Empirical mode decomposition and normalshrink tresholding for speech denoising

TL;DR

This paper introduces a speech denoising method using Empirical Mode Decomposition (EMD) combined with thresholding, demonstrating improved signal quality over traditional methods through extensive simulations.

Contribution

It presents a fully data-driven speech denoising approach based on EMD and thresholding, addressing signal degradation issues and analyzing noise level effects.

Findings

EMD-based denoising improves SNR in speech signals

The method outperforms traditional thresholding techniques

Performance varies with noise level

Abstract

In this paper a signal denoising scheme based on Empirical mode decomposition (EMD) is presented. The denoising method is a fully data driven approach. Noisy signal is decomposed adaptively into intrinsic oscillatory components called Intrinsic mode functions (IMFs) using a decomposition algorithm called sifting process. The basic principle of the method is to decompose a speech signal into segments each frame is categorised as either signal-dominant or noise-dominant then reconstruct the signal with IMFs signal dominant frame previously filtered or thresholded. It is shown, on the basis of intensive simulations that EMD improves the signal to noise ratio and address the problem of signal degradation. The denoising method is applied to real signal with different noise levels and the results compared to Winner and universal threshold of DONOHO and JOHNSTONE [11] with soft and hard tresholding. The effect of level noise value on the performances of the proposed denoising is analysed. The study is limited to signals corrupted by additive white Gaussian random noise.