Condividi tramite


Recognition Mode: Single

  Microsoft Speech Technologies Homepage
<listen mode="single">

Remarks

Single-mode recognition is typically used in multimodal scenarios. A multimodal scenario is one in which users control the period of time they wish to speak to the device by tapping and holding their stylus on a form field. The browser uses a GUI event that occurs, for example, when the stylus is lifted, to control when recognition should stop and results are returned. In this mode, recognition results are returned under control of an explicit Stop method call from the application.

  • Case 1 demonstrates the Stop method call and the possible resulting onreco or onnoreco events.
  • Case 2 illustrates the occurrence of an onnoreco event in response to the babbletimeout interval being exceeded. This event automatically terminates the recognition.
  • Case 3 demonstrates how an onnoreco event occurs when there is an unrecognizable utterance. It should be noted, however, that this does not automatically terminate the recognition.
  • Case 4 demonstrates how an onsilence event occurs if speech is not detected within the initialtimeout interval. Again, this does not automatically terminate the recognition.

For single-mode recognition, the only speech event that automatically stops a recognition, besides a Stop method call, is an onnoreco event that occurs as a result of the babbletimeout interval being exceeded. Of course, non-speech events such as an onerror event or some events of the dtmf element also end a recognition.

The following event diagram illustrates the single-mode recognition time lines.

See Also

listen Element |  Recognition Mode: Single Example | Recognition Mode: Automatic | Recognition Mode: Multiple