Publication detail

Single-Channel Noise Suppression by Wavelets in Spectral Domain

SMÉKAL, Z. SYSEL, P.

Original Title

Single-Channel Noise Suppression by Wavelets in Spectral Domain

English Title

Single-Channel Noise Suppression by Wavelets in Spectral Domain

Type

journal article in Web of Science

Language

en

Original Abstract

The paper describes the design of a new single-channel method for speech enhancement that employs the wavelet transform. Signal decomposition is currently performed in the time domain while noise is removed on individual decomposition levels using thresholding techniques. Here the wavelet transform is applied in the spectral domain. Used as the basis is the method of spectral subtraction, which is suitable for real-time implementation because of its simplicity. The greatest problem in the spectral subtraction method is a trustworthy noise estimate, in particular when non-stationary noise is concerned. Using the wavelet transform we can achieve a more accurate power spectral density also of noise that is non-stationary. Listening tests and SNR measurements yield satisfactory results in comparison with earlier reported experience.

English abstract

The paper describes the design of a new single-channel method for speech enhancement that employs the wavelet transform. Signal decomposition is currently performed in the time domain while noise is removed on individual decomposition levels using thresholding techniques. Here the wavelet transform is applied in the spectral domain. Used as the basis is the method of spectral subtraction, which is suitable for real-time implementation because of its simplicity. The greatest problem in the spectral subtraction method is a trustworthy noise estimate, in particular when non-stationary noise is concerned. Using the wavelet transform we can achieve a more accurate power spectral density also of noise that is non-stationary. Listening tests and SNR measurements yield satisfactory results in comparison with earlier reported experience.

Keywords

Single-channel Speech Enhancement, Power Spectral Density, Wavelet Transform Thresholding

RIV year

2007

Released

08.10.2007

Publisher

Springer

Location

Vietri sul Mare, Italy

ISBN

0302-9743

Periodical

Lecture Notes in Computer Science

Year of study

4775

State

DE

Pages from

150

Pages to

164

Pages count

15

Documents

BibTex


@article{BUT47276,
  author="Zdeněk {Smékal} and Petr {Sysel}",
  title="Single-Channel Noise Suppression by Wavelets in Spectral Domain",
  annote="The paper describes the design of a new single-channel method for speech enhancement that employs the wavelet transform. Signal decomposition is currently performed in the time domain while noise is removed on individual decomposition levels using thresholding techniques. Here the wavelet transform is applied in the spectral domain. Used as the basis is the method of spectral subtraction, which is suitable for real-time implementation because of its simplicity. The greatest problem in the spectral subtraction method is a trustworthy noise estimate, in particular when non-stationary noise is concerned. Using the wavelet transform we can achieve a more accurate power spectral density also of noise that is non-stationary. Listening tests and SNR measurements yield satisfactory results in comparison with earlier reported experience.",
  address="Springer",
  chapter="47276",
  institution="Springer",
  journal="Lecture Notes in Computer Science (IF 0,513)",
  volume="4775",
  year="2007",
  month="october",
  pages="150--164",
  publisher="Springer",
  type="journal article in Web of Science"
}