Hi Simon and company! I need to separate the incoming audio stream and the outcoming, that is, in "my" mediastremer2 library now when I start the audiostream with its RtpSession, this rtpSession is created half-duplex, not full, RECV-ONLY, not SENDRECV, due to i created in a TTS side the rtpSession as SENDONLY.
The rtpSession SENDONLY works fine, when exists audio from the tts I do "rtp_session_send_with_ts" of this audio. this is ok. But in the incoming side, (mediastremer), the MSfilter "recv" which do the rtp_session_recv don't catch audio never.
If the session is created SENDRECV works fine. IT's possible that the synchronous is bad? Thank you to all.