New System Aims to Improve Speech Intelligibility

Thotra, a new startup spun off from the University of Toronto, is preparing to release a system that processes hard-to-understand speech, such as strong accents, stuttering, and the effects of some forms of cerebral palsy. The system analyzes the stream of speech similar to how speech recognition is performed, identifies specific artifacts that make the speech hard to understand, and provides some remediation before sending the speech stream on to the listener. Thotra will be implemented on mobile devices and also in the cloud.

Posted on April 9, 2014 at 7:34 am in Highlights