On the Influence of Inharmonicities in Model-Based Speech Enhancement

Research output: Contribution to journalConference article in JournalResearchpeer-review

2 Citations (Scopus)
196 Downloads (Pure)

Abstract

In relation to speech enhancement, we study the influence of
modifying the harmonic signal model for voiced speech to
include small perturbations in the frequencies of the harmonics.
A perturbed signal model is incorporated in the nonlinear
least squares method, the Capon filter and the amplitude and
phase estimation filter. Results show that it is possible to increase
the performance, in terms of the signal reduction factor
and the output signal-to-noise ratio, at the cost of increased
complexity in the estimation of the model parameters. It is
found that the perturbed signal model performs better than the
harmonic signal model at input signal-to-noise ratios above
approximately -10 dB, and that they are equally good below.
Original languageEnglish
JournalProceedings of the European Signal Processing Conference
ISSN2076-1465
Publication statusPublished - 2013
EventEuropean Signal Processing Conference EUSIPCO 2013 - Marrakech, Moroco
Duration: 9 Sep 201313 Sep 2013

Conference

ConferenceEuropean Signal Processing Conference EUSIPCO 2013
CityMarrakech, Moroco
Period09/09/201313/09/2013

Fingerprint

Speech enhancement
Signal to noise ratio

Keywords

  • Single-channel speech enhancement, perturbed signal models, inharmonicity, parameter estimation

Cite this

@inproceedings{680d341f33d24d55838e121716048caa,
title = "On the Influence of Inharmonicities in Model-Based Speech Enhancement",
abstract = "In relation to speech enhancement, we study the influence ofmodifying the harmonic signal model for voiced speech toinclude small perturbations in the frequencies of the harmonics.A perturbed signal model is incorporated in the nonlinearleast squares method, the Capon filter and the amplitude andphase estimation filter. Results show that it is possible to increasethe performance, in terms of the signal reduction factorand the output signal-to-noise ratio, at the cost of increasedcomplexity in the estimation of the model parameters. It isfound that the perturbed signal model performs better than theharmonic signal model at input signal-to-noise ratios aboveapproximately -10 dB, and that they are equally good below.",
keywords = "Single-channel speech enhancement, perturbed signal models, inharmonicity, parameter estimation, Single-channel speech enhancement, perturbed signal models, inharmonicity, parameter estimation",
author = "N{\o}rholm, {Sidsel Marie} and Jensen, {Jesper Rindom} and Christensen, {Mads Gr{\ae}sb{\o}ll}",
year = "2013",
language = "English",
journal = "Proceedings of the European Signal Processing Conference",
issn = "2076-1465",
publisher = "European Association for Signal Processing (EURASIP)",

}

TY - GEN

T1 - On the Influence of Inharmonicities in Model-Based Speech Enhancement

AU - Nørholm, Sidsel Marie

AU - Jensen, Jesper Rindom

AU - Christensen, Mads Græsbøll

PY - 2013

Y1 - 2013

N2 - In relation to speech enhancement, we study the influence ofmodifying the harmonic signal model for voiced speech toinclude small perturbations in the frequencies of the harmonics.A perturbed signal model is incorporated in the nonlinearleast squares method, the Capon filter and the amplitude andphase estimation filter. Results show that it is possible to increasethe performance, in terms of the signal reduction factorand the output signal-to-noise ratio, at the cost of increasedcomplexity in the estimation of the model parameters. It isfound that the perturbed signal model performs better than theharmonic signal model at input signal-to-noise ratios aboveapproximately -10 dB, and that they are equally good below.

AB - In relation to speech enhancement, we study the influence ofmodifying the harmonic signal model for voiced speech toinclude small perturbations in the frequencies of the harmonics.A perturbed signal model is incorporated in the nonlinearleast squares method, the Capon filter and the amplitude andphase estimation filter. Results show that it is possible to increasethe performance, in terms of the signal reduction factorand the output signal-to-noise ratio, at the cost of increasedcomplexity in the estimation of the model parameters. It isfound that the perturbed signal model performs better than theharmonic signal model at input signal-to-noise ratios aboveapproximately -10 dB, and that they are equally good below.

KW - Single-channel speech enhancement, perturbed signal models, inharmonicity, parameter estimation

KW - Single-channel speech enhancement, perturbed signal models, inharmonicity, parameter estimation

M3 - Conference article in Journal

JO - Proceedings of the European Signal Processing Conference

JF - Proceedings of the European Signal Processing Conference

SN - 2076-1465

ER -