How Tos

Last reviewed: 7/15/2011

Article ID: H071121

HOW: Developing VoCon 3200 V4 applications that manage profiles

The information in this article applies to:

  • ProfileKit 4

Summary

ProfileKit 4 provides support for VoCon 3200 V4 that simplifies the process of developing training sessions for speaker adaptation.

More Information

Chant ProfileKit 4 provides support for training of VoCon 3200 V4 speaker profiles.

Once training session data is persisted, it can be used during recognition to help improve matching accuracy.

A training session is an interactive session with VoCon where the end-user is prompted to speak phrases. Information is retained by VoCon about the session to use with subsequent recognition sessions.

ProfileKit simplifies the process of running a training session with VoCon and saving the data to a .spa file. This file is set as the CSPSpeaker property on SpeechKit ChantSR recognition object sessions with VoCon.

private Chant.ProfileKit.NChantPM NChantPM1;

// Instantiate NChantPM object
NChantPM1 = new NChantPM(this);
//NChantPM1.SetStringProperty(ChantStringProperty.CSPLicense, "LicenseRegistrationNumber");
//NChantPM1.SetStringProperty(ChantStringProperty.CSPSerials, "LicenseSerialNumber");
NChantPM1.SetNumberProperty(ChantNumberProperty.CNPEngineAPI, (int)ChantEngineAPI.CEVoCon4);
NChantPM1.SetStringProperty(ChantStringProperty.CSPEnginePath, "C:\\Nuance\\VoCon Hybrid\\SDK_v4_3\\bin\\winx86rls");
// Set the acoustic model
NChantPM1.SetStringProperty(ChantStringProperty.CSPAcousticModel, "C:\\Nuance\\VoCon Hybrid\\SDK_v4_3\\models\\acmod4_900_enu_gen_car_f16_v1_0_0.dat");
// Set either the common linguistic component or data driven G2P model files
NChantPM1.SetStringProperty(ChantStringProperty.CSPCLCModel, "C:\\Nuance\\VoCon Hybrid\\SDK_v4_3\\models\\clc_enu_cfg3_v1_0_7.dat");

// Set the current speaker
NChantPM1.SetStringProperty(ChantStringProperty.CSPSpeaker, "DefaultSpeaker");

// Example 1
// Training the speaker with ProfileKit grammar training
NChantPM1.SetTrainingProperty(ChantTrainingProperty.CTPTrainingPhraseText, "red\nblue\norange\ngreen\npurple\nyellow\nbrown\n");

// Start training with dialog hidden 
NChantPM1.StartTraining("", 0, ChantRecordingObject.CROMultiMedia, ChantAudioFormat.CAFDefault, false);

// Example 2
// Training the speaker with ProfileKit using a grammar file
NChantPM1.SetTrainingProperty(ChantTrainingProperty.CTPTrainingGrammarVocab, "colors.bnf");

// Start training with dialog hidden using pre-recorded audio file 
NChantPM1.StartTraining("mytraining.wav", 0, ChantRecordingObject.CROFile, ChantAudioFormat.CAFDefault, false);