Performance Analysis and Comparison of Complex LMS, Sign LMS and RLS Algorithms for Speech Enhancement Application

Recent developments in the area of adaptive signal processing have advanced massively due to increase in powerful and cost effective digital signal processors with low cost memory chips. The uses of speech processing system for voice communication and recognition task have become more and more common. These factors lead to promote the use of digital signal processing technology for implementation of emerging applications. The process to remove unwanted interference is common and occurs in many situations. The technique of adaptive filtering is a method by which signal enhancement or noise reduction can be accomplished. An adaptive filter self adjusts its transfer function according to an optimizing algorithm. In this paper we carried out the analysis and experimentation to study the existing adaptive filter algorithms and their application for speech enhancement. The paper describes Least Mean Square (LMS) algorithm and Recursive Least Square (RLS). The complex Least Mean Square (CLMS) algorithm and the modification in CLMS lead to Sign Least Mean Square (SLMS) algorithm. The Sign-Sign Least Mean Square algorithm (SSLMS) is also considered for comparison. Normalization operation is performed on the sample which leads to evolution of NLMS algorithm. The experimentation revels that LMS have fast convergence than RLS. The computational complexity of RLS is very high as compared to LMS