Understanding and Writing Speech Grammars
This 3 day, hands-on course will immerse you in the process and art of writing speech recognition grammars. The two most common grammar languages will be presented: XML Form Grammars and Nuance GSL.
This course will include the following topics:
- An overview of the steps involved in the life cycle of designing and tuning a speech recognition application
- An overview of the speech recognition process
- The role of recognition grammars in the process
- The difference between built-in grammars, source grammars, pre-compiled, and dynamic grammars
- Combining different grammars when necessary
- Which type of grammar to use for maximum runtime efficiency
- How to write a basic grammar
- How to modularize a grammar for readability and reusability
- Writing more complex grammars that allow more than one keyword
- Returning a semantic meaning (or ‘slots’) from the grammar to the application
- How to write hot-word or command grammars (help, stop, cancel)
- Testing that grammars do not allow nonsensical utterances
- How to compile grammars and test for errors
Duration: 3 days
Prerequisites: None