Comparative Performance Analysis of Music Audio Denoising Using Wavelet Technique

Abstract

Mostly communication channels are experiencing noises and it is very difficult to recover from the original signals without any noise. Denoising can be performed with various filtering techniques that can be separate the noise signal from the original signal. Filtering techniques have been proposed for noise removal from the digital musical signals. The overall performance of the filtering technique is less effective. Denoising of an audio signal is an area under development. In this article, music signals are considered input signals. Denoising techniques are compared based on different filter design techniques. A technique of filtering of noise is implemented and compared using Signal-to-Noise Ratio (SNR). Here for comparison, the Butterworth filter and different types of wavelet filters are used for denoising the musical signals. Daubechies wavelet techniques work better than existing methods is proven by SNR.

Keywords: Signal-to-Noise Ratio (SNR), Butterworth filter, Haar wavelet, Mean Squared Error (MSE)