How Tos

Last reviewed: 7/15/2011

Article ID: H071122

HOW: Importing training documents

The information in this article applies to:

  • ProfileKit 4

Summary

ProfileKit 4 supports recognizer training using text documents.

More Information

Speech recognition engines maximize accuracy by adjusting to the acoustical properties of the speaker's audio.

Speaker training provides a way for the recognizer to sample the audio and determine runtime characteristics to aid in separating speech from noise.

Some recognizers also can improve recognition accuracy for dictating uncommon words, or uncommon word groupings, by training for the new words, or word groupings.

Microsoft SAPI 5.3+ provides a way to adapt its active language models by passing text directly to the recognizer.

The ChantPM StartTraining method provides a way to pass text data from which to train.

The following example illustrates training the recognizer with text.

// Start training with dialog hidden using text file 
NChantPM1.StartTraining("mydocument.txt", 0, ChantRecordingObject.CROFile, ChantAudioFormat.CAFText, false);