Why not do multiple cameras and broadcast each of the feeds? Then it's up to the software for each user to do head tracking.