TextFragment Class

Definition

public ref class TextFragment

public class TextFragment

type TextFragment = class

Public Class TextFragment

Inheritance

Object

TextFragment

Examples

The example below is part of a custom speech synthesis implementation inheriting from TtsEngineSsml, and using the use of TextFragment, SpeechEventInfo, FragmentState, and TtsEventId.

The implementation of Speak

1. Receives an array of TextFragment instances and creates a new array of TextFragment instances to be passed to the Speak method on an underlying synthesis engine.

   Particular care is used to respect the TextOffset, TextLength on the original TextFragment when creating the TextToSpeak on the new TextFragment instances.

2. If the TtsEngineAction enumeration value by found from the Action property on the FragmentState returned by the State property of each TextFragment instance is Speak, the implementation

   • Translates Americanism to Britishisms in the text to be spoken.

   • If the EventInterest property on the ITtsEngineSite interfaces provided to the implementation support the WordBoundary event type, a SpeechEventInfo instance is used to create an event to drive a synthesizer progress meter is created.

3. A speech rendering engine is then called with the modified TextFragment array.

private const int WordBoundaryFlag = 1 << (int)TtsEventId.WordBoundary;  
private readonly char[] spaces = new char[] { ' ', '\t', '\r', '\n' };  
internal struct UsVsUk  
{  
  internal string UK;  
  internal string US;  
}  

override public void Speak (TextFragment [] frags, IntPtr wfx, ITtsEngineSite site)  
{  
  TextFragment [] newFrags=new TextFragment[frags.Length];  

  for (int i=0;i<frags.Length;i++){  
    newFrags[i].State=frags[i].State;  
    //truncate  
    newFrags[i].TextToSpeak = frags[i].TextToSpeak.Substring(frags[i].TextOffset,  
                               frags[i].TextLength);  
    newFrags[i].TextLength = newFrags[i].TextToSpeak.Length;  
    newFrags[i].TextOffset = 0;  
    if (newFrags[i].State.Action == TtsEngineAction.Speak) {  
      //Us to UK conversion  
      foreach (UsVsUk term in TransList) {  
      newFrags[i].TextToSpeak.Replace(term.US, term.UK);  
      }  
      //Generate progress meter events if supported  
      if ((site.EventInterest & WordBoundaryFlag) != 0) {  
      string[] subs = newFrags[i].TextToSpeak.Split(spaces);  

      foreach (string s in subs) {  
        int offset = newFrags[i].TextOffset;  
        SpeechEventInfo spEvent = new SpeechEventInfo((Int16)TtsEventId.WordBoundary,   
                (Int16)EventParameterType.Undefined,   
                 s.Length, new IntPtr(offset));  
        offset += s.Length;  
        if (s.Trim().Length > 0) {  
          SpeechEventInfo[] events = new SpeechEventInfo[1];  
          events[0] = spEvent;  
          site.AddEvents(events, 1);  
        }  
      }  
      }  
    }  
  }  

  _baseSynthesize.Speak(newFrags, wfx, site);  

}  

Remarks

The Speech platform infrastructure unpacks the XML based structure of the SSML input and constructs TextFragment objects.

Speech content is available through the TextLength, TextOffset, and TextToSpeak properties of a TextFragment instance.

Speech attribute information, such as emphasis, pitch, and rate, are obtained from the FragmentState object returned by the TextFragmentState property.

Constructors

Properties

Methods

Applies to