Configuring Speech Server

Configuration of Nuance Speech Server is both static and dynamic. Configuration files contain static settings loaded during system initialization or at the start of a session. Dynamic settings occur at runtime when VoiceXML browsers request speech recognition and text-to-speech resources on behalf of applications. Browsers communicate requests through the Speech Server) using an open, client-server protocol known as MRCP. The requests contain configuration settings as well as commands for speech operations.

Although Speech Server generally works well with the installation defaults, voice browsers (and their MRCP clients—see Configuring MRCP clients) might require non-default values.

If using the Management Station, set parameters on the Speech Server instance.

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

• Windows: %NSSSVRSDK%\config\user-NSSxx.txt
• Linux: $NSSSVRSDK/config/user-NSSxx.txt

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