Online speech dereverberation algorithm based on adaptive multichannel linear prediction

Jae Mo Yang, Hong-Goo Kang

Research output: Contribution to journalArticle

6 Citations (Scopus)

Abstract

This paper proposes a real-time acoustic channel equalization method that uses an adaptive multichannel linear prediction technique. In general, multichannel equalization algorithms can eliminate reverberation if they meet the following specific conditions including: the co-primeness between channels and sufficient filter length. It also requires the characteristic of correct channel information, however, it is difficult to estimate accurate acoustic channels in a practical system. The proposed method utilizes a theoretically perfect channel equalization algorithm and considers problems that may arise in the actual system. Linear-predictive multi-input equalization (LIME) is also an appropriate attempt at blind dereverberation by assuring the theoretical basis. However, a huge computational cost is incurred by calculating the large dimensions of a covariance matrix and its inversion. The proposed equalizer is developed as a multichannel linear prediction (MLP) oriented structure with a new formula that is optimized to time-varying acoustical room environments. Moreover, experimental results show that the proposed method works well even if the channel characteristics of each microphone are similar. The results of experiments using various room impulse response (RIR) models, including both the synthesized and real room environments, show that the proposed method is superior to conventional methods.

Original languageEnglish
Pages (from-to)608-619
Number of pages12
JournalIEEE Transactions on Audio, Speech and Language Processing
Volume22
Issue number3
DOIs
Publication statusPublished - 2014 Mar 1

Fingerprint

linear prediction
Acoustics
Reverberation
Equalizers
Microphones
Covariance matrix
Impulse response
rooms
Costs
acoustics
Experiments
reverberation
microphones
impulses
inversions
costs
filters
estimates

All Science Journal Classification (ASJC) codes

  • Acoustics and Ultrasonics
  • Electrical and Electronic Engineering

Cite this

@article{3f63b4be05e441a69f8d7749a2bb915d,
title = "Online speech dereverberation algorithm based on adaptive multichannel linear prediction",
abstract = "This paper proposes a real-time acoustic channel equalization method that uses an adaptive multichannel linear prediction technique. In general, multichannel equalization algorithms can eliminate reverberation if they meet the following specific conditions including: the co-primeness between channels and sufficient filter length. It also requires the characteristic of correct channel information, however, it is difficult to estimate accurate acoustic channels in a practical system. The proposed method utilizes a theoretically perfect channel equalization algorithm and considers problems that may arise in the actual system. Linear-predictive multi-input equalization (LIME) is also an appropriate attempt at blind dereverberation by assuring the theoretical basis. However, a huge computational cost is incurred by calculating the large dimensions of a covariance matrix and its inversion. The proposed equalizer is developed as a multichannel linear prediction (MLP) oriented structure with a new formula that is optimized to time-varying acoustical room environments. Moreover, experimental results show that the proposed method works well even if the channel characteristics of each microphone are similar. The results of experiments using various room impulse response (RIR) models, including both the synthesized and real room environments, show that the proposed method is superior to conventional methods.",
author = "Yang, {Jae Mo} and Hong-Goo Kang",
year = "2014",
month = "3",
day = "1",
doi = "10.1109/TASLP.2013.2294578",
language = "English",
volume = "22",
pages = "608--619",
journal = "IEEE Transactions on Speech and Audio Processing",
issn = "1558-7916",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
number = "3",

}

Online speech dereverberation algorithm based on adaptive multichannel linear prediction. / Yang, Jae Mo; Kang, Hong-Goo.

In: IEEE Transactions on Audio, Speech and Language Processing, Vol. 22, No. 3, 01.03.2014, p. 608-619.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Online speech dereverberation algorithm based on adaptive multichannel linear prediction

AU - Yang, Jae Mo

AU - Kang, Hong-Goo

PY - 2014/3/1

Y1 - 2014/3/1

N2 - This paper proposes a real-time acoustic channel equalization method that uses an adaptive multichannel linear prediction technique. In general, multichannel equalization algorithms can eliminate reverberation if they meet the following specific conditions including: the co-primeness between channels and sufficient filter length. It also requires the characteristic of correct channel information, however, it is difficult to estimate accurate acoustic channels in a practical system. The proposed method utilizes a theoretically perfect channel equalization algorithm and considers problems that may arise in the actual system. Linear-predictive multi-input equalization (LIME) is also an appropriate attempt at blind dereverberation by assuring the theoretical basis. However, a huge computational cost is incurred by calculating the large dimensions of a covariance matrix and its inversion. The proposed equalizer is developed as a multichannel linear prediction (MLP) oriented structure with a new formula that is optimized to time-varying acoustical room environments. Moreover, experimental results show that the proposed method works well even if the channel characteristics of each microphone are similar. The results of experiments using various room impulse response (RIR) models, including both the synthesized and real room environments, show that the proposed method is superior to conventional methods.

AB - This paper proposes a real-time acoustic channel equalization method that uses an adaptive multichannel linear prediction technique. In general, multichannel equalization algorithms can eliminate reverberation if they meet the following specific conditions including: the co-primeness between channels and sufficient filter length. It also requires the characteristic of correct channel information, however, it is difficult to estimate accurate acoustic channels in a practical system. The proposed method utilizes a theoretically perfect channel equalization algorithm and considers problems that may arise in the actual system. Linear-predictive multi-input equalization (LIME) is also an appropriate attempt at blind dereverberation by assuring the theoretical basis. However, a huge computational cost is incurred by calculating the large dimensions of a covariance matrix and its inversion. The proposed equalizer is developed as a multichannel linear prediction (MLP) oriented structure with a new formula that is optimized to time-varying acoustical room environments. Moreover, experimental results show that the proposed method works well even if the channel characteristics of each microphone are similar. The results of experiments using various room impulse response (RIR) models, including both the synthesized and real room environments, show that the proposed method is superior to conventional methods.

UR - http://www.scopus.com/inward/record.url?scp=84898076651&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84898076651&partnerID=8YFLogxK

U2 - 10.1109/TASLP.2013.2294578

DO - 10.1109/TASLP.2013.2294578

M3 - Article

AN - SCOPUS:84898076651

VL - 22

SP - 608

EP - 619

JO - IEEE Transactions on Speech and Audio Processing

JF - IEEE Transactions on Speech and Audio Processing

SN - 1558-7916

IS - 3

ER -