We do this by manually managing the RTCAudioSession. Unfortunately to do this we have to include a couple of RTC headers not exported by the default build of WebRTC.framework (see: Libraries/WebRTC) // FREEBIE
