Linphone is a very good sip phone. I'm using it for making an automatic response system. So i send samples of decoded audio in the dtmfgen filter to an server ASR (Automatic Speech Recognition). I generate the response in text automatically, and with a TTS I generate audio and with the rtp_send api I send the audio back. I don't need files or oss or alsa (sound devices). The linphone is connected to an Asterisk, from wich it receives the calls, opening a Rtpsession.
With a unique call works very well, but How I implement it to support more than a call? I suppose that when my linphone received another call, i should create other paralel audiostrem, (recv->decoder->dtmfgen), with ohter rtpSession. is this possible? How? Where must i modify? Need i open more sip canals or by the sip canal opened with asterisk all calls can incoming?