Automatic Detection and Removal of Impulsive Noise in Audio Signals
Laurent Oudre
published
2015-11-21
reference
Laurent Oudre, Automatic Detection and Removal of Impulsive Noise in Audio Signals, Image Processing On Line, 5 (2015), pp. 267–281. https://doi.org/10.5201/ipol.2015.64

Communicated by Gaël Richard and Rafael Grompone von Gioi
Demo edited by Laurent Oudre

Abstract

This article presents a method for restoring audio signals corrupted by impulsive noise such as clicks, bursts or scratches. The algorithm takes as input a degraded audio signal and automatically detects the locations of the degraded samples and replaces them with more appropriate values. Both steps (detection and interpolation) are based on the assumption that the signal can locally be modeled as a realization of an autoregressive process. Surprisingly, the results obtained on several types of signals (classical, jazz, vocal, etc.) show that a fully automatic method, with a carefully fixed set of parameters, can achieve good performance on a wide range of degraded audio signals.

Download

History