2D TRANSFORM and HYBRID FILTERING APPROACH FOR NOISY SPEECH ENHANCEMENT AND COPARISON BY OBJECT AND SUB MEASURES

Abstract

This thesis presents a detailed study on speech enhancement algorithm in the transform newlinedomain. The effect of transform on speech is discussed. The main objective of newlinethis work is to attenuate the noise from a noisy speech in order to enhance the quality newlineof the processed speech using transform based filtering. The comprehensive evaluation newlineand comparison of their performances is done using different objective measures and newlinealso subjective measure for Discrete Fourier Transform (DFT), Discrete Cosine newlineTransform (DCT), and Discrete Wavelet Transform (DWT newline