Gaussian modeling of mixtures of non-stationary signals in the time-frequency domain (HR-NMF)

Roland Badeau

doi:10.1109/ASPAA.2011.6082264

Gaussian modeling of mixtures of non-stationary signals in the time-frequency domain (HR-NMF)

Roland Badeau

CNRS LTCI

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

Abstract

Nonnegative Matrix Factorization (NMF) is a powerful tool for decomposing mixtures of non-stationary signals in the Time-Frequency (TF) domain. However, unlike the High Resolution (HR) methods dedicated to mixtures of exponentials, its spectral resolution is limited by that of the underlying TF representation. In this paper, we propose a unified probabilistic model called HR-NMF, that permits to overcome this limit by taking both phases and local correlations in each frequency band into account. This model is estimated with a recursive implementation of the EM algorithm, that is successfully applied to source separation and audio inpainting.

Original language	English
Title of host publication	2011 IEEE Workshop on Applications of Signal Processing to Audio and Acoustics, WASPAA 2011
Pages	253-256
Number of pages	4
DOIs	https://doi.org/10.1109/ASPAA.2011.6082264
Publication status	Published - 19 Dec 2011
Externally published	Yes
Event	2011 IEEE Workshop on Applications of Signal Processing to Audio and Acoustics, WASPAA 2011 - New Paltz, NY, United States Duration: 16 Oct 2011 → 19 Oct 2011

Publication series

Name	IEEE Workshop on Applications of Signal Processing to Audio and Acoustics

Conference

Conference	2011 IEEE Workshop on Applications of Signal Processing to Audio and Acoustics, WASPAA 2011
Country/Territory	United States
City	New Paltz, NY
Period	16/10/11 → 19/10/11

Keywords

Expectation-Maximization algorithm
High Resolution methods
Nonnegative Matrix Factorization
audio inpainting
source separation

Access to Document

10.1109/ASPAA.2011.6082264

Cite this

@inproceedings{621f2ddb081149c683f7ea5f9278650a,

title = "Gaussian modeling of mixtures of non-stationary signals in the time-frequency domain (HR-NMF)",

abstract = "Nonnegative Matrix Factorization (NMF) is a powerful tool for decomposing mixtures of non-stationary signals in the Time-Frequency (TF) domain. However, unlike the High Resolution (HR) methods dedicated to mixtures of exponentials, its spectral resolution is limited by that of the underlying TF representation. In this paper, we propose a unified probabilistic model called HR-NMF, that permits to overcome this limit by taking both phases and local correlations in each frequency band into account. This model is estimated with a recursive implementation of the EM algorithm, that is successfully applied to source separation and audio inpainting.",

keywords = "Expectation-Maximization algorithm, High Resolution methods, Nonnegative Matrix Factorization, audio inpainting, source separation",

author = "Roland Badeau",

year = "2011",

month = dec,

day = "19",

doi = "10.1109/ASPAA.2011.6082264",

language = "English",

isbn = "9781457706912",

series = "IEEE Workshop on Applications of Signal Processing to Audio and Acoustics",

pages = "253--256",

booktitle = "2011 IEEE Workshop on Applications of Signal Processing to Audio and Acoustics, WASPAA 2011",

note = "2011 IEEE Workshop on Applications of Signal Processing to Audio and Acoustics, WASPAA 2011 ; Conference date: 16-10-2011 Through 19-10-2011",

}

Badeau, R 2011, Gaussian modeling of mixtures of non-stationary signals in the time-frequency domain (HR-NMF). in 2011 IEEE Workshop on Applications of Signal Processing to Audio and Acoustics, WASPAA 2011., 6082264, IEEE Workshop on Applications of Signal Processing to Audio and Acoustics, pp. 253-256, 2011 IEEE Workshop on Applications of Signal Processing to Audio and Acoustics, WASPAA 2011, New Paltz, NY, United States, 16/10/11. https://doi.org/10.1109/ASPAA.2011.6082264

Gaussian modeling of mixtures of non-stationary signals in the time-frequency domain (HR-NMF). / Badeau, Roland.
2011 IEEE Workshop on Applications of Signal Processing to Audio and Acoustics, WASPAA 2011. 2011. p. 253-256 6082264 (IEEE Workshop on Applications of Signal Processing to Audio and Acoustics).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - Gaussian modeling of mixtures of non-stationary signals in the time-frequency domain (HR-NMF)

AU - Badeau, Roland

PY - 2011/12/19

Y1 - 2011/12/19

N2 - Nonnegative Matrix Factorization (NMF) is a powerful tool for decomposing mixtures of non-stationary signals in the Time-Frequency (TF) domain. However, unlike the High Resolution (HR) methods dedicated to mixtures of exponentials, its spectral resolution is limited by that of the underlying TF representation. In this paper, we propose a unified probabilistic model called HR-NMF, that permits to overcome this limit by taking both phases and local correlations in each frequency band into account. This model is estimated with a recursive implementation of the EM algorithm, that is successfully applied to source separation and audio inpainting.

AB - Nonnegative Matrix Factorization (NMF) is a powerful tool for decomposing mixtures of non-stationary signals in the Time-Frequency (TF) domain. However, unlike the High Resolution (HR) methods dedicated to mixtures of exponentials, its spectral resolution is limited by that of the underlying TF representation. In this paper, we propose a unified probabilistic model called HR-NMF, that permits to overcome this limit by taking both phases and local correlations in each frequency band into account. This model is estimated with a recursive implementation of the EM algorithm, that is successfully applied to source separation and audio inpainting.

KW - Expectation-Maximization algorithm

KW - High Resolution methods

KW - Nonnegative Matrix Factorization

KW - audio inpainting

KW - source separation

U2 - 10.1109/ASPAA.2011.6082264

DO - 10.1109/ASPAA.2011.6082264

M3 - Conference contribution

AN - SCOPUS:83455221408

SN - 9781457706912

T3 - IEEE Workshop on Applications of Signal Processing to Audio and Acoustics

SP - 253

EP - 256

BT - 2011 IEEE Workshop on Applications of Signal Processing to Audio and Acoustics, WASPAA 2011

T2 - 2011 IEEE Workshop on Applications of Signal Processing to Audio and Acoustics, WASPAA 2011

Y2 - 16 October 2011 through 19 October 2011

ER -

Gaussian modeling of mixtures of non-stationary signals in the time-frequency domain (HR-NMF)

Abstract

Publication series

Conference

Keywords

Access to Document

Fingerprint

Cite this