Microsoft Speech Recognition Properties Tweaking

1
votes

Microsoft Speech Recognition can be tweak using properties ( http://msdn.microsoft.com/en-us/library/System.Speech.Recognition.SpeechRecognitionEngine_properties.aspx)

  • initialSilence
  • babble
  • endSilence
  • endSilenceAmbiguous

I'm looking for use case tweaking these properties. For instance what should I do - If I want to match a short word like yes / no - If I want to match a long sentence

How these properties will fit with SubSetMatching mode (http://msdn.microsoft.com/en-us/library/system.speech.recognition.subsetmatchingmode(v=vs.110).aspx) ?

BTW: is there a way to set SubSetMatching mode in the XML of the Grammar ?

1
c#speech-recognition

1 Answers

2
votes

Reducing the BabbleTimeout can be useful in noisy environments. In quieter environments, however, this may result in prematurely finalizing a recognition.

Reducing InitialSilenceTimeout, EndSilenceTimeout, and EndSilenceAmbiguousTimeout can be useful if the commands are short, and you're looking for a faster response from the engine. Again, however, reducing the timeouts too far can result in recognitions being prematurely finalized (i.e., the recognizer finishes before you've finished speaking).

These properties are handled in the audio processing portion of the SR engine, and don't affect SubsetMatchingMode.

And, as far as I know, there's no way to control the SubsetMatchingMode via XML.