For various reasons, I need to use the raw SpeechRecognizer
API instead of the easier RecognizerIntent
(RECOGNIZE_SPEECH) activity.
That means, among other things, that I need to handle RecognitionListener.onError()
myself.
In response to some of the errors, I simply want to re-start listening. This looks straightforward but when I just call SpeechRecognizer.startListening()
upon error, this sometimes seems to trigger two different errors:
ERROR/ServerConnectorImpl(619): Previous session not destroyed
and
"concurrent startListening received - ignoring this call"
Which hints that I should have done some cleanup before attempting to call SpeechRecognizer.startListening()
again.
If this is true, it means that upon a RecognitionListener error, listening is not automatically stopped and/or canceled.
It is also possible that some errors do stop/cancel listening, while others don't. There are really only 9 SpeechRecognizer errors:
- ERROR_NETWORK_TIMEOUT
- ERROR_NETWORK
- ERROR_AUDIO
- ERROR_SERVER
- ERROR_CLIENT
- ERROR_SPEECH_TIMEOUT
- ERROR_NO_MATCH
- ERROR_RECOGNIZER_BUSY
- ERROR_INSUFFICIENT_PERMISSIONS
Since the documentation isn't very detailed about which error cancels listening and which doesn't, do you happen to know, based on your experience, which errors require doing cleanup (and to which extent) before attempting SpeechRecognizer.startListening()
again?