Designing and testing speech recognition apps
Last reviewed: 5/1/2020
HOW Article ID: H052005
The information in this article applies to:
- SpeechKit 9
Chant Developer Workbench provides you powerful prototyping and testing environment to model and validate your application code for managing recognizers.
Managing recognizers for speech recognition is an application runtime function. Chant SpeechKit class libraries handle the complexities of speech recognition and minimize the programming efforts necessary to construct software that listens.
Within Chant Developer Workbench, you can open the Speech Recognizers browser for testing and trace callback events. This enables you to model and test your speech recognizer use before, during, and after integrating code in your applications.
Use the Recognizer browser to test one or more vocabularies as follows:
- Select the View - Speech Recognizers menu item to open the recognizer browser that lists the current speech recognizers detected on the system.
- Select the desired Speech API.
- Select the desired Speech Recognizer.
- Add command vocabularies to test by pressing the Add button.
- Dictation vocabularies are automatically be listed.
- Add grammar vocabularies to test by pressing the Add button.
- Manage the listening context by clicking on the vocabulary to display a check mark and enable the vocabulary.
- Click on the vocabulary to uncheck and disable the vocabulary.
- Select desired audio source by either entering a file name and checking File or checking Microphone for live audio input.
- Press the Start button to start recognition.
- Speak into the microphone for live audio testing.
- Press the Stop button to stop recognition.
To test command vocabularies in the Chant Developer Workbench, create a text file for each vocabulary containing a phrase on each line. Property values are separted by a comma:
red, 1 blue, 2 green, 3 orange, 4 purple, 5 yellow, 6 brown, 7 black, 8 white, 9
Within Chant Developer Workbench UWP, you can open the Speech Recognizers browser for testing and trace callback events. This enables you to model and test your speech recognizer use before, during, and after integrating code in your applications.