ndnrtc

I wish silence suppression would be an easy application-specific issue. However it's not and here's why: if producer stops publishing audio, for consumer this manifests itself as the producer just went offline, so consumer switches back to chasing mode. Chasing mode is not preferred if audio playback resumes as few hundreds of milliseconds (up to 1-1.5 seconds) will be lost during chasing before it'll start playing on consumer side.

Thus a re-design of producer and consumer for silence suppression is required (at least, an introduction of special case when producer notifies consumer that it has suppressed silence and consumer shouldn't switch back to chasing but should maintain same pipeline as it has at the moment of the receipt of the suppression notification).

Peter, let's start on the design of a low-rate silence notification packet that can be returned instead of audio. Can we target this for design discussion next week?