sender"`, `"receiver"` and `"transceiver"` stats may be a good idea, but deserves discussion. Unfortunately https://github.com/w3c/webrtc-stats/pull/628 was merged before this discussion could happen, so I'm opening this issue...
