Setting USAI parameters

Before initializing the Universal Speech Access API, specify configuration parameters in the USAI saiconfig.xml configuration file, located in the following directories:

Note: You must use a text editor to edit configuration entries. No comments are permitted in the file.

Universal Speech Access API configuration file parameters include:

Open tag

Parameter

Close tag

Description

<SpeechServer

ID=servername

 

Specifies the name of a speech server on which to create synthesizer or recognizer resources.

 

Server=speechvendor

 

Specifies the MRCP speech server vendor. Currently not implemented.

 

Deploy=false|true

>

Specifies whether or not the speech server is deployed. Valid entries include:

  • true: The MRCP client can attempt to connect to the server.

  • false: The MRCP client will not try to connect to the server.

<Protocol>

MRCP/version

</Protocol>

Specifies the media resource control protocol version used to communicate with the MRCP server. For example:

MRCP/1.0

Currently only MRCP/1.0 is supported.

<Session>

protocol/version

</Session>

Specifies the session control protocol used to communicate with the MRCP server. For example:

RTSP/1.0

Currently only RTSP/1.0 is supported.

<Transport>

transportype

</Transport>

Specifies the transport protocol used to communicate with the MRCP server. Currently only TCP is supported.

<Address>

IPaddress/RTSPport

</Address>

IPaddress specifies the MRCP server IP address or the hostname.

RTSPport specifies the RTSP port used when sending RTSP messages to the MRCP server (that is, the port on which the server listens for requests).

<Share>

false|true

</share>

OPTIONAL (default = false):Specifies whether or not TTS and ASR will use the same SessionID. (i.e. ASR address/port == TTS address/port)

<ASR>

<Enable> false|true

</Enable>

OPTIONAL (default = true):If the MRCP server is ASR capable, this turns on/off ASR APIs.

 

<ResourceName>Recognizer

</ResourceName>

OPTIONAL (default = recognizer):Recognizer defines the ASR resource name assigned by the MRCP server.

 

 

</ASR>

Ends ASR parameters.

<TTS>

<Enable> false|true

</Enable>

OPTIONAL (default = true):If the MRCP server is TTS capable, this turns on/off TTS APIs.

 

<ResourceName>synthesizer

</ResourceName>

OPTIONAL (default = synthesizer):synthesizer defines the TTS resource name assigned by the MRCP server.

 

 

</TTS>

Ends TTS parameters.

 

 

</SpeechServer>

Ends USAI parameters.


The following example shows sample configuration file entries. SPEECHSERVER1 sets the optional parameters and SPEECHSERVER2 uses the default:

<SpeechServerConfiguration >
<SpeechServer ID="SPEECHSERVER1" Server="Generic" Deploy="false">
<Protocol>MRCP/1.0</Protocol>
<Session>RTSP/1.0</Session>
<Transport>TCP</Transport>
<Address>192.168.0.1:554</Address>
<Share>false</Share>
<ASR>
<Enable>true</Enable>
<ResourceName>recognizer</ResourceName>
</ASR>
<TTS>
<Enable>true</Enable>
<ResourceName>synthesizer</ResourceName>
</TTS>    
</SpeechServer>
<SpeechServer ID="SPEECHSERVER2" Server="Generic"  Deploy="false">
<Protocol>MRCP/1.0</Protocol>
<Session>RTSP/1.0</Session>
<Transport>TCP</Transport>
<Address>HostnameOfSpeechServer2:4900</Address>
</SpeechServer>
</SpeechServerConfiguration>