Configuring Vocalizer

Different people configure Vocalizer at different times. How you set parameters depends on your job function:

  • System administrators set defaults when installing Vocalizer. For example, administrators can set different defaults for each application that uses Vocalizer.
  • Voice browser developers set defaults when integrating Vocalizer into their voice platforms. For example, you can configure the default speaking rate for the synthesized speech.
  • VoiceXML application developers set application-specific values within the session.xml file. For some voice platforms, such as Nuance Voice Platform, these application-specific values can be set via VoiceXML <property> elements where the value uses the "switts." prefixed version of the parameters.
  • In Speech Server, system administrators can set configuration parameters on the Vocalizer host in Management Station. If not using Management Station, set parameters in configuration files.

Configuration precedence

You can set parameters at different levels of precedence. Some are always global (all sessions), and others can be changed for individual sessions.

From the highest to the lowest precedence:

  • Parameters specified via markup (native escapes or SSML markup).
  • Parameters set at the start of a session. This includes setting parameters in a session configuration file (session.xml), which defines defaults for the entire session.
  • Parameters changed on an already open engine instance. This includes some parameters set on a Nuance Speech Server instance during system initialization. Some of those Speech Server parameters affect Vocalizer.
  • Parameters specified while opening an engine instance.
  • Parameters set in Management Station or in Vocalizer configuration files.

    If using the Management Station, set parameters on the Vocalizer instance.

    If not using the Management Station, set parameters in each instance of the Vocalizer configuration file (where xx is the instance number):

    • Windows: %NUANCE_DATA_DIR%\system\config\User-nvsxx.xml
    • Linux: $NUANCE_DATA_DIR/system/config/User-nvsxx.xml

    For service properties exposed in Speech Suite, see Configuring the Vocalizer service.

  • The Vocalizer baseline configuration file (VOCALIZER_SDK\config\baseline.xml) contains the default settings installed with Vocalizer.

Parameters specified via markup

You can update many parameters by inserting markup in the input text. These settings only apply to the current synthesis request, and they are restored to the previous values for the next synthesis request.

Vocalizer supports two markup languages: the native markup (the default), and W3C SSML (with some proprietary extensions).

Parameters set at the start of a session

A session configuration file (session.xml) defines the default configuration invoked at the start of a session (after an engine instance is created). The file has separate sections for each Nuance speech product, including a <speechsynth> section specifically for parameters that affect Vocalizer.

  • A session.xml file does not contain all possible parameters. Many other parameters are set via other mechanisms.
  • During a session, applications can override certain session defaults at any time by setting VoiceXML and Nuance-specific properties via markup.

Parameters changed on an already open engine instance

Some Vocalizer parameters can be changed on an already open TTS engine instance and affect that instance only. This includes parameters that affect Vocalizer that are set on a Nuance Speech Server instance. For these parameters, use the SET-PARAMS method. These parameters are a superset of what can be specified while opening the engine instance.

Parameters specified while opening an engine instance

The interface you use to access Vocalizer determines whether you specify some Vocalizer parameters when opening a TTS engine instance rather than applying them globally. For Speech Server, specify some parameters, like the audio format, when establishing the MRCP connection (SIP INVITE for MRCPv2). These parameters are set in the Speech Server default configuration, on the Speech Server service in Management Station or in the NSS configuration file (if not using Management Station).