TUTCRIS - Tampereen teknillinen yliopisto

TUTCRIS

Low-Latency Sound-Source-Separation using Non-Negative Matrix Factorisation with Coupled Analysis and Synthesis Dictionaries

Tutkimustuotosvertaisarvioitu

Standard

Low-Latency Sound-Source-Separation using Non-Negative Matrix Factorisation with Coupled Analysis and Synthesis Dictionaries. / Barker, Thomas; Virtanen, Tuomas; Pontoppidan, Niels Henrik.

2015 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP). IEEE, 2015. s. 241-245.

Tutkimustuotosvertaisarvioitu

Harvard

Barker, T, Virtanen, T & Pontoppidan, NH 2015, Low-Latency Sound-Source-Separation using Non-Negative Matrix Factorisation with Coupled Analysis and Synthesis Dictionaries. julkaisussa 2015 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP). IEEE, Sivut 241-245, IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH AND SIGNAL PROCESSING, 1/01/00. https://doi.org/10.1109/ICASSP.2015.7177968

APA

Barker, T., Virtanen, T., & Pontoppidan, N. H. (2015). Low-Latency Sound-Source-Separation using Non-Negative Matrix Factorisation with Coupled Analysis and Synthesis Dictionaries. teoksessa 2015 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) (Sivut 241-245). IEEE. https://doi.org/10.1109/ICASSP.2015.7177968

Vancouver

Barker T, Virtanen T, Pontoppidan NH. Low-Latency Sound-Source-Separation using Non-Negative Matrix Factorisation with Coupled Analysis and Synthesis Dictionaries. julkaisussa 2015 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP). IEEE. 2015. s. 241-245 https://doi.org/10.1109/ICASSP.2015.7177968

Author

Barker, Thomas ; Virtanen, Tuomas ; Pontoppidan, Niels Henrik. / Low-Latency Sound-Source-Separation using Non-Negative Matrix Factorisation with Coupled Analysis and Synthesis Dictionaries. 2015 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP). IEEE, 2015. Sivut 241-245

Bibtex - Lataa

@inproceedings{09d719dcbc4542b4bf3f44f1f5df9461,
title = "Low-Latency Sound-Source-Separation using Non-Negative Matrix Factorisation with Coupled Analysis and Synthesis Dictionaries",
abstract = "For real-time or close to real-time applications, sound source separation can be performed on-line, where new frames of incoming data for a mixture signal are processed as they arrive, at very low delay. We propose an approach which generates the separation filters for short synthesis frames to achieve low latency source separation, based on a compositional model mixture of the audio to be separated. Filter parameters are derived from a longer temporal context than the current processing frame through use of a longer analysis frame. A pair of dictionaries are used, one for analysis and one for reconstruction. With this approach we are able to increase separation performance at low latencies whilst retaining the low-latency provided by the use of short synthesis frames. The proposed data handling scheme and parameters can be adjusted to achieve real-time performance, given sufficient computational power. Low-latency output allows a human listener to use the results of such a separation scheme directly, without a perceptible delay. With the proposed method, separated source-to-distortion ratios (SDRs) can be improved by over 1 dB for latencies below 20 ms, without any affect on latency.",
keywords = "low-latency, NMF, Non-negative matrix factorisation, real-time, source separation",
author = "Thomas Barker and Tuomas Virtanen and Pontoppidan, {Niels Henrik}",
year = "2015",
month = "8",
day = "4",
doi = "10.1109/ICASSP.2015.7177968",
language = "English",
isbn = "9781467369978",
pages = "241--245",
booktitle = "2015 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)",
publisher = "IEEE",

}

RIS (suitable for import to EndNote) - Lataa

TY - GEN

T1 - Low-Latency Sound-Source-Separation using Non-Negative Matrix Factorisation with Coupled Analysis and Synthesis Dictionaries

AU - Barker, Thomas

AU - Virtanen, Tuomas

AU - Pontoppidan, Niels Henrik

PY - 2015/8/4

Y1 - 2015/8/4

N2 - For real-time or close to real-time applications, sound source separation can be performed on-line, where new frames of incoming data for a mixture signal are processed as they arrive, at very low delay. We propose an approach which generates the separation filters for short synthesis frames to achieve low latency source separation, based on a compositional model mixture of the audio to be separated. Filter parameters are derived from a longer temporal context than the current processing frame through use of a longer analysis frame. A pair of dictionaries are used, one for analysis and one for reconstruction. With this approach we are able to increase separation performance at low latencies whilst retaining the low-latency provided by the use of short synthesis frames. The proposed data handling scheme and parameters can be adjusted to achieve real-time performance, given sufficient computational power. Low-latency output allows a human listener to use the results of such a separation scheme directly, without a perceptible delay. With the proposed method, separated source-to-distortion ratios (SDRs) can be improved by over 1 dB for latencies below 20 ms, without any affect on latency.

AB - For real-time or close to real-time applications, sound source separation can be performed on-line, where new frames of incoming data for a mixture signal are processed as they arrive, at very low delay. We propose an approach which generates the separation filters for short synthesis frames to achieve low latency source separation, based on a compositional model mixture of the audio to be separated. Filter parameters are derived from a longer temporal context than the current processing frame through use of a longer analysis frame. A pair of dictionaries are used, one for analysis and one for reconstruction. With this approach we are able to increase separation performance at low latencies whilst retaining the low-latency provided by the use of short synthesis frames. The proposed data handling scheme and parameters can be adjusted to achieve real-time performance, given sufficient computational power. Low-latency output allows a human listener to use the results of such a separation scheme directly, without a perceptible delay. With the proposed method, separated source-to-distortion ratios (SDRs) can be improved by over 1 dB for latencies below 20 ms, without any affect on latency.

KW - low-latency

KW - NMF

KW - Non-negative matrix factorisation

KW - real-time

KW - source separation

UR - http://www.cs.tut.fi/~barkert/Publications_files/ICASSP2015_TomBarker_Preprint.pdf

U2 - 10.1109/ICASSP.2015.7177968

DO - 10.1109/ICASSP.2015.7177968

M3 - Conference contribution

SN - 9781467369978

SP - 241

EP - 245

BT - 2015 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)

PB - IEEE

ER -