|Speech Recognition Listener does not detect end of speech in Jelly Bean||abo||7/31/12 6:09 AM|
I wrote an app implementing my own RecognitionListener and it was running fine under 2.3.6 up to ICS on my Nexus S.
After the update to Jelly Bean, the RecognitionListener doesn't detect the end of speech anymore (so, once started the method "onRmsChanged" is called until the app is terminated).
If I start the recognition and do not speak, the onError method with "No match" error is never reached as it did before.
Apart from that, starting the listener produces a sound which was not there before.
mSTT = SpeechRecognizer.createSpeechRecognizer(context);
mSTT.setRecognitionListener(this);intent = new Intent(RecognizerIntent.ACTION_RECOGNIZE_SPEECH);
I do not explicetly set the minimum time as described in the documentation nor do I call stopListening because that was not necessary before.
Any idea what was changed in Jelly Bean or what needs to be changed when programming a recognition listener?
And how to get rid of the sound when starting listening (or can anyone tell me if there are also some new sounds when end of speech or an error is detected)?