Presentation of Information - Aurally

• "Web approach" lowers telephone service development cost
• On small mobile devices (3G)
  • voice input instead of keyboard
  • voice output instead of screen

What a voice browser does for you:

• Browser: "What is your departure airport ?"
• Caller: "Nice"
• Browser: "Where do you want to fly to ?"
• Caller: "Paris"
• Browser: "At what time ?"
• ...

System Architecture

Voice Browser Architecture

W3C XML Markup Languages for Voice

• Dialogue Control: VoiceXML
• Speech Recognition: Speech Recognition Grammar Syntax
• Speech Synthesis: Speech Synthesis Markup Language

VoiceXML

• XML-based programming language for voice applications
• Need to describe
  • System prompt
  • Expected user response
  • Action on expected response
  • Action on inexpected response

Example: Prompt

<prompt>
  <audio>
    Welcome to the <say-as type="acronym">W3C</say-as>
    Voice <say-as type="acronym">XML</say-as> server.
    Would you like to have more information about the
    architecture domain, the document formats domain, the 
    interaction domain, the technology and society domain 
    or the Web Accessibility Initiative ?
  </audio>
</prompt>

Techniques to Improve User Interface

• "Barge-In": User can interupt prompt with answer
• Alternate prompts: Vary prompt for input
• "Mixed Dialogue": User can give response that does not answer question

  Browser: When will you arrive at the hotel ?

  User: I need to rent a car

  Browser: Which company do you prefer ?

  ...

• See VoiceXML 2.0 Working Draft

Speech Recognition Grammar

Example Rule

<rule id="city">
  <one-of>
     <item>Rio de Janeiro</item>
     <item>Rio</item>
     <item>Paris</item>
     ...
  </one-of>
</rule>

Speech Synthesis Markup Language

To be used for Speech Synthesis, Voice Browser or Text-to-Speech applications

You can say what you want spoken and in what language

<?xml version="1.0" encoding="ISO-8859-1"?>
<speak version="1.0" xmlns="http://www.w3.org/2001/10/synthesis"
         xml:lang="en-US">
  <paragraph>I don't speak Japanese.</paragraph>
  <paragraph xml:lang="ja">Nihongo-ga wakarimasen.</paragraph>
</speak>

You can set the speaker tone of the voice

<?xml version="1.0" encoding="ISO-8859-1"?>
<speak version="1.0" xml:lang="en-US"
xmlns="http://www.w3.org/2001/10/synthesis">   
  <voice gender="female">Mary had a little lamb,</voice>
  <!-- now request a different female child's voice -->
  <voice gender="female" variant="2">
    It's fleece was white as snow.
  </voice>
  <!-- platform-specific voice selection -->
  <voice name="Mike">I want to be like Mike.</voice>
</speak>

You can set a contour to the speech to get the right intonation

<?xml version="1.0" encoding="ISO-8859-1"?>
<speak version="1.0" xml:lang="en-US"
xmlns="http://www.w3.org/2001/10/synthesis">
  <prosody contour="(0%,+20)(10%,+30%)(40%,+10)">
    good morning
  </prosody>
</speak>