A Comparative Analysis for Wavelets and ThresholdEstimation Selection for Denoising of Audio Signals of Some Indian Musical Instruments

Main Article Content

Neema Verma
A. K. Verma


It is known that noise is present in all communication channels, therefore, the generated signal, when transmitted through these channels, get corrupted. Denoising of such noisy signals without loosing its features is a challenging task. The wavelet based methods has proved to be one of the best tool for denoising purposes. The proper selection of wavelet function and noise estimation algorithm is a complex task. As not all wavelet function and all noise estimation methods work well for all types of signals. In this paper an effort has been made to find a suitable wavelet function and noise estimation method to give good denoising results of audio signals from some Indian musical instruments such as Tabla, Pakhawaj, Flute, Harmonium and Taanpura. For this purpose Haar, Db10, Coif5 and Bior6.8 wavelets are considered and some well known threshold estimation methods i.e.Sqtwolog, SURE (Rigrsure and Heursure) and Minimaxi are considered for comparative analysis. The quality of denoised musical signal is expressed in terms of PSNR as compared to original signals.

Keywords: Wavelets; Denoising; Haar; Db10; Coif5; Bior6.8;Minimaxi;Rigrsure; Heursure; Square-Root-Log.


Download data is not yet available.

Article Details