Dysphonic speech reconstruction: construction of a novel system for effective and efficient communication.

Abstract

In this article, a MELP-based real-time system that produces synthetic speech for patients with chronic dysphonia while preserving their continual structure of speech, was presented. Dysphonic patients, even though strained, could express themselves within face-to-face communications. However, when speaking in a group or using communication devices such as mobile phones, they face serious difficulties in vocal communication since it becomes harder to understand their speech. This system will make the speech of such a patient much more understandable by reconstructing the patients' dysphonic speech in real time. This system can easily be fitting into an embedded system of the size and shape of a mobile phone, so that it could be carried all the time by the patients.