Summary:
Recent studies again PJSIP, there is a demand, then the appropriate time, the need for a remote client playback, such as: Play an advertisement, or a piece of music. API need to achieve.
text:
Recently PJSIP want to be a robot, a relatively simple idea is to get the customer's voice stream, and then sent to ASR, identification, recognition result to NLP (brain), NLP results returned, then sent to TTS synthesis, synthetic speech, broadcast listen to the customer.
1. Get voice stream, ready to adopt PJSIP the recording interface, simple recording