Novel Shadow Based Hybrid Adaptive Filter Algorithm for Speech Signal

Adaptive filters are successfully using in cancellation of echo and noise signals in radar communications, signal processing and in speech communication systems. Here we are proposing a filter with hybrid LMS algorithm which is a combination of a Fixed size LMS algorithm and Normalized step size LMS algorithm, with shadow technique for speech processing signal to cancel the noise in it.  Here we are varying to parameters namely shadow factor and combined factor to the hybrid algorithm to obtain the better statistical values of speech like signal to noise ratio (SNR) and mean square error (MSE) of the de noise speech at output of the adaptive filter. The above said values are obtained for two different cases i.e by using shadow technique and without shadow technique.