Release v1.0.0

Voice generation is changed to asynchronous in order to minimize latency.
Voice argument is removed from answer_cb and add new callback named answer_with_voice_cb.