Enhancement of Non-Stationary Speech using Harmonic Chirp Filters

Publikation: Bidrag til bog/antologi/rapport/konference proceedingKonferenceartikel i proceedingForskningpeer review

4 Citationer (Scopus)
112 Downloads (Pure)

Resumé

In this paper, the issue of single channel speech enhancement of non-stationary voiced speech is addressed. The non-stationarity of speech is well known, but state of the art speech enhancement methods assume stationarity within frames of 20–30 ms. We derive optimal distortionless filters that take the non-stationarity nature of voiced speech into account via linear constraints. This is facilitated by imposing a harmonic chirp model on the speech signal. As an implicit part of the filter design, the noise statistics are also estimated based on the observed signal and parameters of the harmonic chirp model. Simulations on real speech show that the chirp based filters perform better than their harmonic counterparts. Further, it is seen that the gain of using the chirp model increases when the estimated chirp parameter is big corresponding to periods in the signal where the instantaneous fundamental frequency changes fast.
OriginalsprogEngelsk
TitelINTERSPEECH 2015 : 16th Annual Conference of the International Speech Communication Association Dresden, Germany September 6-10, 2015
ForlagInternational Speech Communications Association
Publikationsdato2015
Sider1755-1759
StatusUdgivet - 2015
BegivenhedINTERSPEECH 2015 16th Annual Conference of the International Speech Communication Association - Dresden, Tyskland
Varighed: 6 sep. 201510 sep. 2015

Konference

KonferenceINTERSPEECH 2015 16th Annual Conference of the International Speech Communication Association
LandTyskland
ByDresden
Periode06/09/201510/09/2015
NavnINTERSPEECH
ISSN1990-9770

Fingerprint

Speech enhancement
Statistics

Citer dette

Nørholm, S. M., Jensen, J. R., & Christensen, M. G. (2015). Enhancement of Non-Stationary Speech using Harmonic Chirp Filters. I INTERSPEECH 2015: 16th Annual Conference of the International Speech Communication Association Dresden, Germany September 6-10, 2015 (s. 1755-1759). International Speech Communications Association. INTERSPEECH
Nørholm, Sidsel Marie ; Jensen, Jesper Rindom ; Christensen, Mads Græsbøll. / Enhancement of Non-Stationary Speech using Harmonic Chirp Filters. INTERSPEECH 2015: 16th Annual Conference of the International Speech Communication Association Dresden, Germany September 6-10, 2015. International Speech Communications Association, 2015. s. 1755-1759 (INTERSPEECH ).
@inproceedings{65af72cc3bd34138bf38749950a568ee,
title = "Enhancement of Non-Stationary Speech using Harmonic Chirp Filters",
abstract = "In this paper, the issue of single channel speech enhancement of non-stationary voiced speech is addressed. The non-stationarity of speech is well known, but state of the art speech enhancement methods assume stationarity within frames of 20–30 ms. We derive optimal distortionless filters that take the non-stationarity nature of voiced speech into account via linear constraints. This is facilitated by imposing a harmonic chirp model on the speech signal. As an implicit part of the filter design, the noise statistics are also estimated based on the observed signal and parameters of the harmonic chirp model. Simulations on real speech show that the chirp based filters perform better than their harmonic counterparts. Further, it is seen that the gain of using the chirp model increases when the estimated chirp parameter is big corresponding to periods in the signal where the instantaneous fundamental frequency changes fast.",
keywords = "speech enhancement, single-channel, non-stationary signals, harmonic chirp model",
author = "N{\o}rholm, {Sidsel Marie} and Jensen, {Jesper Rindom} and Christensen, {Mads Gr{\ae}sb{\o}ll}",
year = "2015",
language = "English",
pages = "1755--1759",
booktitle = "INTERSPEECH 2015",
publisher = "International Speech Communications Association",

}

Nørholm, SM, Jensen, JR & Christensen, MG 2015, Enhancement of Non-Stationary Speech using Harmonic Chirp Filters. i INTERSPEECH 2015: 16th Annual Conference of the International Speech Communication Association Dresden, Germany September 6-10, 2015. International Speech Communications Association, INTERSPEECH , s. 1755-1759, Dresden, Tyskland, 06/09/2015.

Enhancement of Non-Stationary Speech using Harmonic Chirp Filters. / Nørholm, Sidsel Marie; Jensen, Jesper Rindom; Christensen, Mads Græsbøll.

INTERSPEECH 2015: 16th Annual Conference of the International Speech Communication Association Dresden, Germany September 6-10, 2015. International Speech Communications Association, 2015. s. 1755-1759 (INTERSPEECH ).

Publikation: Bidrag til bog/antologi/rapport/konference proceedingKonferenceartikel i proceedingForskningpeer review

TY - GEN

T1 - Enhancement of Non-Stationary Speech using Harmonic Chirp Filters

AU - Nørholm, Sidsel Marie

AU - Jensen, Jesper Rindom

AU - Christensen, Mads Græsbøll

PY - 2015

Y1 - 2015

N2 - In this paper, the issue of single channel speech enhancement of non-stationary voiced speech is addressed. The non-stationarity of speech is well known, but state of the art speech enhancement methods assume stationarity within frames of 20–30 ms. We derive optimal distortionless filters that take the non-stationarity nature of voiced speech into account via linear constraints. This is facilitated by imposing a harmonic chirp model on the speech signal. As an implicit part of the filter design, the noise statistics are also estimated based on the observed signal and parameters of the harmonic chirp model. Simulations on real speech show that the chirp based filters perform better than their harmonic counterparts. Further, it is seen that the gain of using the chirp model increases when the estimated chirp parameter is big corresponding to periods in the signal where the instantaneous fundamental frequency changes fast.

AB - In this paper, the issue of single channel speech enhancement of non-stationary voiced speech is addressed. The non-stationarity of speech is well known, but state of the art speech enhancement methods assume stationarity within frames of 20–30 ms. We derive optimal distortionless filters that take the non-stationarity nature of voiced speech into account via linear constraints. This is facilitated by imposing a harmonic chirp model on the speech signal. As an implicit part of the filter design, the noise statistics are also estimated based on the observed signal and parameters of the harmonic chirp model. Simulations on real speech show that the chirp based filters perform better than their harmonic counterparts. Further, it is seen that the gain of using the chirp model increases when the estimated chirp parameter is big corresponding to periods in the signal where the instantaneous fundamental frequency changes fast.

KW - speech enhancement, single-channel, non-stationary signals, harmonic chirp model

UR - http://interspeech2015.org//wp-content/uploads/direct/INTERSPEECH_2015_AbstractBook.pdf

UR - http://www.isca-speech.org/archive/interspeech_2015/

M3 - Article in proceeding

SP - 1755

EP - 1759

BT - INTERSPEECH 2015

PB - International Speech Communications Association

ER -

Nørholm SM, Jensen JR, Christensen MG. Enhancement of Non-Stationary Speech using Harmonic Chirp Filters. I INTERSPEECH 2015: 16th Annual Conference of the International Speech Communication Association Dresden, Germany September 6-10, 2015. International Speech Communications Association. 2015. s. 1755-1759. (INTERSPEECH ).