Performance Comparison of Transform Domain Adaptive Filters

Abstract In this paper, transform domain adaptive filter and its orthogonality properties such as Discrete Cosine Least Mean Square Transform (DCLMST), Discrete Sine Least Mean Square Transform (DSLMST), Discrete Fourier Least Mean Square Transform (DFLMST), Wavelet Domain Least Mean Square Adaptive Filter (WDLMSAF) and Wavelet Domain Normalized Least Mean Square Adaptive Filter (WDNLMSAF) are studied and compared on the basis of different performance parameters. In this we are using block implementation for FIR filter, to increase the computational speed. Block implementation in frequency domain is performed by fast block LMS algorithm. Their performance is compared on the basis of signal to noise (SNR) ratio, mean square error (MSE) and computational time. The performance comparison is taken for different noise range (5 dB, 0 dB, -5 dB and -10 dB). We have divided the noise level in to different order according to there range: lower order (10-40), moderate order (40-60) and higher order (60-100). WDLMSAF shows best result for MSE and SNR . A graphical and a tabular view is also given to show the comparison.