voice.rst

Voice

This service handles voice control endpoint messages, parses the data and exposes it for external tools. It also allows voice control messages to be sent to a Pebble. It does not implement the state machine for ordering voice control messages correctly: this must be handled by the user of the service.

Events

The service exposes the following events, which can be subscribed to with VoiceServer.register_handler:

• session_setup - Session setup request received
• audio_frame - Audio data frame received
• audio_stop - Audio data stopped

Voice Protocol Sequencing

The correct sequencing for communicating with the Pebble smartwatch is as follows:

Pebble-terminated sessions:

This is the normal sequence of communication. The Server should wait until it receives a stop message from the Pebble before sending the dictation result.

Message	Sender	Event/Function
Session setup request	Pebble	session_setup
Session setup result	Server	VoiceService.send_session_setup_result
Audio data (n frames)	Pebble	audio_frame
Audio stop	Pebble	audio_stop
Dictation result	Server	VoiceService.send_dictation_result

Server-terminated sessions:

If an error occurs a server can terminate the session by sending an audio stop message followed by the dictation result. The dictation result should always be sent.

Message	Sender	Event/Function
Session setup request	Pebble	session_setup
Session setup result	Server	VoiceService.send_session_setup_result
Audio data (n frames)	Pebble	audio_frame
Audio stop	Server	VoiceService.send_stop_audio
Dictation result	Server	VoiceService.send_dictation_result

libpebble2.services.voice