RecognizedWordUnit.LexicalForm Property

public:
 property System::String ^ LexicalForm { System::String ^ get(); };

public string LexicalForm { get; }

member this.LexicalForm : string

Public ReadOnly Property LexicalForm As String

Property Value

Returns a String containing the text of a recognized word, without any normalization.

Examples

The following example shows a utility routine that generates text in one of three formats: lexical (using LexicalForm), normalized (using Text), and phonetic (using Pronunciation). The text output is obtained from a ReadOnlyCollection<T> of RecognizedWordUnit objects, which is obtained from the Words property on the RecognizedPhrase object.

internal enum WordType
{
  Text,
  Normalized = Text,
  Lexical,
  Pronunciation
}

internal static string stringFromWordArray(
         ReadOnlyCollection<RecognizedWordUnit> words,
         WordType type)
{
  string text = "";
  foreach (RecognizedWordUnit word in words)
  {
    string wordText = "";
    if (type == WordType.Text || type == WordType.Normalized)
    {
      wordText = word.Text;
    }
    else if (type == WordType.Lexical)
    {
      wordText = word.LexicalForm;
    }
    else if (type == WordType.Pronunciation)
    {
      wordText = word.Pronunciation;
    }
    else
    {
      throw new InvalidEnumArgumentException(
          String.Format("[0}: is not a valid input", type));
    }

    // Use display attribute
    if ((word.DisplayAttributes & DisplayAttributes.OneTrailingSpace) != 0)
    {
      wordText += " ";
    }
    if ((word.DisplayAttributes & DisplayAttributes.TwoTrailingSpaces) != 0)
    {
      wordText += "  ";
    }
    if ((word.DisplayAttributes & DisplayAttributes.ConsumeLeadingSpaces) != 0)
    {
      wordText = wordText.TrimStart();
    }
    if ((word.DisplayAttributes & DisplayAttributes.ZeroTrailingSpaces) != 0)
    {
    wordText = wordText.TrimEnd();
    }

    text += wordText;

  }
  return text;
}

Remarks

In most cases the values returned by Text and LexicalForm are identical. However, recognition engines may use speech normalization to return more user-friendly or colloquial text representations of audio input.

Speech normalization is the use of special constructs or symbols to express speech in writing. For example, normalization can replace the spoken words "a dollar and sixteen cents" with "$1.16" in output text.

Applies to

See also

• DisplayAttributes