A Noise Reduction Method Based on LMS Adaptive Filter of Audio Signals

Abstract

Noise reduction of audio signals is a key challenge problem in speech enhancement, speech recognition and speech communication applications, etc. It has attracted a considerable amount of research attention over past several dec- ades. The most widely used method is optimal linear filtering method, which achieves clean audio estimate by passing the noise observation through an optimal linear filter or transformation. The representative algorithms include Wiener filter- ing, Kalman filtering, spectral restoration, subspace method, etc. Many theoretical analysis and experiments have been carried out to show that the optimal filtering technique can reduce the level of noise that is present in the audio signals and im- prove the corresponding signal-to-noise ratio (SNR). However, one of the main problems for optimal filtering method is complexity of the algorithm which based upon SVD–decompositions or QR–decompositions. In almost real signal applica- tions it difficult to implement. In this paper, a method for reducing noise from au- dio or speech signals using LMS adaptive filtering algorithm is proposed. The sig- nal is filtered in the time domain, while the filter coefficients are calculated adaptively by steepest-descent algorithm. The simulation results exhibit a higher quality of the processed signal than unprocessed signal in the noise situation.