I am trying to understand how audio streams are initialised in FreeSWITCH. I'm working with ESL (inbound and outbound) and trying to record some audio. Through testing i have proven that audio is working correctly for bridged calls, conferences and file playback however not for single legged audio recording - that is, until I play a file first. In my example the playback application is executed from an outbound socket, the call is then parked and the inbound socket captures the park event and starts recording. With the playback in place i get audio, and without it I simply get a silent file of the correct duration presumably meaning that the the RTP stream is not active. My question therefore is: How can I initialise the RTP stream for recording, without having played a file first? Is there a command for this, or an argument I can pass to the record application? My objective is to understand how RTP streams are initialised in FreeSWITCH and any input would be greatly appreciated.