I have a question related with conference servers. Do you now how is done the mixing in a Conference? Specifically, how input RTP streams are mixed into output string? Should I need to mix in a linear way, interleaving samples of different users?