Copyright © 2003 Elsevier Science B.V. All rights reserved.
Spectral-subtraction speech enhancement in multirate systems with and without non-uniform and adaptive bandwidths
Received 17 June 2002;
References and further reading may be available for this article. To view references and further reading you must purchase this article.
Abstract
The method of spectral subtraction is widely used for single-channel speech enhancement, if the speech signal is corrupted by additive noise. It is based on the manipulation of the magnitude of the noisy-speech spectrum. Most realizations use fixed, uniformly-spaced frequency transformations or, equivalently, filter banks with identical sampling rates in each frequency band. In this paper, we generalize the basic structure: Different filter-bank systems with non-uniform and, especially, non-constant, signal-adaptive spectral resolutions and, therefore, different sampling rates are examined. An efficient realization of a time-varying band allocation is proposed. The arising problems are discussed and the enhancement results are compared to each other and to those obtained with uniform spectral transformations.
Author Keywords: Noise reduction; Spectral subtraction; Filterbanks; Adaptive bandwidths
Article Outline
- 1. Introduction
- 2. Spectral subtraction
- 3. Spectral subtraction in multirate systems
- 4. Spectral subtraction with DFT-based transformations
- 5. Spectral subtraction with wavelet-based transformations
- 5.1. Discrete wavelet transformation
- 5.2. Realizations
- 5.3. Computational load
- 5.4. Experiments
- 5.4.1. Realization with the subtraction rule due to Boll
- 5.4.2. Realization with the subtraction rule due to Ephraim and Malah
- 5.5. Discrete wavelet-packet transformation
- 5.5.1. Discrete wavelet-packet analysis
- 5.5.2. Discrete wavelet-packet basis
- 5.5.3. Discrete wavelet-packet synthesis
- 5.6. Experiments
- 6. Spectral subtraction with adaptive bandwidths
- 7. Conclusion
- Acknowledgements
- References







E-mail Article
Add to my Quick Links

Cited By in Scopus (4)






