A Comparison between Power Spectral Subtraction Method and Multi-Band Spectral Subtraction Method for Enhancing Speech Corrupted by Colored Noise. In this project we investigate the “Multi-Band Spectral Subtraction (MBSS) Method for Enhancing Speech Corrupted by Colored Noise” technique proposed by Sunil D. Kamath and Philipos C. Loizou from the University of Texas at Dallas [1]. This technique is used to filter the colored noise that affects the speech spectrum non-uniformly over different frequencies and to enhance the speech spectrum. In this project we will learn to analyze speech spectrum, study the effect of colored noise to speech signal, and implement the MBSS method for enhancing speech corrupted by colored noise and in addition to the above mentioned methods. In order to evaluate the results we will implement the MBSS technique using MATLAB, and compare the results with the authors‟ results in one side, and on the other side we will implement the other technique namely Power Spectral Subtraction (PSS) and compare the results with the MBSS.
--> To find out how the algorithm performs, listen to the attached audio files before and after spectral subtraction.