Online Parametric NMF for Speech Enhancement

Research output: Contribution to book/anthology/report/conference proceedingArticle in proceedingResearchpeer-review

1 Citation (Scopus)
59 Downloads (Pure)

Abstract

In this paper, we propose a speech enhancement method based on non-negative matrix factorization (NMF) techniques. NMF techniques allow us to approximate the power spectral density (PSD) of the noisy signal as a weighted linear combination of trained speech and noise basis vectors arranged as the columns of a matrix. In this work, we propose to use basis vectors that are parameterised by autoregressive (AR) coefficients. Parametric representation of the spectral basis is beneficial as it can encompass the signal characteristics like, e.g. the speech production model. It is observed that the parametric representation of basis vectors is beneficial while performing online speech enhancement in low delay scenarios.
Original languageEnglish
Title of host publication2018 26th European Signal Processing Conference (EUSIPCO)
Number of pages5
PublisherIEEE
Publication date2018
Article number8553039
ISBN (Print)978-90-827970-0-8
ISBN (Electronic)978-9-0827-9701-5
DOIs
Publication statusPublished - 2018
Event26th European Signal Processing Conference (EUSIPCO 2018) - Rome, Italy
Duration: 3 Sep 20187 Sep 2018
Conference number: 26
http://www.eusipco2018.org

Conference

Conference26th European Signal Processing Conference (EUSIPCO 2018)
Number26
CountryItaly
CityRome
Period03/09/201807/09/2018
Internet address
SeriesProceedings of the European Signal Processing Conference
ISSN2076-1465

Fingerprint

Speech enhancement
Factorization
Power spectral density

Keywords

  • autoregressive modelling, speech enhancement, NMF

Cite this

Kavalekalam, M. S., Nielsen, J. K., Shi, L., Christensen, M. G., & Boldt, J. (2018). Online Parametric NMF for Speech Enhancement. In 2018 26th European Signal Processing Conference (EUSIPCO) [8553039] IEEE. Proceedings of the European Signal Processing Conference https://doi.org/10.23919/EUSIPCO.2018.8553039
Kavalekalam, Mathew Shaji ; Nielsen, Jesper Kjær ; Shi, Liming ; Christensen, Mads Græsbøll ; Boldt, Jesper. / Online Parametric NMF for Speech Enhancement. 2018 26th European Signal Processing Conference (EUSIPCO). IEEE, 2018. (Proceedings of the European Signal Processing Conference).
@inproceedings{6cda013d9d5e4693aaca9a9a90d92b54,
title = "Online Parametric NMF for Speech Enhancement",
abstract = "In this paper, we propose a speech enhancement method based on non-negative matrix factorization (NMF) techniques. NMF techniques allow us to approximate the power spectral density (PSD) of the noisy signal as a weighted linear combination of trained speech and noise basis vectors arranged as the columns of a matrix. In this work, we propose to use basis vectors that are parameterised by autoregressive (AR) coefficients. Parametric representation of the spectral basis is beneficial as it can encompass the signal characteristics like, e.g. the speech production model. It is observed that the parametric representation of basis vectors is beneficial while performing online speech enhancement in low delay scenarios.",
keywords = "autoregressive modelling, speech enhancement, NMF",
author = "Kavalekalam, {Mathew Shaji} and Nielsen, {Jesper Kj{\ae}r} and Liming Shi and Christensen, {Mads Gr{\ae}sb{\o}ll} and Jesper Boldt",
year = "2018",
doi = "10.23919/EUSIPCO.2018.8553039",
language = "English",
isbn = "978-90-827970-0-8",
booktitle = "2018 26th European Signal Processing Conference (EUSIPCO)",
publisher = "IEEE",
address = "United States",

}

Kavalekalam, MS, Nielsen, JK, Shi, L, Christensen, MG & Boldt, J 2018, Online Parametric NMF for Speech Enhancement. in 2018 26th European Signal Processing Conference (EUSIPCO)., 8553039, IEEE, Proceedings of the European Signal Processing Conference, 26th European Signal Processing Conference (EUSIPCO 2018), Rome, Italy, 03/09/2018. https://doi.org/10.23919/EUSIPCO.2018.8553039

Online Parametric NMF for Speech Enhancement. / Kavalekalam, Mathew Shaji; Nielsen, Jesper Kjær; Shi, Liming; Christensen, Mads Græsbøll; Boldt, Jesper.

2018 26th European Signal Processing Conference (EUSIPCO). IEEE, 2018. 8553039 (Proceedings of the European Signal Processing Conference).

Research output: Contribution to book/anthology/report/conference proceedingArticle in proceedingResearchpeer-review

TY - GEN

T1 - Online Parametric NMF for Speech Enhancement

AU - Kavalekalam, Mathew Shaji

AU - Nielsen, Jesper Kjær

AU - Shi, Liming

AU - Christensen, Mads Græsbøll

AU - Boldt, Jesper

PY - 2018

Y1 - 2018

N2 - In this paper, we propose a speech enhancement method based on non-negative matrix factorization (NMF) techniques. NMF techniques allow us to approximate the power spectral density (PSD) of the noisy signal as a weighted linear combination of trained speech and noise basis vectors arranged as the columns of a matrix. In this work, we propose to use basis vectors that are parameterised by autoregressive (AR) coefficients. Parametric representation of the spectral basis is beneficial as it can encompass the signal characteristics like, e.g. the speech production model. It is observed that the parametric representation of basis vectors is beneficial while performing online speech enhancement in low delay scenarios.

AB - In this paper, we propose a speech enhancement method based on non-negative matrix factorization (NMF) techniques. NMF techniques allow us to approximate the power spectral density (PSD) of the noisy signal as a weighted linear combination of trained speech and noise basis vectors arranged as the columns of a matrix. In this work, we propose to use basis vectors that are parameterised by autoregressive (AR) coefficients. Parametric representation of the spectral basis is beneficial as it can encompass the signal characteristics like, e.g. the speech production model. It is observed that the parametric representation of basis vectors is beneficial while performing online speech enhancement in low delay scenarios.

KW - autoregressive modelling, speech enhancement, NMF

U2 - 10.23919/EUSIPCO.2018.8553039

DO - 10.23919/EUSIPCO.2018.8553039

M3 - Article in proceeding

SN - 978-90-827970-0-8

BT - 2018 26th European Signal Processing Conference (EUSIPCO)

PB - IEEE

ER -

Kavalekalam MS, Nielsen JK, Shi L, Christensen MG, Boldt J. Online Parametric NMF for Speech Enhancement. In 2018 26th European Signal Processing Conference (EUSIPCO). IEEE. 2018. 8553039. (Proceedings of the European Signal Processing Conference). https://doi.org/10.23919/EUSIPCO.2018.8553039