Improvement of audio noise reduction system based on RLS algorithm

As various noises seriously affect the quality of our life, the noise cancellation and suppression has been an important task. After decades of development, the researchers have come up with a variety of effective methods to reduce certain types of noises. However, the noise presenting in life are basically a mixture of different kinds of noise. This paper examines a phenomenon of periodic noise mixing with impulse noise. In order to cancel or suppress mixed noise, using the method of spectrogram to detect and eliminate of impulse noise, and then sending the remaining mixed noise into improved RLS adaptive filtering system to eliminate periodic noise. Through the method described in this paper, SNR can achieve 12dB or more, and the noise reduction coefficient can reach more than 0.9.

[1]  H. Wu,et al.  Adaptive impulse detection using center-weighted median filters , 2001, IEEE Signal Processing Letters.

[2]  A Median Filter Based on Judging Impulse Noise by Statistic and Adaptive Threshold , 2008, 2008 Congress on Image and Signal Processing.

[3]  Yuriy V. Zakharov,et al.  Sparse RLS adaptive filter with diagonal loading , 2012, 2012 Conference Record of the Forty Sixth Asilomar Conference on Signals, Systems and Computers (ASILOMAR).

[4]  Zhang Xiaofang.  Analysis and Simulation of a Variable Forgetting Factor RLS Algorithm , 2008 .

[5]  M. Portnoff.  Time-frequency representation of digital signals and systems based on short-time Fourier analysis , 1980 .

[6]  B. Widrow,et al.  Adaptive noise cancelling: Principles and applications , 1975 .

[7]  Qu Qiang.  An Improved HRLS Algorithm , 2008 .

[8]  Rajeev C. Nongpiur,et al.  Impulse noise removal in speech using wavelets , 2008, 2008 IEEE International Conference on Acoustics, Speech and Signal Processing.

[9]  Neil Robertson,et al.  21st European Signal Processing Conference , 2013 .

[10]  Shuqun Zhang,et al.  A new impulse detector for switching median filters , 2002, IEEE Signal Processing Letters.