Legacy Document

Inside Macintosh: Speech Recognition Manager /

Chapter 1 - Speech Recognition Manager / Using the Speech Recognition Manager

Opening Recognition Systems and Recognizers

Before calling any other Speech Recognition Manager routines, you need to call the SROpenRecognitionSystem function to open a recognition system. Then you need to call SRNewRecognizer to create a recognizer associated with that recognition system. Listing 1-3 shows how to use these functions to initialize speech recognition for your application.

Listing 1-3 Initializing speech recognition

SRRecognitionSystem     gRecSystem = NULL;/* our recognition system */
SRRecognizer            gRecognizer = NULL;/* our recognizer */

OSErr MyInitSpeechRecognition (void)
{
   OSErr             myErr;
   
   /* Ensure that the Speech Recognition Mgr is available. */
   if (!MyHasSpeechRecognitionMgr()) 
      return(kSRNotAvailable);

   /* Open the default recognition system. */
   myErr = SROpenRecognitionSystem(&gRecSystem, kSRDefaultRecognitionSystemID);

   /* Use standard feedback window and listening modes. */
   if (!myErr) {
      short myModes = kSRHasFeedbackHasListenModes;
      myErr = SRSetProperty(gRecSystem, kSRFeedbackAndListeningModes, 
                                          &myModes, sizeof(myModes));
   }

   /* Create a new recognizer. */
   if (!myErr)
      myErr = SRNewRecognizer(gRecSystem, &gRecognizer, kSRDefaultSpeechSource);

   /* Set reference constant of rejected word. */
   if (!myErr)
      myErr = MySetRejectedWordRefCon(gRecSystem);

   return(myErr);
}

Note that the MyInitSpeechRecognition function defined in Listing 1-3 explicitly sets the feedback and listening modes property of the recognition system to the value kSRHasFeedbackHasListenModes immediately after it calls the SROpenRecognitionSystem function. This instructs the recognition system to use the standard feedback and listening behavior for any recognizers associated with it, which helps ensure that the user has a consistent experience when using any applications that use the Speech Recognition Manager. In general, you should use other feedback and listening modes values only if your application provides its own feedback mechanism.

The MyInitSpeechRecognition function defined in Listing 1-3 also creates a recognizer by calling the SRNewRecognizer function. SRNewRecognizer takes a reference to an existing recognition system and a speech source ID, which specifies a speech source. SRNewRecognizer returns a reference to the new recognizer through its second parameter (here, gRecognizer). MyInitSpeechRecognition also calls MySetRejectedWordRefCon (shown in Listing 1-8 on page 1-24) to simplify the processing of recognition results.

You can terminate your connection to the Speech Recognition Manager by calling SRReleaseObject to dispose of your recognizer and then SRCloseRecognitionSystem to close your recognition system, as illustrated in Listing 1-4.

Listing 1-4 Ending speech recognition

OSErr MyTerminateSpeechRecognition (void)
{
   OSErr                   myErr;

   /* Stop processing any incoming sound. */
   myErr = SRStopListening(gRecognizer);

   /* Balance call to SRNewRecognizer in MyInitSpeechRecognition. */
   myErr = SRReleaseObject(gRecognizer);

   /* Balance call to SROpenRecognitionSystem in MyInitSpeechRecognition. */
   myErr = SRCloseRecognitionSystem(gRecSystem);

   return(myErr);
}