An Improved Method for Speech Enhancement Based on Ridgelet Transform

In this paper, a speech enhancement technique proposed by Soon and Koh is examined and improved by exploiting different window functions for pre-processing of speech signals. In this method, instead of using two-dimensional (2-D) discrete Fourier Transform (DFT), Ridgelet transform is employed with different window through spectral magnitude subtraction method. A comparative study of performance of different window functions such as Hanning, Hamming, Blackman, Cosh and Exponential windows and other transforms such as discrete Fourier transform (DFT), discrete cosine transform (DCT), discrete wavelet transform (DWT), Ridgelet transform (RT) has been made in terms of signal to noise (SNR) ratio, maximum error (ME), and mean square error (MSE). When compared, Cosh window and Ridgelet transform (RT) gives the best performance than all other known window functions and other transforms.