Experiments on deep learning for speech denoising

Ding Liu, Paris Smaragdis, Minje Kim

Research output: Contribution to journalConference article

Abstract

In this paper we present some experiments using a deep learning model for speech denoising. We propose a very lightweight procedure that can predict clean speech spectra when presented with noisy speech inputs, and we show how various parameter choices impact the quality of the denoised signal. Through our experiments we conclude that such a structure can perform better than some comparable single-channel approaches and that it is able to generalize well across various speakers, noise types and signal-to-noise ratios.

Original languageEnglish (US)
Pages (from-to)2685-2689
Number of pages5
JournalProceedings of the Annual Conference of the International Speech Communication Association, INTERSPEECH
StatePublished - Jan 1 2014
Event15th Annual Conference of the International Speech Communication Association: Celebrating the Diversity of Spoken Languages, INTERSPEECH 2014 - Singapore, Singapore
Duration: Sep 14 2014Sep 18 2014

Fingerprint

Denoising
Experiment
Experiments
Signal to noise ratio
Predict
Generalise
Speech
Learning
Deep learning
Model

Keywords

  • Deep learning
  • Neural networks
  • Source separation
  • Speech denoising

ASJC Scopus subject areas

  • Language and Linguistics
  • Human-Computer Interaction
  • Signal Processing
  • Software
  • Modeling and Simulation

Cite this

Experiments on deep learning for speech denoising. / Liu, Ding; Smaragdis, Paris; Kim, Minje.

In: Proceedings of the Annual Conference of the International Speech Communication Association, INTERSPEECH, 01.01.2014, p. 2685-2689.

Research output: Contribution to journalConference article

@article{1aedbaff5bad4bf5a2cbd50dd1542f00,
title = "Experiments on deep learning for speech denoising",
abstract = "In this paper we present some experiments using a deep learning model for speech denoising. We propose a very lightweight procedure that can predict clean speech spectra when presented with noisy speech inputs, and we show how various parameter choices impact the quality of the denoised signal. Through our experiments we conclude that such a structure can perform better than some comparable single-channel approaches and that it is able to generalize well across various speakers, noise types and signal-to-noise ratios.",
keywords = "Deep learning, Neural networks, Source separation, Speech denoising",
author = "Ding Liu and Paris Smaragdis and Minje Kim",
year = "2014",
month = "1",
day = "1",
language = "English (US)",
pages = "2685--2689",
journal = "Proceedings of the Annual Conference of the International Speech Communication Association, INTERSPEECH",
issn = "2308-457X",

}

TY - JOUR

T1 - Experiments on deep learning for speech denoising

AU - Liu, Ding

AU - Smaragdis, Paris

AU - Kim, Minje

PY - 2014/1/1

Y1 - 2014/1/1

N2 - In this paper we present some experiments using a deep learning model for speech denoising. We propose a very lightweight procedure that can predict clean speech spectra when presented with noisy speech inputs, and we show how various parameter choices impact the quality of the denoised signal. Through our experiments we conclude that such a structure can perform better than some comparable single-channel approaches and that it is able to generalize well across various speakers, noise types and signal-to-noise ratios.

AB - In this paper we present some experiments using a deep learning model for speech denoising. We propose a very lightweight procedure that can predict clean speech spectra when presented with noisy speech inputs, and we show how various parameter choices impact the quality of the denoised signal. Through our experiments we conclude that such a structure can perform better than some comparable single-channel approaches and that it is able to generalize well across various speakers, noise types and signal-to-noise ratios.

KW - Deep learning

KW - Neural networks

KW - Source separation

KW - Speech denoising

UR - http://www.scopus.com/inward/record.url?scp=84910049527&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84910049527&partnerID=8YFLogxK

M3 - Conference article

AN - SCOPUS:84910049527

SP - 2685

EP - 2689

JO - Proceedings of the Annual Conference of the International Speech Communication Association, INTERSPEECH

JF - Proceedings of the Annual Conference of the International Speech Communication Association, INTERSPEECH

SN - 2308-457X

ER -