Ok, but I still don't get how the final recipient gets to know the *original* presentation time (which correspond to audio) if I generate a new one in the transcoder... Is this done via RTCP?! If yes, could you shortly explain it, please? Julian