Degradation type classifier for full band speech contaminated with echo, broadband noise, and reverberation

Leonardo O. Nunes; Luiz W.P. Biscainho; Bowon Lee; Amir Said; Ton Kalker; Ronald W. Schafer

doi:10.1109/TASL.2011.2144973

Degradation type classifier for full band speech contaminated with echo, broadband noise, and reverberation

Leonardo O. Nunes, Luiz W.P. Biscainho, Bowon Lee, Amir Said, Ton Kalker, Ronald W. Schafer

Research output: Contribution to journal › Article › peer-review

4 Scopus citations

Abstract

This paper addresses the problem of identifying impairment types that might be present in a speech signal. In particular, three acoustically induced degradation types that occur in teleconference systems are considered: acoustic echo, reverberation, and broadband noise, as well as combinations among them. The proposed system is double-ended (full reference) and is developed using a database of degraded full-band speech signals created according to a model for teleconference systems. A set of features obtained from both the degraded and non-degraded signals is proposed and shown to adequately capture information associated with each degradation type. A random forest classifier and a support vector machine are successfully employed, achieving a classification error below 2%. Such classifiers can be used to select an appropriate quality assessment tool for a given degraded signal.

Original language	English
Article number	5753923
Pages (from-to)	2516-2526
Number of pages	11
Journal	IEEE Transactions on Audio, Speech and Language Processing
Volume	19
Issue number	8
DOIs	https://doi.org/10.1109/TASL.2011.2144973
State	Published - 2011
Externally published	Yes

Bibliographical note

Funding Information:
Manuscript received December 21, 2010; revised March 28, 2011; accepted March 29, 2011. Date of publication April 21, 2011; date of current version September 23, 2011. The work of L. W. P. Biscainho was supported by CNPq and FAPERJ. The work of L. O. Nunes was supported in part by CNpQ. This R&D project is a cooperation between Hewlett-Packard Brasil Ltd. and COPPE/UFRJ, being supported with resources of Informatics Law (no 8.248, from 1991). The associate editor coordinating the review of this manuscript and approving it for publication was Prof. Lauri Savioja.

Keywords

Pattern classification
speech communication
speech quality assessment
teleconference systems

Access to Document

10.1109/TASL.2011.2144973

Cite this

@article{5317f97e7aee4f4793d67f4e1e1b2607,

title = "Degradation type classifier for full band speech contaminated with echo, broadband noise, and reverberation",

abstract = "This paper addresses the problem of identifying impairment types that might be present in a speech signal. In particular, three acoustically induced degradation types that occur in teleconference systems are considered: acoustic echo, reverberation, and broadband noise, as well as combinations among them. The proposed system is double-ended (full reference) and is developed using a database of degraded full-band speech signals created according to a model for teleconference systems. A set of features obtained from both the degraded and non-degraded signals is proposed and shown to adequately capture information associated with each degradation type. A random forest classifier and a support vector machine are successfully employed, achieving a classification error below 2%. Such classifiers can be used to select an appropriate quality assessment tool for a given degraded signal.",

keywords = "Pattern classification, speech communication, speech quality assessment, teleconference systems",

author = "Nunes, {Leonardo O.} and Biscainho, {Luiz W.P.} and Bowon Lee and Amir Said and Ton Kalker and Schafer, {Ronald W.}",

note = "Funding Information: Manuscript received December 21, 2010; revised March 28, 2011; accepted March 29, 2011. Date of publication April 21, 2011; date of current version September 23, 2011. The work of L. W. P. Biscainho was supported by CNPq and FAPERJ. The work of L. O. Nunes was supported in part by CNpQ. This R&D project is a cooperation between Hewlett-Packard Brasil Ltd. and COPPE/UFRJ, being supported with resources of Informatics Law (no 8.248, from 1991). The associate editor coordinating the review of this manuscript and approving it for publication was Prof. Lauri Savioja.",

year = "2011",

doi = "10.1109/TASL.2011.2144973",

language = "English",

volume = "19",

pages = "2516--2526",

journal = "IEEE Transactions on Audio, Speech and Language Processing",

issn = "1558-7916",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

number = "8",

}

TY - JOUR

T1 - Degradation type classifier for full band speech contaminated with echo, broadband noise, and reverberation

AU - Nunes, Leonardo O.

AU - Biscainho, Luiz W.P.

AU - Lee, Bowon

AU - Said, Amir

AU - Kalker, Ton

AU - Schafer, Ronald W.

N1 - Funding Information: Manuscript received December 21, 2010; revised March 28, 2011; accepted March 29, 2011. Date of publication April 21, 2011; date of current version September 23, 2011. The work of L. W. P. Biscainho was supported by CNPq and FAPERJ. The work of L. O. Nunes was supported in part by CNpQ. This R&D project is a cooperation between Hewlett-Packard Brasil Ltd. and COPPE/UFRJ, being supported with resources of Informatics Law (no 8.248, from 1991). The associate editor coordinating the review of this manuscript and approving it for publication was Prof. Lauri Savioja.

PY - 2011

Y1 - 2011

N2 - This paper addresses the problem of identifying impairment types that might be present in a speech signal. In particular, three acoustically induced degradation types that occur in teleconference systems are considered: acoustic echo, reverberation, and broadband noise, as well as combinations among them. The proposed system is double-ended (full reference) and is developed using a database of degraded full-band speech signals created according to a model for teleconference systems. A set of features obtained from both the degraded and non-degraded signals is proposed and shown to adequately capture information associated with each degradation type. A random forest classifier and a support vector machine are successfully employed, achieving a classification error below 2%. Such classifiers can be used to select an appropriate quality assessment tool for a given degraded signal.

AB - This paper addresses the problem of identifying impairment types that might be present in a speech signal. In particular, three acoustically induced degradation types that occur in teleconference systems are considered: acoustic echo, reverberation, and broadband noise, as well as combinations among them. The proposed system is double-ended (full reference) and is developed using a database of degraded full-band speech signals created according to a model for teleconference systems. A set of features obtained from both the degraded and non-degraded signals is proposed and shown to adequately capture information associated with each degradation type. A random forest classifier and a support vector machine are successfully employed, achieving a classification error below 2%. Such classifiers can be used to select an appropriate quality assessment tool for a given degraded signal.

KW - Pattern classification

KW - speech communication

KW - speech quality assessment

KW - teleconference systems

UR - http://www.scopus.com/inward/record.url?scp=80053194799&partnerID=8YFLogxK

U2 - 10.1109/TASL.2011.2144973

DO - 10.1109/TASL.2011.2144973

M3 - Article

AN - SCOPUS:80053194799

SN - 1558-7916

VL - 19

SP - 2516

EP - 2526

JO - IEEE Transactions on Audio, Speech and Language Processing

JF - IEEE Transactions on Audio, Speech and Language Processing

IS - 8

M1 - 5753923

ER -

Degradation type classifier for full band speech contaminated with echo, broadband noise, and reverberation

Abstract

Bibliographical note

Keywords

Access to Document

Other files and links

Fingerprint

Cite this