Table of Contents

VoiceInteraction ASR

About

VoiceInteraction
Rua Alves Redol, 9
1000-029 Lisboa
Portugal

Website: http://www.voiceinteraction.pt

Definition

VoiceInteraction, founded in 2008 and based in Lisboa (Portugal), is a company specialised in the development of speech technologies. VoiceInteraction develops voice synthesis and speech recognition engines for many web and telephony applications.

Languages

Portuguese · Portuguese (Brazil)

Installation

This is for de Debian Lenny version, for other Linux, read the official Audimus Installation manual. To configure the apt client, just edit and add to the file /etc/apt/sources.list one of the following line set.

For asterisk 1.Y.X

deb http://services.voiceinteraction.pt/repo/Debian/5.0 engines 3rdparty Dixi Audimus
deb http://services.voiceinteraction.pt/repo/Debian/5.0 asterisk.1.Y.X Dixi Audimus

Refresh the package lacal base with:

# apt-get update

And Install the package audimus-asterisk-xx-xx (where xx-xx is the language requested):

# apt-get install audimus-asterisk-es-es

The following extra packages will be installed:

audimus audimus-config audimus-model-es-es-monophones-g2p-phonemodels audimus-model-es-es-monophones-mlp-telephone audimus-model-es-es-monophones-task-asterisk

Activate the license with:

# audimus_activate_license

Please enter your Audimus license:

xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
Your License was successfully activated!

Configuration

To enable speech recognition, change the main speech parameter. The “speech” parameter can get three values, “yes”, “automatic”, “no” or “emulation” (don’t generate errors if you enable speech grammars).

…
speech=automatic
speechprovider=verbio
…

In the VoiceXML browser configuration file:

############################
# ASR server configuration #
############################

client.rec.resource.0.cacheDir              VXIString   /tmp/cacheContent
client.rec.resource.0.format                VXIString   txt
client.rec.resource.0.syntax                VXIString   doctype

You need to restart the Vxi and Asterisk to get all the changes.

Logs file from the ASR engine:

# tail -f /var/log/VI/VI.log

The following VoiceXML example uses the speech recognition, with the built in grammar, ‘digits.’

<?xml version="1.0" encoding="iso-8859-1"?>
<vxml version="2.0" xmlns="http://www.w3.org/2001/vxml" xml:lang="en-UK">
 <form>
  <property name="inputmodes" value="voice"/>
  <property name="timeout" value="30s"/>
  <field name="text" type="digits">
   <catch event="noinput nomatch">
    <reprompt/>
   </catch>
   <prompt>
    Speak to me:
   </prompt>
  </field>
  <filled>
   <prompt>
    You say me:
    <value expr="text" />
   </prompt>
   <clear namelist="text" />
  </filled>
 </form>
</vxml>