Maximize speech recognition accuracy by ensuring end users have completed training sessions with the recognizer. Speaker trainingeven for command and grammar-based recognitionis essential for accuracy.
A recognition profile is a collection of information used by a speech recognition engine to increase its recognition accuracy for a specific individual's voice and environment.
A speech recognition engine saves training and background noise information to use in recognizing speech. Some speech recognition engines, such as Microsoft's Recognizer, capture and save information over time to adapt the profile for optimum speech recognition accuracy.
What is Profile Management?
Profile management enables you to:
- create speech recognition profiles,
- delete speech recognition profiles, and
- train speech recognition profiles.
Applications benefits include:
- enhanced speech recognition accuracy,
- added flexibility to adapt to different acoustical environments, and
- expanded reliability to recover from system errors.
What is ProfileKit?
Chant ProfileKit handles the complexities of managing profiles.
ProfileKit includes C++, C++Builder, Delphi, Java, and .NET Framework class library formats to support all your programming languages and sample projects for popular IDEssuch as the latest Visual Studio from Microsoft and RAD Studio from Embarcadero.
The class libraries can be integrated with 32-bit and 64-bit applications for Windows platforms.
Profile Management Architecture
ProfileKit provides a simple way to create, delete, and train speaker profiles. Applications integrate speaker profile training as part of its speech recognition administrative features.
Applications can select or create a speaker profile, train it, and delete a speaker profile. ProfileKit supports the following speech APIs for profile management:
Applications receive training notifications through event callbacks.
ProfileKit Architecture for Profile Management
ProfileKit encapsulates all of the technologies necessary to make the process of managing profiles simple and efficient.
ProfileKit simplifies the process of managing profiles by handling the low-level activities directly with the speech application program interface (API).
Instantiate ProfileKit to manage a speaker profile and destroy ProfileKit to release its resources when speaker profile management is no longer needed.
The goal of profile management is to maximize speech recognition accuracy. With ProfileKit you can:
- Create and delete speaker profiles on demand;
- Launch recognizer built-in dialogs; and
- Customize speaker training.
Chant ProfileKit handles the complexities of working with speaker profiles. This allows your application to enhance the quality of speech recognition and offer administrative features for maintaining speaker profiles.
Recognizers provide speaker training via built-in dialogs and/or custom training via API calls. ProfileKit supports the following recognizers and speaker training options:
|Recognizer||Speech API||Training Dialogs||Custom Training|
|Microsoft SAPI 5 (all languages)||SAPI 5||Yes||Yes|
|Microsoft Speech Platform (all languages)||MSP||Yes||Yes|
Within Chant Developer Workbench, you can:
- Create and delete speaker profiles;
- Enumerate speaker profiles; and
- Invoke recognizer built-in speaker training.
You may explore the capabilities of Chant ProfileKit for 30 days. To continue to use the product after 30 days, you must purchase a license for the software or stop using the software and remove it from your system.
A valid purchased license gives you the right to construct executable applications that use the applicable class library and distribute it with executable applications without royalty obligations to Chant.
The Chant ProfileKit license is a single end-user license. Each developer who installs and uses ProfileKit to develop applications must have their own license.
ProfileKit class library names vary by platform: Windows 32-bit and 64-bit. This helps ensure the correct library is deployed with your application.
ProfileKit System Requirements
- Intel processor or equivalent,
- Microsoft Windows 10,
- 120 MB of hard drive space,
- CD-ROM drive,
- VGA or higher-resolution monitor,
- Microsoft SAPI 5 recognizer,
- C++, C++Builder, Delphi, Java, or .NET Framework, and
- close-talk microphone.