Question

I am trying to get a certain grammar working in my speech recognition.

My grammar definition is as follows:

<rule id="showFlight">
<example>Show me Alaska Airlines flight number 2117</example>
<example>Where is US Airways flight 45</example>
<item>
  <one-of>
    <item>show me</item>
    <item>where is</item>
  </one-of>
</item>
<item>
  <ruleref uri="#airline" />
  <tag>out.Carrier = rules.airline;</tag>
</item>
flight
<item repeat="0-1">number</item>
<item repeat="1-">
  <ruleref uri="#digit" />
  <tag>out.Number = rules.digit;</tag>
</item>
</rule>

My problem resides with the very last — the digits. I define that 1-or-more digits can exist in the grammer, and this works. But when I go to extract the value in my OnSpeechRecognized callback, I only get the last digit spoken.

    public override bool OnSpeechRecognized(object sender, Microsoft.Speech.Recognition.SpeechRecognizedEventArgs e)
    {
        String output = String.Format("Recognition Summary:\n" +
            "  Recognized phrase: {0}\n" +
            "  Confidence score {1}\n" +
            "  Grammar used: {2}\n",
            e.Result.Text, e.Result.Confidence, e.Result.Grammar.Name);
        Console.WriteLine(output);

        // Display the semantic values in the recognition result.
        Console.WriteLine("  Semantic results:");
        //Console.WriteLine(e.Result.Semantics["Flight"].Value);

        foreach (KeyValuePair<String, SemanticValue> child in e.Result.Semantics["ShowFlight"])
        {
            Console.WriteLine("    {0} is {1}",
              child.Key, child.Value.Value ?? "null");
        }
        Console.WriteLine();

...

Or, more directly:

e.Result.Semantics["ShowFlight"]["Number"].Value.ToString()

If I say “two-one-one-seven”, the only digit in [“Number”] is 7. Likewise, if I say “four-five” the only digit I get returned is 5.

How can I extract all the numbers that are spoken that are part of the flight number?

Also, is there a secret internal grammar I can load that will allow me to recognize both “four-five” and “fortyfive” easily?

Answer 1

You can simply replace the last ‘item’ element with the following:

  <tag>out.Number = &quot;&quot;</tag>
  <item repeat="1-">
    <ruleref uri="#digit" />
    <tag>out.Number += rules.digit;</tag>
  </item>

This will concatenate all the recognized digits to out.Number.

Regarding the second question, there is no such “secret internal grammar”, unfortunately. You will have to code it yourself.