Tampere University of Technology

TUTCRIS Research Portal

Separation of Moving Sound Sources Using Multichannel NMF and Acoustic Tracking

Research output: Contribution to journalArticleScientificpeer-review

Standard

Separation of Moving Sound Sources Using Multichannel NMF and Acoustic Tracking. / Nikunen, Joonas; Diment, Aleksandr; Virtanen, Tuomas.

In: IEEE/ACM Transactions on Audio Speech and Language Processing, Vol. 26, No. 2, 2018, p. 281-295.

Research output: Contribution to journalArticleScientificpeer-review

Harvard

Nikunen, J, Diment, A & Virtanen, T 2018, 'Separation of Moving Sound Sources Using Multichannel NMF and Acoustic Tracking', IEEE/ACM Transactions on Audio Speech and Language Processing, vol. 26, no. 2, pp. 281-295. https://doi.org/10.1109/TASLP.2017.2774925

APA

Nikunen, J., Diment, A., & Virtanen, T. (2018). Separation of Moving Sound Sources Using Multichannel NMF and Acoustic Tracking. IEEE/ACM Transactions on Audio Speech and Language Processing, 26(2), 281-295. https://doi.org/10.1109/TASLP.2017.2774925

Vancouver

Nikunen J, Diment A, Virtanen T. Separation of Moving Sound Sources Using Multichannel NMF and Acoustic Tracking. IEEE/ACM Transactions on Audio Speech and Language Processing. 2018;26(2):281-295. https://doi.org/10.1109/TASLP.2017.2774925

Author

Nikunen, Joonas ; Diment, Aleksandr ; Virtanen, Tuomas. / Separation of Moving Sound Sources Using Multichannel NMF and Acoustic Tracking. In: IEEE/ACM Transactions on Audio Speech and Language Processing. 2018 ; Vol. 26, No. 2. pp. 281-295.

Bibtex - Download

@article{45afd50c53ef40cda25a7c4dee4b380e,
title = "Separation of Moving Sound Sources Using Multichannel NMF and Acoustic Tracking",
abstract = "In this paper we propose a method for separation of moving sound sources. The method is based on first tracking the sources and then estimation of source spectrograms using multichannel non-negative matrix factorization (NMF) and extracting the sources from the mixture by single-channel Wiener filtering. We propose a novel multichannel NMF model with time-varying mixing of the sources denoted by spatial covariance matrices (SCM) and provide update equations for optimizing model parameters minimizing squared Frobenius norm. The SCMs of the model are obtained based on estimated directions of arrival of tracked sources at each time frame. The evaluation is based on established objective separation criteria and using real recordings of two and three simultaneous moving sound sources. The compared methods include conventional beamforming and ideal ratio mask separation. The proposed method is shown to exceed the separation quality of other evaluated blind approaches according to all measured quantities. Additionally, we evaluate the method's susceptibility towards tracking errors by comparing the separation quality achieved using annotated ground truth source trajectories.",
keywords = "acoustic source tracking, Acoustics, Array signal processing, Direction-of-arrival estimation, Estimation, Mathematical model, microphone arrays, Microphones, moving sound sources, Sound source separation, Spectrogram, time-varying mixing model",
author = "Joonas Nikunen and Aleksandr Diment and Tuomas Virtanen",
year = "2018",
doi = "10.1109/TASLP.2017.2774925",
language = "English",
volume = "26",
pages = "281--295",
journal = "Ieee-Acm transactions on audio speech and language processing",
issn = "2329-9290",
publisher = "IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC",
number = "2",

}

RIS (suitable for import to EndNote) - Download

TY - JOUR

T1 - Separation of Moving Sound Sources Using Multichannel NMF and Acoustic Tracking

AU - Nikunen, Joonas

AU - Diment, Aleksandr

AU - Virtanen, Tuomas

PY - 2018

Y1 - 2018

N2 - In this paper we propose a method for separation of moving sound sources. The method is based on first tracking the sources and then estimation of source spectrograms using multichannel non-negative matrix factorization (NMF) and extracting the sources from the mixture by single-channel Wiener filtering. We propose a novel multichannel NMF model with time-varying mixing of the sources denoted by spatial covariance matrices (SCM) and provide update equations for optimizing model parameters minimizing squared Frobenius norm. The SCMs of the model are obtained based on estimated directions of arrival of tracked sources at each time frame. The evaluation is based on established objective separation criteria and using real recordings of two and three simultaneous moving sound sources. The compared methods include conventional beamforming and ideal ratio mask separation. The proposed method is shown to exceed the separation quality of other evaluated blind approaches according to all measured quantities. Additionally, we evaluate the method's susceptibility towards tracking errors by comparing the separation quality achieved using annotated ground truth source trajectories.

AB - In this paper we propose a method for separation of moving sound sources. The method is based on first tracking the sources and then estimation of source spectrograms using multichannel non-negative matrix factorization (NMF) and extracting the sources from the mixture by single-channel Wiener filtering. We propose a novel multichannel NMF model with time-varying mixing of the sources denoted by spatial covariance matrices (SCM) and provide update equations for optimizing model parameters minimizing squared Frobenius norm. The SCMs of the model are obtained based on estimated directions of arrival of tracked sources at each time frame. The evaluation is based on established objective separation criteria and using real recordings of two and three simultaneous moving sound sources. The compared methods include conventional beamforming and ideal ratio mask separation. The proposed method is shown to exceed the separation quality of other evaluated blind approaches according to all measured quantities. Additionally, we evaluate the method's susceptibility towards tracking errors by comparing the separation quality achieved using annotated ground truth source trajectories.

KW - acoustic source tracking

KW - Acoustics

KW - Array signal processing

KW - Direction-of-arrival estimation

KW - Estimation

KW - Mathematical model

KW - microphone arrays

KW - Microphones

KW - moving sound sources

KW - Sound source separation

KW - Spectrogram

KW - time-varying mixing model

U2 - 10.1109/TASLP.2017.2774925

DO - 10.1109/TASLP.2017.2774925

M3 - Article

VL - 26

SP - 281

EP - 295

JO - Ieee-Acm transactions on audio speech and language processing

JF - Ieee-Acm transactions on audio speech and language processing

SN - 2329-9290

IS - 2

ER -