I believe the limitation to 3 or 4 streams, is imposed by the DSPs, which can only handle this much (the mixing is done on the digital voice streams). In a way it makes sense: if more than 4 people speak at the same time, nobody understands anything....
