Noise estimation for speech enhancement using minimum-spectral-average and vowel-presence detection approach

Ching Ta Lu, Yung-Yu Chen, Jun Hong Shen, Ling Ling Wang, Chung Lin Lei

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

Abstract

The accuracy of noise estimation is important for the performance of a speech enhancement system. This study proposes using variable segment length for noise tracking and variable thresholds for the determination of speech-presence probability. Initially, the fundamental frequency is estimated to determine whether a frame is a vowel. In the case of a vowel frame, the segment length increases; meanwhile the threshold for speech-presence is decreased. So the noise magnitude is adequately underestimated. The speech distortion is accordingly reduced in enhanced speech. Conversely, the segment length is rapidly decreased during noise-dominant regions. This enables the noise estimate to be updated quickly and the noise variation to be well tracked, yielding background noise being efficiently removed by the process of speech enhancement. Experimental results show that the proposed method can efficiently track the variation of background noise, enabling the performance of speech enhancement to be improved.

Original languageEnglish
Title of host publicationFrontier Computing - Theory, Technologies and Applications, FC 2016
EditorsNeil Y. Yen, Jason C. Hung
PublisherSpringer Verlag
Pages317-327
Number of pages11
ISBN (Print)9789811031861
DOIs
Publication statusPublished - 2018 Jan 1
Event 5th International Conference on Frontier Computing, FC 2016 - Tokyo, Japan
Duration: 2016 Jul 132016 Jul 15

Publication series

NameLecture Notes in Electrical Engineering
Volume422
ISSN (Print)1876-1100
ISSN (Electronic)1876-1119

Other

Other 5th International Conference on Frontier Computing, FC 2016
CountryJapan
CityTokyo
Period16-07-1316-07-15

Fingerprint

Speech enhancement

All Science Journal Classification (ASJC) codes

  • Industrial and Manufacturing Engineering

Cite this

Lu, C. T., Chen, Y-Y., Shen, J. H., Wang, L. L., & Lei, C. L. (2018). Noise estimation for speech enhancement using minimum-spectral-average and vowel-presence detection approach. In N. Y. Yen, & J. C. Hung (Eds.), Frontier Computing - Theory, Technologies and Applications, FC 2016 (pp. 317-327). (Lecture Notes in Electrical Engineering; Vol. 422). Springer Verlag. https://doi.org/10.1007/978-981-10-3187-8_32
Lu, Ching Ta ; Chen, Yung-Yu ; Shen, Jun Hong ; Wang, Ling Ling ; Lei, Chung Lin. / Noise estimation for speech enhancement using minimum-spectral-average and vowel-presence detection approach. Frontier Computing - Theory, Technologies and Applications, FC 2016. editor / Neil Y. Yen ; Jason C. Hung. Springer Verlag, 2018. pp. 317-327 (Lecture Notes in Electrical Engineering).
@inproceedings{05b7f1a98c0f45b69166d48209422bbf,
title = "Noise estimation for speech enhancement using minimum-spectral-average and vowel-presence detection approach",
abstract = "The accuracy of noise estimation is important for the performance of a speech enhancement system. This study proposes using variable segment length for noise tracking and variable thresholds for the determination of speech-presence probability. Initially, the fundamental frequency is estimated to determine whether a frame is a vowel. In the case of a vowel frame, the segment length increases; meanwhile the threshold for speech-presence is decreased. So the noise magnitude is adequately underestimated. The speech distortion is accordingly reduced in enhanced speech. Conversely, the segment length is rapidly decreased during noise-dominant regions. This enables the noise estimate to be updated quickly and the noise variation to be well tracked, yielding background noise being efficiently removed by the process of speech enhancement. Experimental results show that the proposed method can efficiently track the variation of background noise, enabling the performance of speech enhancement to be improved.",
author = "Lu, {Ching Ta} and Yung-Yu Chen and Shen, {Jun Hong} and Wang, {Ling Ling} and Lei, {Chung Lin}",
year = "2018",
month = "1",
day = "1",
doi = "10.1007/978-981-10-3187-8_32",
language = "English",
isbn = "9789811031861",
series = "Lecture Notes in Electrical Engineering",
publisher = "Springer Verlag",
pages = "317--327",
editor = "Yen, {Neil Y.} and Hung, {Jason C.}",
booktitle = "Frontier Computing - Theory, Technologies and Applications, FC 2016",
address = "Germany",

}

Lu, CT, Chen, Y-Y, Shen, JH, Wang, LL & Lei, CL 2018, Noise estimation for speech enhancement using minimum-spectral-average and vowel-presence detection approach. in NY Yen & JC Hung (eds), Frontier Computing - Theory, Technologies and Applications, FC 2016. Lecture Notes in Electrical Engineering, vol. 422, Springer Verlag, pp. 317-327, 5th International Conference on Frontier Computing, FC 2016, Tokyo, Japan, 16-07-13. https://doi.org/10.1007/978-981-10-3187-8_32

Noise estimation for speech enhancement using minimum-spectral-average and vowel-presence detection approach. / Lu, Ching Ta; Chen, Yung-Yu; Shen, Jun Hong; Wang, Ling Ling; Lei, Chung Lin.

Frontier Computing - Theory, Technologies and Applications, FC 2016. ed. / Neil Y. Yen; Jason C. Hung. Springer Verlag, 2018. p. 317-327 (Lecture Notes in Electrical Engineering; Vol. 422).

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

TY - GEN

T1 - Noise estimation for speech enhancement using minimum-spectral-average and vowel-presence detection approach

AU - Lu, Ching Ta

AU - Chen, Yung-Yu

AU - Shen, Jun Hong

AU - Wang, Ling Ling

AU - Lei, Chung Lin

PY - 2018/1/1

Y1 - 2018/1/1

N2 - The accuracy of noise estimation is important for the performance of a speech enhancement system. This study proposes using variable segment length for noise tracking and variable thresholds for the determination of speech-presence probability. Initially, the fundamental frequency is estimated to determine whether a frame is a vowel. In the case of a vowel frame, the segment length increases; meanwhile the threshold for speech-presence is decreased. So the noise magnitude is adequately underestimated. The speech distortion is accordingly reduced in enhanced speech. Conversely, the segment length is rapidly decreased during noise-dominant regions. This enables the noise estimate to be updated quickly and the noise variation to be well tracked, yielding background noise being efficiently removed by the process of speech enhancement. Experimental results show that the proposed method can efficiently track the variation of background noise, enabling the performance of speech enhancement to be improved.

AB - The accuracy of noise estimation is important for the performance of a speech enhancement system. This study proposes using variable segment length for noise tracking and variable thresholds for the determination of speech-presence probability. Initially, the fundamental frequency is estimated to determine whether a frame is a vowel. In the case of a vowel frame, the segment length increases; meanwhile the threshold for speech-presence is decreased. So the noise magnitude is adequately underestimated. The speech distortion is accordingly reduced in enhanced speech. Conversely, the segment length is rapidly decreased during noise-dominant regions. This enables the noise estimate to be updated quickly and the noise variation to be well tracked, yielding background noise being efficiently removed by the process of speech enhancement. Experimental results show that the proposed method can efficiently track the variation of background noise, enabling the performance of speech enhancement to be improved.

UR - http://www.scopus.com/inward/record.url?scp=85031416105&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85031416105&partnerID=8YFLogxK

U2 - 10.1007/978-981-10-3187-8_32

DO - 10.1007/978-981-10-3187-8_32

M3 - Conference contribution

AN - SCOPUS:85031416105

SN - 9789811031861

T3 - Lecture Notes in Electrical Engineering

SP - 317

EP - 327

BT - Frontier Computing - Theory, Technologies and Applications, FC 2016

A2 - Yen, Neil Y.

A2 - Hung, Jason C.

PB - Springer Verlag

ER -

Lu CT, Chen Y-Y, Shen JH, Wang LL, Lei CL. Noise estimation for speech enhancement using minimum-spectral-average and vowel-presence detection approach. In Yen NY, Hung JC, editors, Frontier Computing - Theory, Technologies and Applications, FC 2016. Springer Verlag. 2018. p. 317-327. (Lecture Notes in Electrical Engineering). https://doi.org/10.1007/978-981-10-3187-8_32