Saoud Safa ¹, Bennasr Mouhamed ¹, Cherif Adnen ¹

Affiliations
1 Science Faculty of Tunis, University of Tunis El-Manar, Laboratory of Analysis and Processing of Signals and Electric and Energy System, 1060 Tunis, TN

Abstract

Background/Objectives: The speech signal is an attractive research issue that solves signal noises. For this reason, the paper main goal is to effectively ensure speech quality. An accurate Optimized Speech Enhancement Algorithm (OSEA) has been proposed and implemented to improve speech intelligibility and quality. The proposed approach was developed and implemented through a Discrete Wavelet Transform (DWT) and kalman filter methods. To assess speech quality, various noisy types for each SNR level were used. The obtained results were compared with a wavelet based speech enhancement using objective assessments like SNR, NRMSE and PESQ. Indeed, the obtained results demonstrated a better speech quality .All prepared tests have been implemented on a TMS320C6416 fixed –point digital signal processor which appears to be suitable platform for real-time requirements.

Keywords

Digital Signal Processor, Discrete Wavelet Transform, Kalman Filter, Speech Enhancement, Real-Time

Full Text