locked
Need Kinect Speech Grammar Clarification? RRS feed

  • Question

  • Hi,

    I am working with Kinect speech recognition and designed a SRGS/SISR based grammar file. (only part of the grammar displayed here). If the user says "Listen or Start Listening" (the items in first rule), I am able to get the recognized text as "Listen". But my requirement is to get the exact user utterance and display that. For eg: If user says "Start listening to my voice", I should display "Start listening to my voice" as user utterance and "Listen" as normalized grammar text.

    Can anyone help how to get and display the exact user utterance which is not included as part of the  grammar file.

    <rule id="BaseRule" scope="public"> <one-of> <item> <ruleref uri="#SpeechActionListen"/> </item> <item> <ruleref uri="#SpeechActionStop"/> </item> <item> </one-of> <tag>out.Obj = rules.latest().Text</tag> </rule> <rule id="SpeechActionListen" scope="public"> <one-of> <item>Listen</item> <item>Start Listening</item> </one-of> <tag>out.Text="Listen </rule> <rule id="SpeechActionStop" scope="public"> <one-of> <item>Go To Sleep</item> <item>Sleep</item> <item>Stop Listening</item> </one-of> <tag>out.Text="Sleep"</tag> </rule>

     

     private void sre_SpeechRecognized(object sender, SpeechRecognizedEventArgs e)
            {
     string strSpeech = e.Result.Text;
    }

    Thanks,

    Bharat.


    Tuesday, November 6, 2012 7:05 AM

Answers

  • Dictation (recognizing words not in a grammar) is not currently supported with Speech Server 11 on Kinect for Windows. We are currently limitted to command and control style scenarios, where we know all words that the user is expected to speak.
    Friday, November 30, 2012 10:01 PM

All replies

  • Hi All

    Can anyone help on this part?

    Thanks,

    Bharat.

    Friday, November 23, 2012 10:17 AM
  • Dictation (recognizing words not in a grammar) is not currently supported with Speech Server 11 on Kinect for Windows. We are currently limitted to command and control style scenarios, where we know all words that the user is expected to speak.
    Friday, November 30, 2012 10:01 PM
  • How about the normal Windows Server / desktop? The SDK manual shows the

    grammarbuilder.AppendDictation() method, so it should work?

    Sunday, January 27, 2019 3:50 AM