currently talking w/o their input? Might end up eating too many resources.
Or you could do the muxing clientside and have the server just provide the audiostreams of everyone who is currently talking. The clients can then just mute themselves.
Обсуждают сегодня