Creating a New Engine Configuration

  • Article

Creating a New Engine Configuration

Use the Speech Engine Configurations tab of the Microsoft Management Console (MMC) snap-in for Microsoft Speech Server (MSS) to create a new engine configuration.

For example, you can dedicate an engine configuration to a specific application in order to distribute resource load.

To Create a New Engine Configuration

  1. Open the MMC Administration console. (For instructions, see Administering Speech Server Using MMC.)

  2. In the console tree, expand the applicable group.

  3. In the details pane, double-click the computer running Speech Engine Services (SES) to open the SES properties page.

  4. Click the Speech Engine Configurations tab.

  5. Create an engine configuration for each engine required by the application:

    Create a configuration for the telephony recognizer — The telephony recognizer engine manages grammar resources for telephony (voice-only) applications. Create a configuration for a multimodal recognizer — The multimodal recognizer engine manages grammar resources for multimodal applications. Create a configuration for the prompt engine — The prompt engine manages prompt databases for either telephony or multimodal applications. Create a configuration for a DTMF parser — The DTMF engine manages Dual Tone Multi-Frequency (DTMF) grammar files. DTMF grammars are generally small; in most cases it makes sense to let the DefaultDTMFParser engine configuration handle the DTMF instances for all applications.  

  6. When finished, click OK.

To Create a Configuration for the Telephony Recognizer

  1. On the Speech Engine Configurations tab, click Add, and type a name for the telephony recognizer engine configuration, for example, "MyAppTelephonyRecognizer."

  2. Under Settings for Selected Configuration, next to Instances, specify the number of engine instances to create at startup for the telephony recognizer.

    Note  For Microsoft Speech Server 2004 Standard Edition, the total number of instances for all recognition engines (DefaultTelephonyRecognizer, DefaultMultimodalRecognizer, and any application-specific recognizers) cannot exceed 24.

  3. Next to Engine Class, select Recognition.

  4. Under Engine Name, select Microsoft English (U.S. Telephony) v 7.0 Server; or, if a language pack is installed, select the language-appropriate engine.

    Alternatively, select an installed third-party telephony recognition engine from the list.

  5. Under Application Settings for Selected Configuration, next to Resource Set to Preload, select TelephonyRecognizer. This must match the associated ResourceSetType entry in the application manifest file. See About Application Manifest Files.

  6. In the Application List, delete any previous entry (such as Default) and type the name of the application.

  7. Click Apply.

To Create a Configuration for a Multimodal Recognizer

  1. On the Speech Engine Configurations tab, click Add, and type a name for the multimodal recognizer engine configuration, for example, "MyAppMultimodalRecognizer."

  2. Under Settings for Selected Configuration, next to Instances, specify the number of engine instances to create at startup for the multimodal recognizer.

Note For Microsoft Speech Server 2004 Standard Edition, the total number of instances for all recognition engines (DefaultTelephonyRecognizer, DefaultMultimodalRecognizer, and any application-specific recognizers) cannot exceed 24.

  1. Next to Engine Class, select Recognition.

  2. Under Engine Name, select Microsoft English (U.S. Wideband) v 7.0 Server. Alternatively, select an installed third-party multimodal recognition engine from the list.

  3. Under Application Settings for Selected Configuration, next to Resource Set to Preload, select MultimodalRecognizer. This must match the associated ResourceSetType entry in the application manifest file. See About Application Manifest Files.

  4. In the Application List, delete any previous entry (such as Default) and type the name of the application.

  5. Click Apply.

To Create a Configuration for the Prompt Engine

  1. On the Speech Engine Configurations tab, click Add, and type a name for the prompt engine configuration, for example, "MyAppVoice."

  2. Under Settings for Selected Configuration, next to Instances, specify the number of engine instances to create at startup for the prompt engine.

    Note  For Microsoft Speech Server 2004 Standard Edition, the total number of instances for all speech output engines (DefaultVoice, as well as any application-specific voice configurations) cannot exceed 24.

  3. Next to Engine Class, select Prompt.

  4. Under Engine Name, specify the voice to use for the application. By default, Speechify Jill is selected. To change the voice, select another one from the drop-down list.

  5. In the Application List, delete any previous entry (such as Default) and type the name of the application.

  6. Click Apply.

To Create a Configuration for a DTMF Parser

  1. On the Speech Engine Configurations tab, click Add, and type a name for the DTMF parser configuration, for example, "MyAppDTMFParser."

  2. Under Settings for Selected Configuration, next to Instances, specify the number of engine instances to create at startup for the DTMF parser.

    Note  For Microsoft Speech Server 2004 Standard Edition, the total number of instances for all DTMF parser engines (DefaultDTMFParser, as well as any application-specific DTMF configurations) cannot exceed 24.

  3. Next to Engine Class, select DTMF.

  4. Under Engine Name, accept the default (Microsoft English (U.S. Wideband) v 7.0 Server). Alternatively, select an installed third-party DTMF recognition engine from the drop-down.

  5. In the Application List, delete any previous entry (such as Default) and type the name of the application.

  6. Click Apply.

See Also

Capacity Guidelines | About Engine Configurations | Configuring Speech Engines | Configuring Support for Additional Languages