Applications for Speech Synthesis and Recognition

ABSTRACT

In the late '90's, Sun released an initial version of the Java Speech API (JSAPI). This API was designed to uniformly implement speech applications using commercial off the shelf (COTS) software to leverage existing speech research. Software partners have leveraged speech research advances in recognition (interpreting spoken words) and synthesis (text to speech processing).

This colloquium will be a high-level assessment of the work resulting from a recent FEAD grant, including an overview of:

• the content of the developing specifications.
• considerations when converting applications for use by people in special environments (hands-free) or those with disabilities or physical limitations.
• speech UI issues. Speech user interfaces present a new problem set, especially for applications where recognition tools historically need a fair amount of training in order to reliability interpret speech. The focus will be on UI techniques to reduce the amount of necessary recognition without sacrificing the ability to provide a powerful application front-end.
• prototyped applications using Sun's JSAPI (Java Speech API) and the W3C'S (World Wide Web Consortium) grammar and markup language specifications.

Short Bio

Trudy Howles is an Assistant Professor of Computer Science at the Rochester Institute of Technology. She served as a long time software consultant and software development engineer prior to joining the faculty at RIT. She is a member of the American Society for Quality and is a Certified Software Quality Engineer.

Colloquia Series page.