A new log-spectral amplitude estimator using the weighted Euclidean distortion measure for speech enhancement

Anh Duc Nguyen, Kensuke Naoe, Yoshiyasu Takefuji

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

1 Citation (Scopus)

Abstract

This paper considers a perceptually motivated estimator for single-channel speech enhancement based on statistics and Bayesian estimation in the frequency domain. In detail, we propose a new speech log-spectral amplitude estimator where the cost function is the weighted Euclidean distortion measure of the speech log-spectral amplitude. That cost function is motivated by auditory masking effects of the human hearing system. The statistical assumptions used to develop the proposed estimator are the complex Gaussian distribution and independence of speech, noise discrete Fourier transform coefficients. We evaluate this estimator with speech signals contaminated by various noise sources at different input signal-to-noise ratios and find that it achieves better performance than the well-known minimum mean square error log-spectral amplitude estimator in terms of both noise reduction and speech quality.

Original languageEnglish
Title of host publication2010 IEEE 26th Convention of Electrical and Electronics Engineers in Israel, IEEEI 2010
Pages675-679
Number of pages5
DOIs
Publication statusPublished - 2010
Event2010 IEEE 26th Convention of Electrical and Electronics Engineers in Israel, IEEEI 2010 - Eilat, Israel
Duration: 2010 Nov 172010 Nov 20

Other

Other2010 IEEE 26th Convention of Electrical and Electronics Engineers in Israel, IEEEI 2010
CountryIsrael
CityEilat
Period10/11/1710/11/20

Fingerprint

Speech enhancement
Cost functions
Speech intelligibility
Gaussian distribution
Audition
Noise abatement
Discrete Fourier transforms
Mean square error
Signal to noise ratio
Statistics

Keywords

  • Bayesian estimation
  • Noise reduction
  • Short-time spectral amplitude
  • Speech enhancement
  • Weighted Euclidean distortion measure

ASJC Scopus subject areas

  • Electrical and Electronic Engineering

Cite this

Nguyen, A. D., Naoe, K., & Takefuji, Y. (2010). A new log-spectral amplitude estimator using the weighted Euclidean distortion measure for speech enhancement. In 2010 IEEE 26th Convention of Electrical and Electronics Engineers in Israel, IEEEI 2010 (pp. 675-679). [5662129] https://doi.org/10.1109/EEEI.2010.5662129

A new log-spectral amplitude estimator using the weighted Euclidean distortion measure for speech enhancement. / Nguyen, Anh Duc; Naoe, Kensuke; Takefuji, Yoshiyasu.

2010 IEEE 26th Convention of Electrical and Electronics Engineers in Israel, IEEEI 2010. 2010. p. 675-679 5662129.

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

Nguyen, AD, Naoe, K & Takefuji, Y 2010, A new log-spectral amplitude estimator using the weighted Euclidean distortion measure for speech enhancement. in 2010 IEEE 26th Convention of Electrical and Electronics Engineers in Israel, IEEEI 2010., 5662129, pp. 675-679, 2010 IEEE 26th Convention of Electrical and Electronics Engineers in Israel, IEEEI 2010, Eilat, Israel, 10/11/17. https://doi.org/10.1109/EEEI.2010.5662129
Nguyen AD, Naoe K, Takefuji Y. A new log-spectral amplitude estimator using the weighted Euclidean distortion measure for speech enhancement. In 2010 IEEE 26th Convention of Electrical and Electronics Engineers in Israel, IEEEI 2010. 2010. p. 675-679. 5662129 https://doi.org/10.1109/EEEI.2010.5662129
Nguyen, Anh Duc ; Naoe, Kensuke ; Takefuji, Yoshiyasu. / A new log-spectral amplitude estimator using the weighted Euclidean distortion measure for speech enhancement. 2010 IEEE 26th Convention of Electrical and Electronics Engineers in Israel, IEEEI 2010. 2010. pp. 675-679
@inproceedings{3db668394c6043cfb8cdef3404e01eb2,
title = "A new log-spectral amplitude estimator using the weighted Euclidean distortion measure for speech enhancement",
abstract = "This paper considers a perceptually motivated estimator for single-channel speech enhancement based on statistics and Bayesian estimation in the frequency domain. In detail, we propose a new speech log-spectral amplitude estimator where the cost function is the weighted Euclidean distortion measure of the speech log-spectral amplitude. That cost function is motivated by auditory masking effects of the human hearing system. The statistical assumptions used to develop the proposed estimator are the complex Gaussian distribution and independence of speech, noise discrete Fourier transform coefficients. We evaluate this estimator with speech signals contaminated by various noise sources at different input signal-to-noise ratios and find that it achieves better performance than the well-known minimum mean square error log-spectral amplitude estimator in terms of both noise reduction and speech quality.",
keywords = "Bayesian estimation, Noise reduction, Short-time spectral amplitude, Speech enhancement, Weighted Euclidean distortion measure",
author = "Nguyen, {Anh Duc} and Kensuke Naoe and Yoshiyasu Takefuji",
year = "2010",
doi = "10.1109/EEEI.2010.5662129",
language = "English",
isbn = "9781424486809",
pages = "675--679",
booktitle = "2010 IEEE 26th Convention of Electrical and Electronics Engineers in Israel, IEEEI 2010",

}

TY - GEN

T1 - A new log-spectral amplitude estimator using the weighted Euclidean distortion measure for speech enhancement

AU - Nguyen, Anh Duc

AU - Naoe, Kensuke

AU - Takefuji, Yoshiyasu

PY - 2010

Y1 - 2010

N2 - This paper considers a perceptually motivated estimator for single-channel speech enhancement based on statistics and Bayesian estimation in the frequency domain. In detail, we propose a new speech log-spectral amplitude estimator where the cost function is the weighted Euclidean distortion measure of the speech log-spectral amplitude. That cost function is motivated by auditory masking effects of the human hearing system. The statistical assumptions used to develop the proposed estimator are the complex Gaussian distribution and independence of speech, noise discrete Fourier transform coefficients. We evaluate this estimator with speech signals contaminated by various noise sources at different input signal-to-noise ratios and find that it achieves better performance than the well-known minimum mean square error log-spectral amplitude estimator in terms of both noise reduction and speech quality.

AB - This paper considers a perceptually motivated estimator for single-channel speech enhancement based on statistics and Bayesian estimation in the frequency domain. In detail, we propose a new speech log-spectral amplitude estimator where the cost function is the weighted Euclidean distortion measure of the speech log-spectral amplitude. That cost function is motivated by auditory masking effects of the human hearing system. The statistical assumptions used to develop the proposed estimator are the complex Gaussian distribution and independence of speech, noise discrete Fourier transform coefficients. We evaluate this estimator with speech signals contaminated by various noise sources at different input signal-to-noise ratios and find that it achieves better performance than the well-known minimum mean square error log-spectral amplitude estimator in terms of both noise reduction and speech quality.

KW - Bayesian estimation

KW - Noise reduction

KW - Short-time spectral amplitude

KW - Speech enhancement

KW - Weighted Euclidean distortion measure

UR - http://www.scopus.com/inward/record.url?scp=78651254026&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=78651254026&partnerID=8YFLogxK

U2 - 10.1109/EEEI.2010.5662129

DO - 10.1109/EEEI.2010.5662129

M3 - Conference contribution

SN - 9781424486809

SP - 675

EP - 679

BT - 2010 IEEE 26th Convention of Electrical and Electronics Engineers in Israel, IEEEI 2010

ER -