Speech modeling and noise removal using a perceptually modified Wiener filter

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

Algorithms for spectral subtraction suffer from musical noise effects due to the large gaps in the frequency spectrum created by the subtractive process. Proposed methods to solve this problem used the auditory-masking model in the Wiener filter. Since the auditory-masking threshold (AMT) curve reveals that spectral components above it are perceptible, it can serve as a lower bound in the estimate of the short-term speech spectrum. We propose an improvement of the Wiener filter estimate using perceptual constraints that exploit the auditory masking curve. Using an LPC model, from psychoacoustics we derive an estimate of the spectral density of speech that tends to lower and spread the energy of the musical noise onto other frequencies in the critical band. Objective and subjective evaluations indicate a slightly improved performance over ordinary spectral subtraction and Wiener filtering methods.

Original languageEnglish (US)
Title of host publicationProceedings of the 4th IASTED International Conference on Signal Processing, Pattern Recognition, and Applications, SPPRA 2007
Pages304-308
Number of pages5
StatePublished - 2007
Event4th IASTED International Conference on Signal Processing, Pattern Recognition, and Applications, SPPRA 2007 - Innsbruck, Austria
Duration: Feb 14 2007Feb 16 2007

Other

Other4th IASTED International Conference on Signal Processing, Pattern Recognition, and Applications, SPPRA 2007
CountryAustria
CityInnsbruck
Period2/14/072/16/07

Fingerprint

Spectral density

Keywords

  • DSP
  • Speech processing

ASJC Scopus subject areas

  • Computer Science Applications
  • Computer Vision and Pattern Recognition
  • Signal Processing

Cite this

Pinto, R., Spanias, A., & Berisha, V. (2007). Speech modeling and noise removal using a perceptually modified Wiener filter. In Proceedings of the 4th IASTED International Conference on Signal Processing, Pattern Recognition, and Applications, SPPRA 2007 (pp. 304-308)

Speech modeling and noise removal using a perceptually modified Wiener filter. / Pinto, Ryan; Spanias, Andreas; Berisha, Visar.

Proceedings of the 4th IASTED International Conference on Signal Processing, Pattern Recognition, and Applications, SPPRA 2007. 2007. p. 304-308.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Pinto, R, Spanias, A & Berisha, V 2007, Speech modeling and noise removal using a perceptually modified Wiener filter. in Proceedings of the 4th IASTED International Conference on Signal Processing, Pattern Recognition, and Applications, SPPRA 2007. pp. 304-308, 4th IASTED International Conference on Signal Processing, Pattern Recognition, and Applications, SPPRA 2007, Innsbruck, Austria, 2/14/07.
Pinto R, Spanias A, Berisha V. Speech modeling and noise removal using a perceptually modified Wiener filter. In Proceedings of the 4th IASTED International Conference on Signal Processing, Pattern Recognition, and Applications, SPPRA 2007. 2007. p. 304-308
Pinto, Ryan ; Spanias, Andreas ; Berisha, Visar. / Speech modeling and noise removal using a perceptually modified Wiener filter. Proceedings of the 4th IASTED International Conference on Signal Processing, Pattern Recognition, and Applications, SPPRA 2007. 2007. pp. 304-308
@inproceedings{256dd6dcec374bc98a9e5e4e7ba1e7bc,
title = "Speech modeling and noise removal using a perceptually modified Wiener filter",
abstract = "Algorithms for spectral subtraction suffer from musical noise effects due to the large gaps in the frequency spectrum created by the subtractive process. Proposed methods to solve this problem used the auditory-masking model in the Wiener filter. Since the auditory-masking threshold (AMT) curve reveals that spectral components above it are perceptible, it can serve as a lower bound in the estimate of the short-term speech spectrum. We propose an improvement of the Wiener filter estimate using perceptual constraints that exploit the auditory masking curve. Using an LPC model, from psychoacoustics we derive an estimate of the spectral density of speech that tends to lower and spread the energy of the musical noise onto other frequencies in the critical band. Objective and subjective evaluations indicate a slightly improved performance over ordinary spectral subtraction and Wiener filtering methods.",
keywords = "DSP, Speech processing",
author = "Ryan Pinto and Andreas Spanias and Visar Berisha",
year = "2007",
language = "English (US)",
isbn = "9780889866461",
pages = "304--308",
booktitle = "Proceedings of the 4th IASTED International Conference on Signal Processing, Pattern Recognition, and Applications, SPPRA 2007",

}

TY - GEN

T1 - Speech modeling and noise removal using a perceptually modified Wiener filter

AU - Pinto, Ryan

AU - Spanias, Andreas

AU - Berisha, Visar

PY - 2007

Y1 - 2007

N2 - Algorithms for spectral subtraction suffer from musical noise effects due to the large gaps in the frequency spectrum created by the subtractive process. Proposed methods to solve this problem used the auditory-masking model in the Wiener filter. Since the auditory-masking threshold (AMT) curve reveals that spectral components above it are perceptible, it can serve as a lower bound in the estimate of the short-term speech spectrum. We propose an improvement of the Wiener filter estimate using perceptual constraints that exploit the auditory masking curve. Using an LPC model, from psychoacoustics we derive an estimate of the spectral density of speech that tends to lower and spread the energy of the musical noise onto other frequencies in the critical band. Objective and subjective evaluations indicate a slightly improved performance over ordinary spectral subtraction and Wiener filtering methods.

AB - Algorithms for spectral subtraction suffer from musical noise effects due to the large gaps in the frequency spectrum created by the subtractive process. Proposed methods to solve this problem used the auditory-masking model in the Wiener filter. Since the auditory-masking threshold (AMT) curve reveals that spectral components above it are perceptible, it can serve as a lower bound in the estimate of the short-term speech spectrum. We propose an improvement of the Wiener filter estimate using perceptual constraints that exploit the auditory masking curve. Using an LPC model, from psychoacoustics we derive an estimate of the spectral density of speech that tends to lower and spread the energy of the musical noise onto other frequencies in the critical band. Objective and subjective evaluations indicate a slightly improved performance over ordinary spectral subtraction and Wiener filtering methods.

KW - DSP

KW - Speech processing

UR - http://www.scopus.com/inward/record.url?scp=56549113073&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=56549113073&partnerID=8YFLogxK

M3 - Conference contribution

SN - 9780889866461

SP - 304

EP - 308

BT - Proceedings of the 4th IASTED International Conference on Signal Processing, Pattern Recognition, and Applications, SPPRA 2007

ER -