Avaya Logo

Previous Topic

Next Topic

Book Contents

Book Index

Using TTS and speech recognition languages with VoiceXML

Consider the following information for VoiceXML applications using Text-to-Speech (TTS) and speech recognition.

Languages

VoiceXML applications for Avaya IR work with languages determined by the speech engine and the corresponding TTS or speech recognition package:

To make languages work with VoiceXML, apply these settings:

If the lang attribute is set to something other than English and English is used in the prompt, the TTS engine attempts to translate the English into the specified language.

The defaults.xml file

If not otherwise specified in the VoiceXML /vs/data/vxml/defaults.xml file or in the VoiceXML application, the default TTS voice for VoiceXML is US English and female. To change the default gender to male at the system level, add the following line to the defaults.xml file:

<property name="promptgender" value="male"/>

To change the default language at the system level, modify the xml:lang property in the defaults.xml file. Note that the TTS voice administered as part of Text-to-Speech must match the defaults for VoiceXML. For example, if the VoiceXML defaults of US English and female are used, there must be a TTS voice available that is US English and female.

Note:
Changes to the xml:lang property affect both TTS and NLSR.

The xml:lang property is only valid within the defaults.xml file. To set the language for a VoiceXML document, use the xml:lang attribute of the <vxml> tag. To set the language for a specific prompt, use the xml:lang attribute of the <prompt> tag. To set the language for a specific grammar, use the xml:lang attribute of the <grammar> tag.

Changes to the defaults.xml file do not go into effect until you restart the voice system.

© 2006 Avaya Inc. All Rights Reserved.