Improving dysarthric speech intelligibility through re-synthesized and grafted units

In this paper, an assistive speech-enabled system is proposed to help dysarthric speakers. This system aims at improving the intelligibility of dysarthric speech making it as natural as possible and close to the original voice of the dysarthric speaker. The resynthesized utterances use new basic units, a new concatenating algorithm and a grafting technique to correct the bad pronounced phonemes. The Nemours database is used to evaluate the proposed assistive communication system. The results of subjective evaluations show that the proposed technique reaches the very satisfactory MOS (mean opinion score) of 4.

1. INTRODUCTION

Dysarthria is a motor speech disorder affecting millions of people. A dysarthric speaker has much difficulty to communicating. This disorder induces bad or not pronounced phonemes, variable speech amplitude, poor articulation, etc. According to Aronson [1], dysarthria covers various speech troubles resulting from neurological disorders. These troubles are linked to the disturbance of brain and nerve stimulus of the muscles involved in the production of speech. As a result, dysarthric speakers suffer from weakness, slowness, and impaired muscle tone during the production of speech. The organs of speech production may be affected to varying degrees. Thus, the reduction of intelligibility is a common disruption to the various forms of dysarthria.