RE: ASR text

You are not able to do this because natural language recognition is not up
to speed yet.
Grammars are used to constrain the span of allowable (spoken) inputs for a
field with a
high degree of accuracy and confidence level.

Ranjan


-----Original Message-----
From: Sivadasan Plakote [mailto:splakote@infoglyptic.com]
Sent: Wednesday, September 25, 2002 6:30 AM
To: W3C Voice Mailing List
Subject: Re: ASR text



Hi all
please see following code:

  <field name="SAYANYTHING">
<!-- NO GRAMMAR -->
       <prompt>
    Say your name.         
</prompt>

       <filled>
         <if cond="SAYANYTHING$.utterance == 'John' ">
         ...........................
         </if>
        </filled>
</field>

Why we are not able to do this  with out grammar?.

Thanks
Siva




> Hi Murali,
> 
> yes, just use the shadow variables.
> 
> With FIELD$.utterance you should get, what you want.
> Just put the name of the field in the VoiceXML code for FIELD
> 
> regards
> 
> Stefan
> 
> -----------------------------
> Stefan W. Hamerich
> TEMIC Speech Dialog Systems
> Research Department
> Soeflinger Str. 100
> 89077 Ulm
> Germany
> 
> Tel:      +49/731/3994-123
> Fax:      +49/731/3994-250
> Mail:     stefan.hamerich@temic-sds.com
> Internet: http://www.temic-sds.com
> 
> > -----Original Message-----
> > From: Muralidhar [mailto:muralidhar@infoglyptic.com]
> > Sent: Wednesday, September 25, 2002 11:10 AM
> > To: W3C Voice Mailing List
> > Subject: Re: <transfer> with data
> > 
> > 
> > 
> > hai all,
> > 
> > is there any way to get the user input (ASR text),which is
> > converted into text format to validate
> > with the grammar file text.
> > 
> > 
> > with regards,
> > Murali

Received on Wednesday, 25 September 2002 08:34:17 UTC