Unable to PlaybackWords without generating alternate recognition results
Last reviewed: 2/12/2003
PRB Article ID: P020303
The information in this article applies to:
- SpeechKit 4
The PlaybackWords method enables you to play the audio of the words as they were originally spoken during recognition to facilitate word correction. This requires setting the CNPMaxAlternates property to a value greater than 0.
There may be application scenarios where you want to play the audio from recognition results and not support word correction (i.e. not incur the overhead of alternate generation).
Setting the CNPMaxAlternates property to a value greater than 0 causes the SAPI 5 and SMAPI component interfaces to set the recognizer properties to retain audio for playback. However there is no property specific for retaining audio.
Enhancement in SpeechKit 4 build 188.8.131.52.
A new ChantNumberProperty enumeration constant CNPRetainAudio indicates whether the recognizers should retain audio for recognition result playback using the PlaybackWords method. Set to a value greater than 0 causes audio to be retained for playback.
|CNPRetainAudio||26||Microsoft SAPI 5 and IBM SMAPI recognizer property. Indicates whether the recognizers should retain audio for recognition result playback using the PlaybackWords method.|
Because of new ChantNumberProperty constant, you may need to do the following:
- Copy new VCL libraries to DelphiX\lib and updated CDLL to Delphi\bin directory;
- Compiled languages (e.g., C/C++, C#, Java, Pascal (Delphi), Visual Basic) applications must be recompiled;
- ActiveX applications (e.g., Director, Flash, FoxPro, PowerBuilder) that may have constant declarations imbedded, need the updated definitions.
Component Formats Impacted