Speech enhancement hybrid of wiener filter and haar wavelet algorithm using DSP

Abstract

The most common, natural and effective form of communication between human is speech. Generally, during speech communication a lot of unwanted noise were corrupted and make speech signal weak. Due to this enhancing or de-noising to speech signal become very important. In this project the speech enhancement hybrid of wiener filter and haar wavelet algorithm will perform degrading the unwanted noise which play in speech signal. The algorithm will implement on the floating point Digital Signal Processing processor by using c programming language. MATLAB is used to get the coefficients of process. The main objective of this project is to enhance the speech signal by using hybrid of wiener filtering and compressing with haar wavelet. In the filtering the use of TSNR and HRNR techniques reduced the noise and regenerated the harmonic signal in speech. Then by using Haar wavelet transform algorithm to compress the noise. By using PSNR measurement to compare the result. This result can open the doors for new future studies to try for applying other process like speech recognition and sound classification.