![]() ![]() Six Objective quality measures are considered in this study to test the performance of the algorithm for enhanced speech quality and compared. Analysis is done using Daubechies and Symlet wavelets with different real world noisy environments as well as white Gaussian Noise. This Algorithm incorporates the noisy speech signal divided in to overlapping frames and each frame is subject to the Multitaper Power Spectrum estimation using either SLEPIAN(Discrete Prolate Spheroidal Sequences) or SINE Tapers and applied to the Wavelet Thresholding the Multitaper Spectrum based Speech Enhancement Algorithm and the enhanced speech is reconstructed in its time domain. The results showed that, the use of Multitaper Spectrum estimation combined with the proposed Thresholding procedures yields better quality in terms of six objective quality measures. It is observed that, further improvement is observed with the proposed schemes when applied to noisy speech signals of low SNR (0dB) conditions. The performance of the new Thresholding methods is compared with the other thresholding methods. This paper investigates further improvement of Speech quality in terms of six objective quality measures using Discrete Wavelet Transform Thresholding the Multitaper Power Spectrum with the above proposed Thresholding methods. Loizou have proposed the Wavelet Thresholding the Multitaper Spectrum Speech Enhancement with Hard and Soft Thresholding schemes. To address the "musical noise" encountered in most of the frequency domain speech enhancement algorithms, Yi. The two Hybrid Thresholding schemes are formulated by incorporating the modified arctan function and exponential function suitably in Hard and Soft thresholding methods. In this paper, Modified Improved Arctan Thresholding and two Hybrid Thresholding schemes (TH-I and TH-II) have been proposed and successfully used in Wavelet Thresholding the Multitaper Spectrum for Speech Enhancement. ![]()
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |