Modified LMS algorithms for speech processing with an adaptive noise canceller

A desired signal corrupted by additive noise can often be recovered by an adaptive noise canceller using the least mean squares (LMS) algorithm. A major disadvantage of the LMS algorithm is its excess mean-squared error, or misadjustment, which increases linearly with the desired signal power, This...

Full description

Saved in:
Bibliographic Details
Published inIEEE transactions on speech and audio processing Vol. 6; no. 4; pp. 338 - 351
Main Author Greenberg, J.E.
Format Journal Article
LanguageEnglish
Published New York, NY IEEE 01.07.1998
Institute of Electrical and Electronics Engineers
Subjects
Online AccessGet full text
ISSN1063-6676
DOI10.1109/89.701363

Cover

Abstract A desired signal corrupted by additive noise can often be recovered by an adaptive noise canceller using the least mean squares (LMS) algorithm. A major disadvantage of the LMS algorithm is its excess mean-squared error, or misadjustment, which increases linearly with the desired signal power, This leads to degrading performance when the desired signal exhibits large power fluctuations and is a serious problem in many speech processing applications. This work considers two modified LMS algorithms, the weighted sum and sum methods, designed to solve this problem by reducing the size of the steps in the weight update equation when the desired signal is strong. The weighted sum method is derived from an optimal method (also developed in this work), which is not generally applicable because it requires quantities unavailable in a practical system. The previously proposed, but ad hoc, sum method is analyzed and compared to the weighted sum method. Analysis of the two modified LMS algorithms indicates that either one provides substantial improvements in the presence of strong desired signals and similar performance in the presence of weak desired signals, relative to the unmodified LMS algorithm. Computer simulations with both uncorrelated Gaussian noise and speech signals confirm the results of the analysis and demonstrate the effectiveness of the modified algorithms. The modified LMS algorithms are particularly suited for signals (such as speech) that exhibit large fluctuations in short-time power levels.
AbstractList A desired signal corrupted by additive noise can often be recovered by an adaptive noise canceller using the least mean squares (LMS) algorithm. A major disadvantage of the LMS algorithm is its excess mean-squared error, or misadjustment, which increases linearly with the desired signal power, This leads to degrading performance when the desired signal exhibits large power fluctuations and is a serious problem in many speech processing applications. This work considers two modified LMS algorithms, the weighted sum and sum methods, designed to solve this problem by reducing the size of the steps in the weight update equation when the desired signal is strong. The weighted sum method is derived from an optimal method (also developed in this work), which is not generally applicable because it requires quantities unavailable in a practical system. The previously proposed, but ad hoc, sum method is analyzed and compared to the weighted sum method. Analysis of the two modified LMS algorithms indicates that either one provides substantial improvements in the presence of strong desired signals and similar performance in the presence of weak desired signals, relative to the unmodified LMS algorithm. Computer simulations with both uncorrelated Gaussian noise and speech signals confirm the results of the analysis and demonstrate the effectiveness of the modified algorithms. The modified LMS algorithms are particularly suited for signals (such as speech) that exhibit large fluctuations in short-time power levels
A desired signal corrupted by additive noise can often be recovered by an adaptive noise canceller using the least mean squares (LMS) algorithm. A major disadvantage of the LMS algorithm is its excess mean-squared error, or misadjustment, which increases linearly with the desired signal power, This leads to degrading performance when the desired signal exhibits large power fluctuations and is a serious problem in many speech processing applications. This work considers two modified LMS algorithms, the weighted sum and sum methods, designed to solve this problem by reducing the size of the steps in the weight update equation when the desired signal is strong. The weighted sum method is derived from an optimal method (also developed in this work), which is not generally applicable because it requires quantities unavailable in a practical system. The previously proposed, but ad hoc, sum method is analyzed and compared to the weighted sum method. Analysis of the two modified LMS algorithms indicates that either one provides substantial improvements in the presence of strong desired signals and similar performance in the presence of weak desired signals, relative to the unmodified LMS algorithm. Computer simulations with both uncorrelated Gaussian noise and speech signals confirm the results of the analysis and demonstrate the effectiveness of the modified algorithms. The modified LMS algorithms are particularly suited for signals (such as speech) that exhibit large fluctuations in short-time power levels.
Author Greenberg, J.E.
Author_xml – sequence: 1
  givenname: J.E.
  surname: Greenberg
  fullname: Greenberg, J.E.
  organization: MIT, Cambridge, MA, USA
BackLink http://pascal-francis.inist.fr/vibad/index.php?action=getRecordDetail&idt=2398765$$DView record in Pascal Francis
BookMark eNptkL1PwzAUxD0UibYwsDJ5QEgMae24_hpRxZfUwgDMkeu8tEapHewUxH9PqlQdENOT3v3udLoRGvjgAaELSiaUEj1VeiIJZYIN0JASwTIhpDhFo5Q-CCGKytkQPS9D6SoHJV4sX7Gp1yG6drNNuAoRpwbAbnATg4WUnF_j707ExmNTmqZ1X4B9cAmwNd5CXUM8QyeVqROcH-4Yvd_fvc0fs8XLw9P8dpFZRkSbcWWsYFrksCo1Z4xzI_VMas0k55yKXMBK0xkjpDRWq1zmSgihuk9uK1ESNkbXfW7X7XMHqS22Lu0rGA9hl4pcMS5olzxGVwfQJGvqKnZNXSqa6LYm_hQ500oK3mE3PWZjSClCdSQoKfZjFkoX_ZgdO_3DWtea1gXfRuPqfx2XvcMBwDH5IP4CP86ANA
CODEN IESPEJ
CitedBy_id crossref_primary_10_1049_iet_spr_2016_0111
crossref_primary_10_1016_j_compeleceng_2003_04_001
crossref_primary_10_1121_1_5007278
crossref_primary_10_1109_TSP_2007_894406
crossref_primary_10_1049_iet_spr_2008_0258
crossref_primary_10_1049_el_20081432
crossref_primary_10_1186_1687_6180_2013_187
crossref_primary_10_3844_ajeassp_2010_710_717
crossref_primary_10_1109_LSP_2009_2027638
crossref_primary_10_1121_1_1536624
crossref_primary_10_1121_1_1624064
crossref_primary_10_1109_TSP_2002_1011199
crossref_primary_10_1109_TSP_2006_887155
crossref_primary_10_1109_TSA_2004_832993
crossref_primary_10_1016_j_amc_2004_06_087
crossref_primary_10_1155_2008_529480
crossref_primary_10_1109_TASL_2007_896670
crossref_primary_10_1016_j_jfranklin_2006_08_002
crossref_primary_10_1109_TIM_2024_3372219
crossref_primary_10_7840_KICS_2011_36C_3_183
crossref_primary_10_1088_1009_1963_16_2_014
crossref_primary_10_1121_1_1316095
crossref_primary_10_1007_s10772_017_9425_1
crossref_primary_10_1109_TSP_2005_855108
crossref_primary_10_1016_j_apacoust_2004_02_004
crossref_primary_10_1109_LSP_2015_2415875
crossref_primary_10_1002_acs_3011
crossref_primary_10_1109_ACCESS_2019_2951145
crossref_primary_10_32604_cmc_2021_014981
crossref_primary_10_1007_s11760_023_02771_0
crossref_primary_10_1121_1_3003933
crossref_primary_10_1007_s11465_010_0001_5
crossref_primary_10_1109_JLT_2024_3418816
crossref_primary_10_1016_j_compbiomed_2013_12_009
crossref_primary_10_1109_TIM_2009_2017150
crossref_primary_10_4028_www_scientific_net_AMM_121_126_4259
crossref_primary_10_1080_00207217_2010_497670
crossref_primary_10_1016_S0016_0032_00_00007_7
crossref_primary_10_1109_LSP_2008_2008584
crossref_primary_10_1109_TASL_2012_2234112
crossref_primary_10_1121_1_2139073
crossref_primary_10_1109_TCST_2006_880185
crossref_primary_10_1250_ast_28_153
crossref_primary_10_1587_transfun_E94_A_1576
crossref_primary_10_3233_JIFS_189159
Cites_doi 10.1007/BF01012112
10.1109/TCS.1986.1085982
10.1109/78.236504
10.1016/0167-6393(90)90019-6
10.1080/00016489.1990.12088412
10.1109/31.31337
10.1109/TASSP.1986.1164798
10.1109/ICASSP.1983.1172047
10.1109/31.1709
10.1109/TASSP.1987.1165232
10.1109/TCOM.1980.1094711
10.1016/0165-1684(90)90013-O
10.1109/89.397095
10.1109/PROC.1980.11774
10.1109/TAC.1967.1098599
10.1109/78.193228
10.1109/78.139261
10.1109/TSP.1993.193134
10.1109/89.221372
10.1109/JSAC.1984.1146062
10.1109/TASSP.1986.1164777
10.1109/TASSP.1981.1163596
10.1109/TASSP.1986.1164814
10.1109/TASSP.1982.1163933
10.1109/TASSP.1987.1165167
10.1109/TAP.1982.1142739
10.1121/1.402446
10.1121/1.381436
10.1121/1.406676
10.1109/78.286951
10.1109/78.286952
10.1109/TASSP.1986.1164914
10.1109/78.218137
10.1109/ICASSP.1980.1170939
ContentType Journal Article
Copyright 1998 INIST-CNRS
Copyright_xml – notice: 1998 INIST-CNRS
DBID RIA
RIE
AAYXX
CITATION
IQODW
7SC
8FD
JQ2
L7M
L~C
L~D
DOI 10.1109/89.701363
DatabaseName IEEE All-Society Periodicals Package (ASPP) 1998–Present
IEEE Electronic Library (IEL)
CrossRef
Pascal-Francis
Computer and Information Systems Abstracts
Technology Research Database
ProQuest Computer Science Collection
Advanced Technologies Database with Aerospace
Computer and Information Systems Abstracts – Academic
Computer and Information Systems Abstracts Professional
DatabaseTitle CrossRef
Computer and Information Systems Abstracts
Technology Research Database
Computer and Information Systems Abstracts – Academic
Advanced Technologies Database with Aerospace
ProQuest Computer Science Collection
Computer and Information Systems Abstracts Professional
DatabaseTitleList Computer and Information Systems Abstracts

Database_xml – sequence: 1
  dbid: RIE
  name: IEEE Electronic Library (IEL)
  url: https://proxy.k.utb.cz/login?url=https://ieeexplore.ieee.org/
  sourceTypes: Publisher
DeliveryMethod fulltext_linktorsrc
Discipline Engineering
Applied Sciences
EndPage 351
ExternalDocumentID 2398765
10_1109_89_701363
701363
GroupedDBID -~X
0R~
29I
5GY
6IK
97E
AAJGR
AAWTH
ABAZT
ABJNI
ABQJQ
ABVLG
ACGFS
AETIX
AGQYO
AHBIQ
AI.
AIBXA
ALLEH
ALMA_UNASSIGNED_HOLDINGS
BEFXN
BFFAM
BGNUA
BKEBE
BPEOZ
CS3
DU5
EBS
EJD
HZ~
H~9
ICLAB
IFIPE
IFJZH
IPLJI
JAVBF
LAI
M43
O9-
OCL
RIA
RIE
RNS
TN5
VH1
AAYXX
CITATION
IQODW
RIG
7SC
8FD
JQ2
L7M
L~C
L~D
ID FETCH-LOGICAL-c306t-58ac63962ebd953355a794799375551626eb914300dac98272866681432cf6d03
IEDL.DBID RIE
ISSN 1063-6676
IngestDate Fri Sep 05 07:03:49 EDT 2025
Mon Jul 21 09:17:02 EDT 2025
Thu Apr 24 23:44:01 EDT 2025
Wed Oct 01 01:23:19 EDT 2025
Wed Aug 27 02:56:47 EDT 2025
IsPeerReviewed false
IsScholarly false
Issue 4
Keywords Additive noise
Gaussian noise
Correlation analysis
Least squares method
Noise reduction
Adaptive filtering
Computational complexity
Speech processing
Implementation
Language English
License https://ieeexplore.ieee.org/Xplorehelp/downloads/license-information/IEEE.html
CC BY 4.0
LinkModel DirectLink
MergedId FETCHMERGED-LOGICAL-c306t-58ac63962ebd953355a794799375551626eb914300dac98272866681432cf6d03
Notes ObjectType-Article-2
SourceType-Scholarly Journals-1
ObjectType-Feature-1
content type line 23
PQID 28356153
PQPubID 23500
PageCount 14
ParticipantIDs crossref_primary_10_1109_89_701363
ieee_primary_701363
crossref_citationtrail_10_1109_89_701363
proquest_miscellaneous_28356153
pascalfrancis_primary_2398765
ProviderPackageCode CITATION
AAYXX
PublicationCentury 1900
PublicationDate 1998-07-01
PublicationDateYYYYMMDD 1998-07-01
PublicationDate_xml – month: 07
  year: 1998
  text: 1998-07-01
  day: 01
PublicationDecade 1990
PublicationPlace New York, NY
PublicationPlace_xml – name: New York, NY
PublicationTitle IEEE transactions on speech and audio processing
PublicationTitleAbbrev T-SAP
PublicationYear 1998
Publisher IEEE
Institute of Electrical and Electronics Engineers
Publisher_xml – name: IEEE
– name: Institute of Electrical and Electronics Engineers
References douglas (ref37) 1992
ref35
ref13
ref34
ref12
ref15
ref36
ref14
ieee (ref39) 1969
ref30
ref33
ref11
ref32
ref2
ref17
ref16
ref19
ref18
ref24
ref23
ref26
zurek (ref38) 0
ref25
ref20
ref42
ref41
greenberg (ref22) 1994
ref21
ref28
ref27
weiss (ref5) 1987; 24
ref29
ref8
dillier (ref10) 1993; 30
ref7
ref4
ref3
ref6
haykin (ref31) 1986
widrow (ref1) 1985
ref40
van compernolle (ref9) 1990; 469
References_xml – year: 1994
  ident: ref22
  publication-title: Improved design of microphone-array hearing aids
– ident: ref36
  doi: 10.1007/BF01012112
– volume: 30
  start-page: 95
  year: 1993
  ident: ref10
  article-title: digital signal processing (dsp) applications for multiband loudness correction digital hearing aids and cochlear implants
  publication-title: J Rehab Res Develop
– ident: ref15
  doi: 10.1109/TCS.1986.1085982
– ident: ref20
  doi: 10.1109/78.236504
– ident: ref4
  doi: 10.1016/0167-6393(90)90019-6
– year: 1985
  ident: ref1
  publication-title: Adaptive Signal Processing
– volume: 24
  start-page: 93
  year: 1987
  ident: ref5
  article-title: use of an adaptive noise canceler as an input preprocessor for a hearing aid
  publication-title: J Rehab Res Develop
– volume: 469
  start-page: 76
  year: 1990
  ident: ref9
  article-title: hearing aids using binaural processing principles
  publication-title: Acta Otolaryngologica Suppl
  doi: 10.1080/00016489.1990.12088412
– ident: ref17
  doi: 10.1109/31.31337
– ident: ref27
  doi: 10.1109/TASSP.1986.1164798
– ident: ref33
  doi: 10.1109/ICASSP.1983.1172047
– ident: ref23
  doi: 10.1109/31.1709
– ident: ref25
  doi: 10.1109/TASSP.1987.1165232
– ident: ref7
  doi: 10.1109/TCOM.1980.1094711
– ident: ref24
  doi: 10.1016/0165-1684(90)90013-O
– ident: ref6
  doi: 10.1109/89.397095
– ident: ref2
  doi: 10.1109/PROC.1980.11774
– ident: ref32
  doi: 10.1109/TAC.1967.1098599
– ident: ref21
  doi: 10.1109/78.193228
– year: 1992
  ident: ref37
  publication-title: Nonlinear stochastic gradient adaptation Analysis and optimization
– ident: ref28
  doi: 10.1109/78.139261
– ident: ref14
  doi: 10.1109/TSP.1993.193134
– ident: ref18
  doi: 10.1109/89.221372
– year: 1969
  ident: ref39
  publication-title: IEEE Recommended Practice for Speech Quality Measurements
– ident: ref3
  doi: 10.1109/JSAC.1984.1146062
– year: 0
  ident: ref38
– ident: ref8
  doi: 10.1109/TASSP.1986.1164777
– ident: ref12
  doi: 10.1109/TASSP.1981.1163596
– ident: ref16
  doi: 10.1109/TASSP.1986.1164814
– ident: ref26
  doi: 10.1109/TASSP.1982.1163933
– ident: ref13
  doi: 10.1109/TASSP.1987.1165167
– ident: ref42
  doi: 10.1109/TAP.1982.1142739
– ident: ref11
  doi: 10.1121/1.402446
– ident: ref40
  doi: 10.1121/1.381436
– ident: ref41
  doi: 10.1121/1.406676
– ident: ref29
  doi: 10.1109/78.286951
– year: 1986
  ident: ref31
  publication-title: Adaptive Filter Theory
– ident: ref30
  doi: 10.1109/78.286952
– ident: ref35
  doi: 10.1109/TASSP.1986.1164914
– ident: ref19
  doi: 10.1109/78.218137
– ident: ref34
  doi: 10.1109/ICASSP.1980.1170939
SSID ssj0008174
Score 1.4208391
Snippet A desired signal corrupted by additive noise can often be recovered by an adaptive noise canceller using the least mean squares (LMS) algorithm. A major...
SourceID proquest
pascalfrancis
crossref
ieee
SourceType Aggregation Database
Index Database
Enrichment Source
Publisher
StartPage 338
SubjectTerms Additive noise
Algorithm design and analysis
Applied sciences
Degradation
Exact sciences and technology
Fluctuations
Information, signal and communications theory
Least squares approximation
Noise cancellation
Signal analysis
Signal processing
Speech analysis
Speech processing
Telecommunications and information theory
Title Modified LMS algorithms for speech processing with an adaptive noise canceller
URI https://ieeexplore.ieee.org/document/701363
https://www.proquest.com/docview/28356153
Volume 6
hasFullText 1
inHoldings 1
isFullTextHit
isPrint
journalDatabaseRights – providerCode: PRVIEE
  databaseName: IEEE Electronic Library (IEL)
  issn: 1063-6676
  databaseCode: RIE
  dateStart: 19930101
  customDbUrl:
  isFulltext: true
  dateEnd: 20051231
  titleUrlDefault: https://ieeexplore.ieee.org/
  omitProxy: false
  ssIdentifier: ssj0008174
  providerName: IEEE
link http://utb.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwjV07T8MwELZoJxh4FBDlaSEGlrR52XFGhKgqRLsAElvk2BdaUZKqSRd-Pb44rXgNbJFzjpTz6zvf3XeEXHnKE9zT3PGBKScMY-0IA2sd0OhQlAH6sjDaYsyHz-H9C3tpeLbrXBgAqIPPoIePtS9fF2qJV2X9CAnGghZpRYLbVK31piss4bIxcAIHwzYbEiHPjfsi7tmO346eupYKRkLK0igjs1Usfm3I9Skz2LHp22VNTojBJW-9ZZX21McP6sZ__sAu2W7QJr2x02OPbEDeIVtfOAj3yXhU6GlmgCh9GD1SOXstFtNq8l5Sg2ZpOQdQEzq32QRGnuK9LZU5lVrOcaekeTEtgSqcPJhVeECeB3dPt0OnKbLgKGMtVA4TUhmUwn1INYaaMibNEo0QtjB0ovkc0tiAKtfVUsXCj3xhLB5hWnyVce0Gh6SdFzkcEZql6EVkkIVMhp7UaeaKUGkOka9UGkVdcr3Sf6IaBnIshDFLakvEjRMRJ1ZFXXK5Fp1b2o2_hDqo4rXAqvXs25iuXyPTYcRZl1ysxjgxSwmVI3MolmWC1HOIf4___O4J2bT5iBiqe0ra1WIJZwaQVOl5PRU_AcLT3Lc
linkProvider IEEE
linkToHtml http://utb.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwjV07T8MwELZ4DMDAo4AoTwsxsKQkqZ04I0KgAm0XQOoWOfaFVpSkatKFX48vTiteA1vknCPl_PrOd_cdIRee8kTg6cDxgSuHsUg7wsBaBzQ6FGUbfVkYbdEPOi_sYcAHNc92lQsDAFXwGbTwsfLl61zN8KrsKkSCsfYyWeWMMW6TtRbbrrCUy8bEaTsYuFnTCHludCWilu367fCpqqlgLKQsjDpSW8fi15ZcnTN3WzaBu6joCTG85K01K5OW-vhB3vjPX9gmmzXepNd2guyQJcgaZOMLC-Eu6fdyPUoNFKXd3hOV49d8OiqH7wU1eJYWEwA1pBObT2DkKd7cUplRqeUE90qa5aMCqMLpg3mFe-Tl7vb5puPUZRYcZeyF0uFCKoNTAh8SjcGmnEuzSEMELhzdaH4ASWRgletqqSLhh74wNo8wLb5KA-2298lKlmdwQGiaoB-RQ8q4ZJ7USeoKpnQAoa9UEoZNcjnXf6xqDnIshTGOK1vEjWIRxVZFTXK-EJ1Y4o2_hBqo4oXAvPXk25guXiPXYRjwJjmbj3FsFhMqR2aQz4oYyecQAR_--d0zstZ57nXj7n3_8Yis2-xEDNw9JivldAYnBp6UyWk1LT8BdoHgBA
openUrl ctx_ver=Z39.88-2004&ctx_enc=info%3Aofi%2Fenc%3AUTF-8&rfr_id=info%3Asid%2Fsummon.serialssolutions.com&rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Ajournal&rft.genre=article&rft.atitle=Modified+LMS+algorithms+for+speech+processing+with+an+adaptive+noise+canceller&rft.jtitle=IEEE+transactions+on+speech+and+audio+processing&rft.au=Greenberg%2C+J+E&rft.date=1998-07-01&rft.issn=1063-6676&rft.volume=6&rft.issue=4&rft.spage=338&rft.epage=351&rft_id=info:doi/10.1109%2F89.701363&rft.externalDBID=NO_FULL_TEXT
thumbnail_l http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/lc.gif&issn=1063-6676&client=summon
thumbnail_m http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/mc.gif&issn=1063-6676&client=summon
thumbnail_s http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/sc.gif&issn=1063-6676&client=summon