WebRTC Metrics

A comprehensive overview of WebRTC statistics, derived calculations, extracted features, and observable signals, to better understand call quality, connectivity, and user experience in rtcStats

Back

trackaudiovideoinboundoutbound

Track identity (feature)

Track-level identity features that set the grain of a track row: media kind, direction, track identifier, and when the track was added.

Description

Track-level features, extracted per media track into features_track.

These features identify what a features_track row is and how it joins to the rest of a session:

kind - media type of the track, audio or video.
direction - outbound (sent by this peer) or inbound (received by this peer).
trackIdentifier - the track id, ties the row back to a specific MediaStreamTrack.
startTime - when the track was added to the connection.

Extracted by the open-source rtcstats-features package.

Why it matters

kind and direction are the two filters you reach for on almost every feature query. A single connection carries audio and video, sent and received, so every metric in features_track is meaningful only once you have split by these two columns. Audio concealment, video freezes, encoder cost: each belongs to one specific slice, and mixing slices produces numbers that mean nothing.

trackIdentifier gives you a stable handle on one track across the session, useful when you correlate a track row with raw getStats dumps or with signaling logs. startTime tells you when the track joined, which matters for tracks added mid-call (a screen share started later, a camera toggled on) and for normalizing cumulative counters by how long the track was actually live.

Treat features_track.connection_id = features_connection.id as the join key, then filter on kind and direction to reach the grain you want.

SQL example

Count tracks by kind and direction:

SELECT
  track.kind,      -- 'audio' or 'video'
  track.direction, -- 'inbound' or 'outbound'
  COUNT(*) AS tracks
FROM "rtcstats-server" AS server
JOIN features_metadata   AS metadata   ON metadata.dump_id    = server.id
JOIN features_connection AS connection ON connection.dump_id  = metadata.id
JOIN features_track      AS track      ON track.connection_id = connection.id
GROUP BY track.kind, track.direction
ORDER BY track.kind, track.direction;

Related features

Codec - the negotiated codec for each identified track.