The idea here is that every consumer of a video track uses a direct listener, and a video track's AppendToTrack does both: 1 call straight into any track listeners 2 queue an event to set the frames for the track in the MSG (just state keeping to help future listeners to the track know what's currently been pushed to the track) Certain care will need to be taken around track disabling, where the track's state will have to be communicated back from the MSG to the producer somehow.