About Engine Configurations

About Engine Configurations

Speech Engine Services (SES) creates default configurations for each of the speech engines that might be required by an application—Recognition, for telephony and multimodal applications (one each); Prompt, for prompt playback; and DTMF for applications supporting Dual Tone Multi-Frequency (DTMF) input.

Settings that can be configured for speech engines include:

  • The number of engine instances to handle incoming calls.

    The number of engine instances corresponds roughly to the number of simultaneous calls the system can support. Increasing or decreasing available engine instances can help optimize memory usage or enable the system to handle increased call volume.

  • The engine class and name for each type of engine required by the application.

    Engine class is either Recognition, Prompt, or DTMF. See Architecture: Speech Engine Services.

    Engine name is the actual name provided by the engine vendor (for example, Microsoft or ScanSoft Inc.), and is read directly from the registry keys installed by Setup. If a language pack is installed, the non-English language engine(s) also appear in this list.

  • The type of resource to load into the specified engine. This is either:

    • TelephonyRecognizer, to load grammar resources for telephony (voice-only) applications,

    • MultimodalRecognizer, to load grammar resources for multimodal applications,

      - or -

    • Voice, to load prompt databases for either telephony or multimodal applications.

  • The voice the user hears when speech must be synthesized.

    When an application needs to synthesize speech, callers hear a system-produced voice, rather than the one recorded for application prompts. See Configuring Speechify.

  • The name of the application that this engine configuration supports.

    This is Default unless you specify an application name. Leaving this setting as Default means all application resources are loaded into the default engine configurations.

Default Engine Configurations

The following table shows the initial number of instances specified by the default engine configurations.

Engine Type

Default # instances

TelephonyRecognizer

24 (services grammar files for voice input)

Voice

24 (services prompt databases for voice output)

DtmfRecognizer

6 (services grammar files for DTMF input)

MultiModalRecognizer

0 (services resources for multimodal applications; must be set by the administrator)

For non-English language engines, the initial instances are set to zero. See Configuring Support for Additional Languages.

See Also

Optimizing SES Memory Usage | Creating a New Engine Configuration | Modifying the Number of Available Engines | Modifying the Default Voice