Download pdf - Journal AES 2002 Abr Vol 50 Num 4

sustainingmemberorganizations AESAES

VO

LU

ME

50,NO

.4JO

UR

NA

L O

F T

HE

AU

DIO

EN

GIN

EE

RIN

G S

OC

IET

Y2002 A

PR

IL

JOURNAL OF THE AUDIO ENGINEERING SOCIETYAUDIO / ACOUSTICS / APPLICATIONSVolume 50 Number 4 2002 April

In this issue…

Synthesis As Music Restoration

TelecommunicationsReproduction Quality

Impulse Response Methods

Virtual Auditory-SpaceHeadphone Issues

Features…

21st ConferenceSt. Petersburg, Russia—Preview

22nd ConferenceEspoo, Finland—Preview

Update: Sections Directory

The Audio Engineering Society recognizes with gratitude the financialsupport given by its sustaining members, which enables the work ofthe Society to be extended. Addresses and brief descriptions of thebusiness activities of the sustaining members appear in the Octoberissue of the Journal.

The Society invites applications for sustaining membership. Informa-tion may be obtained from the Chair, Sustaining Memberships Com-mittee, Audio Engineering Society, 60 East 42nd St., Room 2520,New York, New York 10165-2520, USA, tel: 212-661-8528. Fax: 212-682-0477.

ACO Pacific, Inc.Air Studios Ltd.AKG Acoustics GmbHAKM Semiconductor, Inc.Amber Technology LimitedAMS Neve plcATC Loudspeaker Technology Ltd.Audio LimitedAudiomatica S.r.l.Audio Media/IMAS Publishing Ltd.Audio Precision, Inc.AudioScience, Inc.Audio-Technica U.S., Inc.AudioTrack CorporationAudio-Visual Land Pte Ltd.Autograph Sound Ltd.B & W Loudspeakers LimitedBMP RecordingBritish Broadcasting CorporationBSS Audio Cadac Electronics PLCCalrec AudioCanford Audio plcCEDAR Audio Ltd.Celestion International LimitedCentre for Signal ProcessingCerwin-Vega, IncorporatedCommunity Professional Loudspeakers, Inc.Cox Audio EngineeringCrystal Audio Products/Cirrus Logic Inc.D.A.S. Audio, S.A.D.A.T. Ltd.dCS Ltd.Deltron Emcon LimitedDigigramDigital Audio Disc CorporationDolby Laboratories, Inc.DRA LaboratoriesDTS, Inc.DYNACORD, EVI Audio GmbHEastern Acoustic Works, Inc.

Event Electronics, LLCFerrotec (USA) CorporationFocusrite Audio Engineering Ltd.Fostex America, a division of Foster Electric

U.S.A., Inc.FreeSystems Private LimitedFTG Sandar TeleCast ASGentner Communications Corp.Harman BeckerHHB Communications Ltd.Innova SONInnovative Electronic Designs (IED), Inc.International Federation of the Phonographic

IndustryJBL ProfessionalJensen Transformers Inc.Kawamura Electrical LaboratoryKEF Audio (UK) LimitedKenwood U.S.A. CorporationKlark Teknik Group (UK) PlcKlipsch L.L.C.Lectret Precision Pte. Ltd.Leitch Technology CorporationLindos ElectronicsMagnetic Reference Laboratory (MRL) Inc.Martin Audio Ltd.Meridian Audio LimitedMetropolis Studios and MasteringMiddle Atlantic Products Inc.Mosses & MitchellM2 Gauss Corp.National Semiconductor CorporationGeorg Neumann GmbH Neutrik AGNVisionNXT (New Transducers Ltd.)1 LimitedOntario Institute of Audio Recording TechnologyOutline sncPRIMEDIA Business Magazines & Media Inc.Prism Sound

Pro-Bel LimitedPro-Sound NewsRadio Free AsiaRane CorporationRecording ConnectionRocket NetworkRoyal National Institute for the BlindRycote Microphone Windshields Ltd.SADiESanctuary Studios Ltd.Sekaku Electron Ind. Co., Ltd.Sennheiser Electronic CorporationShure Inc.Snell & Wilcox Ltd.Solid State Logic, Ltd.Sony Broadcast & Professional EuropeSound Devices LLCSound On Sound Ltd.Soundcraft Electronics Ltd.Soundtracs plcSowter Audio TransformersSRS Labs, Inc.Stage AccompanySterling Sound, Inc.Studer North America Inc.Studer Professional Audio AGTannoy LimitedTASCAMTHAT CorporationTOA Electronics, Inc.TommexTouchtunes Music Corp.United Entertainment Media, Inc.Uniton AGUniversity of Essex, Dept. of Electronic

Systems EngineeringUniversity of SalfordUniversity of Surrey, Dept. of Sound RecordingWenger CorporationJ. M. Woodgate and AssociatesYamaha Research and Development

AUDIO ENGINEERING SOCIETY, INC.INTERNATIONAL HEADQUARTERS

60 East 42nd Street, Room 2520, New York, NY 10165-2520, USATel: +1 212 661 8528 . Fax: +1 212 682 0477E-mail: [email protected] . Internet: http://www.aes.org

Roger K. Furness Executive DirectorSandra J. Requa Executive Assistant to the Executive Director

ADMINISTRATION

STANDARDS COMMITTEE

GOVERNORS

OFFICERS 2001/2002

Karl-Otto BäderMarina Bosi

Theresa LeonardDavid Robinson

Annemarie StaepelaereRoland Tan

Juergen WahlNick Zacharov

Gerald Gibson Chair Dietrich Schuller Vice Chair

Mendel Kleiner Chair Mark Ureda Vice Chair

SC-04-01 Acoustics and Sound Source Modeling Richard H. Campbell, Wolfgang Ahnert

SC-04-02 Characterization of Acoustical MaterialsPeter D’Antonio, Trevor J. Cox

SC-04-03 Loudspeaker Modeling and Measurement David Prince, Steve Hutt

SC-04-04 Microphone Measurement and CharacterizationDavid Josephson, Jackie Green

SC-04-07 Listening Tests: David Clark, T. Nousaine

SC-06-01 Audio-File Transfer and Exchange Mark Yonge, Brooks Harris

SC-06-02 Audio Applications Using the High Performance SerialBus (IEEE: 1394): John Strawn, Bob Moses, Yoshi Sawada

SC-06-04 Internet Audio Delivery SystemKarlheinz Brandenburg, Scott Levine

SC-06-06 Audio MetadataC. Chambers

Garry Margolis President

Kees A. Immink President-Elect

Roy Pritts Past President

Jim Anderson Vice President Eastern Region, USA/Canada

James A. Kaiser Vice PresidentCentral Region, USA/Canada

Bob Moses Vice President,Western Region, USA/Canada

Søren Bech Vice PresidentNorthern Region, Europe

Markus ErneVice President, Central Region, Europe

Daniel Zalay Vice President, Southern Region, Europe

Mercedes Onorato Vice President,Latin American Region

Neville ThieleVice President, International Region

Ronald Streicher Secretary

Marshall Buck Treasurer

TECHNICAL COUNCIL

Wieslaw V. Woszczyk ChairJürgen Herre and

Robert Schulein Vice Chairs

COMMITTEES

SC-02-01 Digital Audio Measurement Techniques Richard C. Cabot, I. Dennis, M. Keyhl

SC-02-02 Digital Input-Output Interfacing: Robert A. Finger Julian Dunn

SC-02- 05 Synchronization: Robin Caine

John P. Nunn Chair Robert A. Finger Vice Chair

Robin Caine Chair Steve Harris Vice Chair

John P. Nunn Chair

Tim Shelton Vice Chair

J. G. (Jay) McKnightVice Chair, Western Hemisphere

Mark YongeSecretary, Standards Manager

Daniel QueenStandards Advisor

Yoshizo Sohma Vice Chair, International

SC-02 SUBCOMMITTEE ON DIGITAL AUDIO

Working Groups

SC-03 SUBCOMMITTEE ON THE PRESERVATION AND RESTORATIONOF AUDIO RECORDING

Working Groups

SC-04 SUBCOMMITTEE ON ACOUSTICS

Working Groups

SC-06 SUBCOMMITTEE ON NETWORK AND FILE TRANSFER OF AUDIO

Working Groups

TECHNICAL COMMITTEES SC-03-01 Analog Recording: J. G. McKnight

SC-03-02 Transfer Technologies: Lars Gaustad, Greg Faris, Dave Wickstrom

SC-03-04 Storage and Handling of Media: Ted Sheldon, Gerd Cyrener

SC-03-06 Digital Library and Archives Systems: William Storm Joe Bean, Werner Deutsch

SC-03-12 Forensic Audio: Tom Owen, M. McDermott, Eddy Bogh Brixen

STANDARDSJohn P. Nunn Chair

TELLERSChristopher V. Freitag Chair

WOMEN IN AUDIOCosette Collier Chair

Correspondence to AES officers and committee chairs should be addressed to them at the society’s international headquarters.

Irving Joel Chair John Woodgate Vice ChairVice Chair

SC-05-02 Single-programme Connections Ray Rayburn, Werner Bachmann

SC-05-03 Multiple-programme ConnectionsDavid Tosti-Lane, J. Chester

SC-05-05 Grounding and EMC Practices Bruce Olson, Jim Brown

SC-05 SUBCOMMITTEE ON INTERCONNECTIONS

Working Groups

ACOUSTICS & SOUNDREINFORCEMENT

Mendel Kleiner ChairKurt Graffy Vice Chair

ARCHIVING, RESTORATION ANDDIGITAL LIBRARIES

David Ackerman Chair

CODING OF AUDIO SIGNALSJames Johnston and

Jürgen Herre Cochairs

AUTOMOTIVE AUDIORichard S. Stroud Chair

Tim Nind Vice Chair

HIGH-RESOLUTION AUDIOMalcolm Hawksford Chair

Vicki R. Melchior andTakeo Yamamoto Vice Chairs

LOUDSPEAKERS & HEADPHONESDavid Clark Chair

Juha Backman Vice Chair

MICROPHONES & APPLICATIONSDavid Josephson Chair

Wolfgang Niehoff Vice Chair

MULTICHANNEL & BINAURALAUDIO TECHNOLOGIESFrancis Rumsey Chair

Gunther Theile Vice Chair

NETWORK AUDIO SYSTEMSJeremy Cooperstock ChairRobert Rowe and Thomas

Sporer Vice Chairs

OPTICAL RECORDINGTakeo Yamamoto Chair

Reinier ten Kate Vice Chair

PERCEPTION & SUBJECTIVEEVALUATION OF AUDIO SIGNALS

Durand Begault ChairSøren Bech and Eiichi Miyasaka

Vice Chairs

SIGNAL PROCESSINGRonald Aarts Chair

James Johnston and Christoph M.Musialik Vice Chairs

STUDIO PRACTICES & PRODUCTIONGeorge Massenburg Chair

Alan Parsons, David Smith andMick Sawaguchi Vice Chairs

TRANSMISSION & BROADCASTINGStephen Lyman Chair

AWARDSDavid Robinson Chair

CONFERENCE POLICYSøren Bech Chair

CONVENTION POLICY & FINANCEMarshall Buck Chair

EDUCATIONDon Puluse Chair

FUTURE DIRECTIONSGarry Margolis Chair

HISTORICALJ. G. (Jay) McKnight Chair

Ted Sheldon Vice ChairDonald J. Plunkett Chair Emeritus

LAWS & RESOLUTIONSKees A. Immink Chair

MEMBERSHIP/ADMISSIONSFrancis Rumsey Chair

NOMINATIONSRoy Pritts Chair

PUBLICATIONS POLICYRichard H. Small Chair

REGIONS AND SECTIONSSubir Pramanik Chair

AES Journal of the Audio Engineering Society(ISSN 0004-7554), Volume 50, Number 4, 2002 AprilPublished monthly, except January/February and July/August when published bi-monthly, by the Audio Engineering Society, 60 East 42nd Street, New York, NewYork 10165-2520, USA, Telephone: +1 212 661 8528. Fax: +1 212 682 0477. E-mail: [email protected]. Periodical postage paid at New York, New York, and at anadditional mailing office. Postmaster: Send address corrections to Audio Engineer-ing Society, 60 East 42nd Street, New York, New York 10165-2520.

The Audio Engineering Society is not responsible for statements made by itscontributors.

COPYRIGHTCopyright © 2002 by the Audio Engi-neering Society, Inc. It is permitted toquote from this Journal with custom-ary credit to the source.

COPIESIndividual readers are permitted tophotocopy isolated articles for re-search or other noncommercial use.Permission to photocopy for internalor personal use of specific clients isgranted by the Audio Engineering So-ciety to libraries and other users regis-tered with the Copyright ClearanceCenter (CCC), provided that the basefee of $1.00 per copy plus $0.50 perpage is paid directly to CCC, 222Rosewood Dr., Danvers, MA 01923,USA. 0004-7554/95. Photocopies ofindividual articles may be orderedfrom the AES Headquarters office at$5.00 per article.

REPRINTS AND REPUBLICATIONMultiple reproduction or republica-tion of any material in this Journal requires the permission of the AudioEngineering Society. Permissionmay also be required from the author(s). Send inquiries to AES Edi-torial office.

SUBSCRIPTIONSThe Journal is available by subscrip-tion. Annual rates are $160.00 surfacemail, $205.00 air mail. For information,contact AES Headquarters.

BACK ISSUESSelected back issues are available:From Vol. 1 (1953) through Vol. 12(1964), $10.00 per issue (members),$15.00 (nonmembers); Vol. 13 (1965)to present, $6.00 per issue (members),$11.00 (nonmembers). For informa-tion, contact AES Headquarters office.

MICROFILMCopies of Vol. 19, No. 1 (1971 Jan-uary) to the present edition are avail-able on microfilm from University Microfilms International, 300 NorthZeeb Rd., Ann Arbor, MI 48106, USA.

ADVERTISINGCall the AES Editorial office or send e-mail to: [email protected].

MANUSCRIPTSFor information on the presentationand processing of manuscripts, seeInformation for Authors.

Patricia M. Macdonald Executive EditorWilliam T. McQuaide Managing EditorGerri M. Calamusa Senior EditorAbbie J. Cohen Senior EditorMary Ellen Ilich Associate EditorPatricia L. Sarch Art Director

EDITORIAL STAFF

Europe ConventionsZevenbunderslaan 142/9, BE-1190 Brussels, Belgium, Tel: +32 2 3457971, Fax: +32 2 345 3419, E-mail for convention information:[email protected] ServicesB.P. 50, FR-94364 Bry Sur Marne Cedex, France, Tel: +33 1 4881 4632,Fax: +33 1 4706 0648, E-mail for membership and publication sales:[email protected] KingdomBritish Section, Audio Engineering Society Ltd., P. O. Box 645, Slough,SL1 8BJ UK, Tel: +441628 663725, Fax: +44 1628 667002, E-mail:[email protected] Japan Section, 1-38-2 Yoyogi, Room 703, Shibuyaku-ku, Tokyo 151-0053, Japan, Tel: +81 3 5358 7320, Fax: +81 3 5358 7328, E-mail: [email protected].

PURPOSE: The Audio Engineering Society is organized for the purposeof: uniting persons performing professional services in the audio engi-neering field and its allied arts; collecting, collating, and disseminatingscientific knowledge in the field of audio engineering and its allied arts;advancing such science in both theoretical and practical applications; andpreparing, publishing, and distributing literature and periodicals relative tothe foregoing purposes and policies.MEMBERSHIP: Individuals who are interested in audio engineering maybecome members of the Society. Applications are considered by the Ad-missions Committee. Grades and annual dues are: Full member, $80.00;Associate member, $80.00; Student member, $40.00. A membership appli-cation form may be obtained from headquarters. Sustaining membershipsare available to persons, corporations, or organizations who wish to supportthe Society. A subscription to the Journal is included with all memberships.

Ronald M. AartsJames A. S. AngusGeorge L. AugspurgerJeffrey BarishJerry BauckJames W. BeauchampSøren BechDurand BegaultBarry A. BlesserJohn S. BradleyRobert Bristow-JohnsonJohn J. BubbersMarshall BuckMahlon D. BurkhardRichard C. CabotEdward M. CherryRobert R. CordellAndrew DuncanJohn M. EargleLouis D. FielderEdward J. Foster

Mark R. GanderEarl R. GeddesDavid GriesingerMalcolm O. J. HawksfordJürgen HerreTomlinson HolmanAndrew HornerJames D. JohnstonArie J. M. KaizerJames M. KatesD. B. Keele, Jr.Mendel KleinerDavid L. KlepperPaul W. KlipschW. Marshall Leach, Jr.Stanley P. LipshitzRobert C. MaherDan Mapes-RiordanJ. G. (Jay) McKnightGuy W. McNallyD. J. MearesRobert A. MoogJames A. Moorer

Martin PolonD. PreisDaniel QueenFrancis RumseyKees A. Schouhamer

ImminkManfred R. SchroederRobert B. SchuleinRichard H. SmallJulius O. Smith IIIGilbert SoulodreHerman J. M. SteenekenJohn StrawnG. R. (Bob) ThurmondFloyd E. TooleEmil L. TorickJohn VanderkooyDaniel R. von

RecklinghausenRhonda WilsonJohn M. WoodgateWieslaw V. Woszczyk

REVIEW BOARD

Flávia Elzinga AdvertisingIngeborg M. StochmalCopy Editor

Barry A. BlesserConsulting Technical Editor

Stephanie Paynes Writer

Daniel R. von Recklinghausen Editor

Eastern Region, USA/CanadaSections: Atlanta, Boston, District of Columbia, New York, Philadelphia, TorontoStudent Sections: American University, Berklee College of Music, CarnegieMellon University, Duquesne University, Fredonia, Full Sail Real WorldEducation, Hampton University, Institute of Audio Research, McGillUniversity, Pennsylvania State University, University of Hartford, Universityof Massachusetts-Lowell, University of Miami, University of North Carolina atAsheville, William Patterson University, Worcester Polytechnic UniversityCentral Region, USA/CanadaSections: Central Indiana, Chicago, Detroit, Kansas City, Nashville, NewOrleans, St. Louis, Upper Midwest, West MichiganStudent Sections: Ball State University, Belmont University, ColumbiaCollege, Michigan Technological University, Middle Tennessee StateUniversity, Music Tech College, SAE Nashville, Northeast CommunityCollege, Ohio University, Ridgewater College, Hutchinson Campus,Southwest Texas State University, University of CincinnatiWestern Region, USA/CanadaSections: Alberta, Colorado, Los Angeles, Pacific Northwest, Portland, San Diego, San Francisco, Utah, VancouverStudent Sections: American River College, Brigham Young University,California State University–Chico, Citrus College, Cogswell PolytechnicalCollege, Conservatory of Recording Arts and Sciences, Denver, Ex’pressionCenter for New Media, Long Beach City College, San Diego State University,San Francisco State University, Cal Poly San Luis Obispo, Stanford University,The Art Institute of Seattle, University of Southern California, VancouverNorthern Region, Europe Sections: Belgian, British, Danish, Finnish, Moscow, Netherlands, Norwegian, St. Petersburg, SwedishStudent Sections: All-Russian State Institute of Cinematography, Danish,Netherlands, St. Petersburg, University of Lulea-PiteaCentral Region, EuropeSections: Austrian, Belarus, Czech, Central German, North German, South German, Hungarian, Lithuanian, Polish, Slovakian Republic, Swiss,UkrainianStudent Sections: Berlin, Czech Republic, Darmstadt, Detmold,Düsseldorf, Graz, Ilmenau, Technical University of Gdansk (Poland), Vienna,Wroclaw University of TechnologySouthern Region, EuropeSections: Bosnia-Herzegovina, Bulgarian, Croatian, French, Greek, Israel,Italian, Portugal, Romanian, Slovenian, Spanish, Turkish, YugoslavianStudent Sections: Croatian, Conservatoire de Paris, Italian, Louis-Lumière School, StrasbourgLatin American Region Sections: Argentina, Brazil, Chile, Colombia (Medellin), Mexico, Uruguay,VenezuelaStudent Sections: Taller de Arte Sonoro (Caracas), University of Javeriana,Bogota, ColombiaInternational RegionSections: Adelaide, Brisbane, Hong Kong, India, Japan, Korea, Malaysia,Melbourne, Philippines, Singapore, Sydney

AES REGIONAL OFFICES

AES REGIONS AND SECTIONS

AES JOURNAL OF THE

AUDIO ENGINEERING SOCIETY

AUDIO/ACOUSTICS/APPLICATIONS

VOLUME 50 NUMBER 4 2002 APRIL

CONTENT

PAPERSRestoration and Enhancement of Solo Guitar Recordings Based on Sound Source Modeling .................................................................................Paulo A. A. Esquef, Vesa Välimäki, and Matti Karjalainen 227By combining techniques for music synthesis and analysis, the authors demonstrated that the restoration of guitar recordings had been corrupted by added noise and limited spectral bandwidth. During the analysis phase the parameters of the synthesis model were extracted and then used to define the synthesis model,which is able to produce a simulation without noise or spectral limits. The initial corruption and the musical complexity limit the accuracy of the parameter extraction. The results of this work should be considered a pilot study.

On the Quality of Hearing One’s Own Voice .....................................Ronald Appel and John G. Beerends 237Evaluating the quality of sound reproduction in a telecommunications environment over long distances is unique and unrelated to the classical problem of reproduction quality. Specifically, the experiments attempt to calibrate the acceptable quality level for the listener’s own voice through the side channel when it has been delayed and distorted. The proposed model represents degradation as a single number to predict the subjective experience of comfort while talking and listening to one’s own voice.

Comparison of Different Impulse Response Measurement Techniques..................................................Guy-Bart Stan, Jean-Jacques Embrechts, and Dominique Archambeau 249Although the impulse response of a linear, time-invariant system is mathematically defined, the choice of empirical methods strongly influences the results when applied to a real room with noise and other imperfections. The authors explored the relationship between the environment under test and four different approaches: maximum-length sequences, inverse repeated sequences, time-stretched pulses, and SineSweep. Recommended choices are suggested based on noise level, accuracy, and calibration effort.

ENGINEERING REPORTSVariability in the Headphone-to-Ear-Canal Transfer Function.......................................................................................................Ken I. McAnally and Russell L. Martin 263The quality of virtual auditory space simulation using headphones is determined, in part, by the accuracy of the head-related transfer function that models the influence of the head and pinna as a function of sound location. A potential problem with the resulting filters is that the transfer function is influenced by the variability of listener headphone placement, which might change the localization accuracy. After careful study the authors concluded that the variability is not sufficient to significantly influence perception.

STANDARDS AND INFORMATION DOCUMENTSAES Standards Committee News........................................................................................................... 267Call for comment communications; digital interfacing; preservation and restoration

FEATURES21st Conference Preview, St. Petersburg, Russia................................................................................. 274

Calendar ................................................................................................................................................. 276Program.................................................................................................................................................. 278Registration Form ................................................................................................................................. 289

22nd Conference Preview, Espoo, Finland............................................................................................ 290Calendar ................................................................................................................................................. 292Program.................................................................................................................................................. 293Registration Form ................................................................................................................................. 301

Updates and Corrections to the 2001/2002 International Sections Directory.................................... 302

DEPARTMENTSReview of Acoustical Patents...........................271News of the Sections ........................................303Sound Track........................................................307New Products and Developments....................308Upcoming Meetings ..........................................310Available Literature ...........................................310

Membership Information...................................311Advertiser Internet Directory............................312AES Special Publications .................................315In Memoriam ......................................................320Sections Contacts Directory ............................322AES Conventions and Conferences ................328

PAPERS

0 INTRODUCTION

Audio enhancement is a wide concept and is closelyrelated to audio restoration. An intuitive idea of audioenhancement is associated with any audio processing thatis able to improve the perceptual quality of an audio sig-nal. The goal of the digital audio restoration field [1] is,ideally, to improve the quality of audio signals extractedfrom old recordings, such as wax cylinders, 78 rpm, long-playing records, magnetic tape, and even digital mediamatrices. The usual approach consists of finding the bestway to capture and transfer the recorded sound from theoriginal matrices to a digital medium and, after that,applying digital signal-processing techniques to removeany disturbance or noise produced by the recording andreproducing system.

The most common tasks of audio restoration algorithmsare to remove impulsive noise and reduce broad-bandnoise from the degraded audio sources. Whereas localizeddisturbances, at least those of short duration, are relativelyeasy to treat, dealing with global types of degradation isstill a challenging task. In particular, in the broad-bandnoise-reduction problem the goal is to find better tradeoffsbetween effective noise reduction and signal distortion[1], [2]. Although the perceptual quality of the restoredsignals plays an important role in this matter, only recently

have psychoacoustic criteria been proposed for audioenhancement purposes [3], [4], still bounded by the lackof an observable clean reference signal.

Usually audio restoration algorithms employ signalmodeling techniques, which deal with the informationavailable in the surface presentation of audio signals, thatis, the attempt to model the waveform representation ofthe audio signal. In sound source modeling (SSM) tech-niques, however, the goal is to model the phenomenonthat has generated the waveform. As a natural conse-quence, a structured audio representation [5] is requiredin SSM.

In addition to SSM, models for the propagation mediumand the receptor characteristics have been increasinglyemployed in audio signal processing. In [6] a generalframework for audio and musical signal processing isdescribed. It shows the hierarchical scales and relation-ships among several levels of audio representations. Infact, actual challenging audio signal-processing applica-tions seem to move toward the incorporation of higherrepresentation levels of audio signals, such as the object-and content-based ones. Among those applications it ispossible to cite sound source recognition [7], sound sourceseparation [8], music retrieval, automatic transcription ofmusic [9], object-based sound source modeling [10], andsound synthesis [11].

Due to the requirement of a structured audio represen-tation when using SSM, its practical use for the analysisand synthesis of audio signals is still limited to specificcases. It is easy to see that for general cases, such as analy-sis and synthesis of polyphonic music, the SSM-basedsystem faces difficult tasks. The analysis part requires

J. Audio Eng. Soc., Vol. 50, No. 4, 2002 April 227

Restoration and Enhancement of Solo GuitarRecordings Based on Sound Source Modeling*

PAULO A. A. ESQUEF,1 VESA VÄLIMÄKI,2,1 AES Member, AND MATTI KARJALAINEN,1 AES Fellow

1Helsinki University of Technology, Laboratory of Acoustics and Audio Signal Processing,FIN-02015 HUT, Espoo, Finland

2Tampere University of Technology, Pori School of Technology and Economics, FIN-28101 Pori, Finland

New propositions to audio restoration and enhancement based on sound source modelingare presented. A case based on the commuted waveguide synthesis algorithm for plucked-string tones is described. The main motivation is to take advantage of prior information ofgenerative models of sound sources when restoring or enhancing musical signals.

* An earlier version of this paper was presented at the 110thConvention of the Audio Engineering Society, Amsterdam, TheNetherlands, 2001 May 12–15, under the title, “Restoration andEnhancement of Instrumental Recordings Based on SoundSource Modeling.” Manuscript received 2001 June 25; revised2001 December 10.

ESQUEF ET AL. PAPERS

instrument recognition, sound source separation, detectionof musical events, and extraction of their features, amongothers. For the synthesis part it is required that the types ofinstruments present in the signal be known and synthesismodels be available for them. Furthermore one must takeinto account that the signal analysis performance may bedegraded when dealing with real recordings, which maycontain spurious background noise, nonlinear distortions,and reverberation.

Nevertheless, SSM has been employed in morerestrained situations. For instance, in [10] an object-basedSSM system for analysis and synthesis of the acoustic gui-tar is presented. The system, which deals with two-voicepolyphony examples, is able to analyze the signal, isolatethe voices, and recreate the signal again using a guitar syn-thesizer. The analysis part involves signal modeling tech-niques, such as sinusoidal modeling [12]–[14], combinedwith auditory modeling to pitch determination and signalseparation. The synthesis part employs the physical mod-eling approach, for example, the digital waveguidemethod [15], which has been used successfully to synthe-sis realistic instrument sounds by taking into accountphysical properties associated with the instruments andtheir particular playing techniques [16].

A similar approach as that adopted in [10] can be usedfor audio restoration purposes if both the analysis and thesynthesis parts can be made robust to the presence of noisein the signal. In addition SSM allows taking advantage ofprevious knowledge of the model parameters associatedwith a high-quality instrument sound. This informationcan be useful when attempting to enhance the sound qual-ity of a poorly recorded instrument. For instance, in thispaper it is shown that it is possible to reconstruct missinghigh-frequency harmonics of guitar tones, since high-quality synthesis models for plucked-string tones areavailable, providing prior knowledge of what would betheir frequency content.

Based on the considerations previously outlined onemust note that the restoration or enhancement of audiosignals within the SSM framework is still restricted tosimple cases. Therefore in this paper, only single and iso-lated acoustic guitar notes are considered. This choiceobviously simplifies both the analysis and the synthesisstages involved in the method. For the SSM of pluckedstrings, the computed waveguide synthesis (CWS) algo-rithm [17], [18] is used. This choice allows obtaining themodel parameters by analyzing recorded tones [19]. Thestudy presented here is divided basically into two parts: aproposition to extend the bandwidth of originally band-limited guitar tones, and a denoising scheme for guitartones which mixes a traditional spectral-based dehissingmethod and SSM.

The paper is organized as follows. In Section 1 theCWS algorithm for plucked-string instruments used in thiswork is reviewed. The SSM-based method to extend thebandwidth of guitar tones is proposed in Section 2. InSection 3 the dehissing of guitar tones is discussed. Expe-rimental results are described in Section 4. Discussion,conclusions, and directions to future works are given inSection 5.

1 CWS METHOD FOR PLUCKED-STRINGINSTRUMENTS

Physical modeling techniques for digital sound synthe-sis of musical instruments have become a popularapproach in recent years. In particular, the digital wave-guide synthesis, first introduced by Smith [15], and its fur-ther improvements and extensions [20] have proved to bewell suited to high-quality synthesis of string instruments.

In the case of plucked-string instruments, a naturalstructure of a physical-based synthesizer system wouldconsist of an impulsive excitation signal as the input of aplucking event model cascaded with a string model and abody model of the instrument. If the plucking, string, andbody models are considered linear and time-invariant sys-tems, it is possible to commute them and combine theplucking and body responses into only one input signal.This is the basic principle of the CWS method [17], [18]for plucked strings. For more detailed information on thedevelopment of a guitar synthesizer based on the CWSmethod, see [21].

1.1 String ModelThe function of the vibrating string model is to simulate

the generation of string modes after the plucking event.Considering an isolated string, its behavior can be effi-ciently simulated by the string model illustrated in Fig. 1,whose transfer function is given by

( )( ) ( )

S zz F z H z1

1

L i

(1)

where z-Li and F(z) are, respectively, the integer and frac-tional parts of the delay line associated with the length ofthe string L. Transfer function H(z) is called the loop filter,and it is in charge of simulating the frequency-dependentlosses of the harmonic modes.

In this work the loop filter is implemented as a one-polelow-pass filter with the transfer function given by

( ) .H z gazaz

a

1

1

1

(2)

The magnitude response of H(z) must not exceed unity inorder to guarantee the stability of S(z). This constraintimposes that 0 < g < 1 and 1 < a < 0.

The presence of the fractional delay filter F(z) isintended to provide a fine-tuning of the fundamental fre-quency by precisely adjusting the length of the string.Here it is implemented as a fourth-order Lagrange inter-polator FIR filter [22]. In this configuration the string-model transfer function S(z) is completely defined by thelength of the loop delay L and the loop filter parameters

228 J. Audio Eng. Soc., Vol. 50, No. 4, 2002 April

Fig. 1. Block diagram of string model.

PAPERS SOUND SOURCE MODELING

g and a. In fact, these string-model parameters dependon the fundamental frequency and the fret number.Therefore they must be estimated for each tone to besynthesized.

1.2 Estimation of String-Model ParametersIn this section the estimation of the string parameters in

the CWS model is discussed. The estimation procedurecan be performed automatically by analyzing recordedtones, as shown in [19], [23].

The first step consists in estimating the fundamentalfrequency of the tone, for instance, through the autocorre-lation method. In this case the analysis is performed overa signal excerpt taken after the attack part of the tone,since the value of the fundamental frequency of plucked-string tones takes some time to stabilize after the pluckinginstant. Then, given an estimate of the fundamental fre-quency f0 the length of the delay line in samples isobtained as

Lf

s

0ft(3)

where fs is the sampling rate of the analyzed signal.The next step consists in estimating the loop filter

parameters. The magnitude response of the loop filteractually defines how the energy of the vibrating stringmodes decays as a function of time. However, the stringmodel defined by S(z) can only simulate the exponentiallydecaying behavior of the ideal string modes. In this casethe time constants of the decaying exponentials have adirect relationship with the magnitude response of theloop filter.

The estimation of the loop filter parameters is carriedout in three basic steps. First the decaying envelope ofeach harmonic is obtained through a pitch-synchronizedSTFT analysis, followed by a magnitude peak pickingalgorithm. Then linear curves are fitted to the envelopeson a logarithmic scale. The resulting set of slopes defineswhat would be the values of the loop gains [or the magni-tude of H(z)] at the harmonic frequencies. Finally H(z) isdesigned via a weighted least-squares procedure in whichthe error between its magnitude and the previously esti-mated values of the loop gains is minimized. A detaileddescription of the procedures used to estimate the string-model parameters is found in [19].

2 BANDWIDTH EXTENSION OF GUITAR TONES

In this section the problem of reconstructing missingspectral information in guitar tones is addressed within theSSM approach. The connections between bandwidthextension and audio restoration appear in two cases: toovercome the intrinsic bandwidth limitations of oldrecording systems in capturing the audio source and, evenmore interestingly, to reconstruct the spectral informationlost during a denoising procedure. The latter case will bediscussed in Section 3.

The test signal used in this study is a single guitar tonewhich was low-pass filtered in order to remove the high-

frequency harmonics while preserving the fundamentalfrequency as well as a few harmonics.

The first step of the bandwidth extension procedure isto estimate the string-model parameters as described inSection 1.

Estimating the fundamental frequency is not problem-atic, assuming that it was preserved in the low-pass-filtered tone. The loop filter design is more critical, sincethere are no harmonics available above a certain frequencyto have their decay rate estimated. Nevertheless, consider-ing the simplicity of the loop filter employed in the stringmodel and the fact that for this type of string model, vari-ations between 25 and 40% in the time constant of thedecay are not perceived [24], it is acceptable to estimatethe decay rate of the missing harmonics by analyzing asimilar full-band guitar tone.

As seen in Section 1, the string model is basically acomb filter tuned at the fundamental frequency and itsharmonics. Thus the main effect of inverse filtering theguitar tone through the string model S(z) is to attenuate thestring modes. The resulting excitation eCWS(k) usually hasa large number of resonances associated with all otherinformation except that of the vibrating string, such asnonlinearities associated with the plucking event, bodyresonances, and coupling between strings.

In the bandwidth extension problem, the analyzed sig-nal is already low-pass filtered, resulting in an excitationwith a low-pass characteristic as well. Therefore it is onlyable to excite the string modes corresponding to the har-monic frequencies originally present in the analyzed sig-nal. However, if an extra amount of energy is added to theexcitation in a proper way, it is possible to excite all themodes of the string model. Thus by altering the excitationsignal it is possible to resynthesize a new tone whosebandwidth is greater than that of the analyzed one.

Typically the frequency response of acoustic guitarbodies exhibits a few slowly decaying resonance modes inthe low-frequency range [25]. Toward higher frequenciesthe number of resonance modes increases, but their decaytime decreases. This characteristic motivates the use ofexponentially decaying white noise to efficiently modelthe high-frequency response of guitar bodies [26].

In this sense a possible strategy to fulfill the informa-tion that is missing in the low-pass-filtered excitationwould consist of adding an artificially generated noiseburst epluck(k) directly to the string model, triggered withthe attack part of eCWS(k), as illustrated in Fig. 2. Thisnoise burst, which can be considered either a rough modelfor the missing high-frequency modes of the guitar bodyor an extra plucking event, must have enough energywithin the entire frequency range in order to fully excite


Fig. 2. Bandwidth extension scheme.


the string model.If the artificial plucking can really excite the string

modes, the resynthesized tone will exhibit harmonics inthe full frequency range, although the decay rates of thepreviously nonexistent modes will be defined only by thestring-model characteristics.

The simplest choice for the artificially generated pluck-ing signal would be an impulse. However, it is known thatthe finger–string interaction is not really impulsive, and abetter option is to generate an impulsive noise burst, forinstance, by windowing a zero-mean Gaussian white-noise sequence. A noise burst of about 10 ms seems a rea-sonable choice to simulate the duration of a typical fin-ger–string interaction.

It is important to notice that the power spectrum densityof the noise burst described before is flat and thus willexcite almost equally all the string modes. However, itwould be desired that the additional noise burst, composedwith the filtered excitation, could emulate a typical spec-tral behavior of the attack part of an excitation correspon-ding to a full-bandwidth tone. A simple option to realizethat is to color the noise burst in a proper way, forinstance, according to known information about typicalspectral characteristics of guitar bodies. Alternatively, onecan obtain this information through the excitation of afull-bandwidth tone, for instance, by estimating the spec-tral envelope of its attack part.

In addition it would be desirable to leave the harmonicsoriginally present in the low-pass-filtered tone undis-torted. However, in real cases it is not a trivial task, sinceno previous information about the bandwidth limitation ofthe analyzed tone is available. If it can be roughly inferred,an arbitrary attenuation in the spectrum of the noise burstcan be included to compensate for the unnecessary extraenergy within the original bandwidth.

The generation of the noise burst, which simulates aplucking event, can be carried out as depicted in Fig. 3.The input sequence n(k) is a zero-mean Gaussian white-noise sequence, the filter E(z) is a coloring filter the mag-nitude response of which must approximate the spectralenvelope of the very beginning of a full-bandwidth excita-tion. The high-pass filter Hhp(z) is optional and can beincluded to compensate for the unnecessary addition ofenergy within the effective bandwidth of the analyzedtone. The gain factor a controls the local signal-to-noiseratio (SNR) at the part of the excitation to be modified.

Naturally, the noise burst is windowed before its addi-tion to the excitation. In this context, the characteristics ofthe synthetic pluck depend on its length in samples, themagnitude response of the filters E(z) and Hhp(z), and thevalue of the gain factor. Further details concerning thechoices of the filters E(z) and Hhp(z), the length of thenoise burst, and the value of gain a are given in Section4.1.

3 SSM AND DENOISING OF GUITAR TONES

In this section a processing scheme that involves theSSM of guitar tones and traditional methods of audiorestoration is proposed to improve the perceptual qualityof dehissed guitar tones.

The basic principle behind spectral-based audio dehiss-ing methods is to split the noisy signal into a certain num-ber of frequency bands and to attenuate the signal on thosebands where the SNR is below a given value [27], [28].The spectral analysis and the corresponding signal recon-struction can be realized via either filterbanks or short-time Fourier transform. Usually dehissing methods sufferfrom a difficult tradeoff between the reduction of the noiseeffects and the distortion of the signal to be restored [2],[1], [27]. As both the noise and the signal share the samespectral range, any attempt to have the noise reduced leadsto a degradation of the signal information to some extent.

One of the most common audio dehissing methods isbased on digital Wiener filtering [1], in which the signal issegmented in short-time frames and the magnitude spec-trum of each frame is weighted according to local estimatesof the SNR at each frequency bin. The lower the SNR at acertain bin, the more attenuated is its magnitude value.When dehissing acoustical musical signals, signal losses aremore prominent at high frequencies, since lower SNRs areobserved in this range. Naturally, auditory properties suchas masking and critical bands help to explain the improvedperceptual quality of dehissed audio signals. However, thelack of a reference signal prevents the appealing use ofauditory-based approaches as well as the employment ofeither purely objective or perceptual-based measures toevaluate the perceptual quality of the dehissed signals.

The results of bandwidth extension, described in Section2, provide a useful appeal to the dehissing problem. Thehard tradeoff between the preservation of the valuable sig-nal information and the noise reduction can be softened onthe grounds that the signal information can be recon-structed afterward if a sound source model is available forthe signal. This possibility poses an alternative view on thedehissing problem as discussed in the following sections.

3.1 Aggressive Dehissing and BandwidthExtension

The possibility of reconstructing lost frequency infor-mation of guitar tones, as described in Section 2, can servethe dehissing problem in the following way. A spectral-based dehissing method with an overestimated value fornoise variance can be employed to perform an aggressivetype of denoising. This choice will surely reduce the per-ceptual effects of the residual noise, but it will lead to anoversmoothed restored signal. A remedy to the over-smoothing problem is to apply a postprocessing stage suchas the SSM-based bandwidth extension to recover the sig-nal information that was lost due to the aggressive dehiss-ing procedure.

In this case the estimation of the string-model parame-ters faces similar problems as those discussed in Section2, when dealing with band-limited guitar tones. Here thehigh-frequency harmonics are either masked by the cor-


Fig. 3. Generation of synthetic plucking event.


rupting noise or absent due to the aggressive dehissingprocedure. Thus their decay rate estimation is prevented.Anyway, the same considerations as in Section 2 regardingthe estimation of the loop filter parameters are applicablein this case. Further details on the implementation of thepreviously described approach are given in Section 4.2.

3.2 Integrated Dehissing and BandwidthExtension

Another strategy for the dehissing problem consists ofintegrating the dehissing and signal reconstruction proce-dures into a single stage. This can be achieved by adapt-ing the dehissing method to process the excitation corre-sponding to the noisy signal. In fact, in the noisyexcitation the energy at the harmonic frequencies is atten-uated and the corrupting noise is colored. Nevertheless, aspectral-based dehissing method is still applicable to theexcitation signal, since it has important resonances asso-ciated with the body modes. Of course, an aggressivedehissing procedure will lead to losses mainly in thehigh-frequency content of the excitation signal. As wasseen in Section 2, the attack part of the excitation has animportant role in the reconstruction of the frequencyinformation of the tone. In this context, if only the attackpart of the excitation is spared from the aggressive dehiss-ing procedure, it will provide enough energy to properlyexcite the string model in order to resynthesize a non-smoothed and noise-free tone.

A possible way to protect the attack part of the excita-tion from the aggressive dehissing is to control the noisevariance estimation used in the spectral-based dehissingmethod artificially. For instance, a gain can be assigned tothe noise variance estimate in such a way that it is set to ahigh value elsewhere except at the attack part, where thevalue of the gain should be set to unity.

Considering that the highest local variance of the exci-tation is observed during its attack part, an automatic pro-cedure can be devised within the frame-by-frame dehiss-ing procedure. First a local estimate of the excitationvariance at the attack part, s2

attack, should be obtained.Then for each frame index i, the estimate of the noise vari-ance is multiplied by a gain given by

,minmine

m ms

s

maxmax

attackattack

2

2J

L

KKK

N

P

OOO

(4)

where s2 is a locally estimated excitation variance withina given frame i, e is a small positive value to prevent adivision by zero, and mmax is a constant value which rep-resents the maximum value of m allowed. Since for mostframes but those associated with the attack part of theexcitation s2 < < s2

attack, their ratio will assume higher val-ues than mmax, implying m mmax. Further details on theimplementation of the integrated dehissing procedure aredescribed in Section 4.2.

4 EXPERIMENTAL RESULTS

In this section experimental results in the bandwidthextension and the dehissing of single guitar tones are

described. The test signal used in both cases is an F4 tonewith a fundamental frequency of 347 Hz. The tone wasrecorded in an anechoic chamber and sampled at 22.05kHz.

4.1 Bandwidth ExtensionThe test signal used in the bandwidth extension experi-

ments was low-pass filtered using a 101th-order equirip-ple FIR filter with a cutoff frequency at 1 kHz, a transitionband of 1 kHz, and an attenuation of 80 dB on the rejec-tion band. In this case, regardless of the filtering proce-dure, the fundamental and the next two harmonic frequen-cies of the tone were preserved.

In the experiment described here, the string-modelparameters were estimated using the original tone. Thischoice was taken as an attempt to isolate the problemsassociated with the estimation of the model parametersand the bandwidth extension procedure. In the followingstep, the excitation corresponding to the low-pass-filteredtone was obtained by inverse filtering.

The artificially generated plucking event epluck(k) wasobtained as shown in Fig. 3. In this experiment n(k) waschosen as a zero mean white Gaussian noise sequence,and E(z) as a second-order resonator tuned at 200 Hz. Thisfrequency corresponds to the lowest mode of the top plateof the guitar body [25]. The radius of the poles was arbi-trarily set to 0.8. It can be seen from Fig. 4 that with theseparameters the magnitude response of E(z) approximatesquite well the spectral envelope associated with the attackpart of a full-bandwidth excitation.

The high-pass filter Hhp was not included to keep thegenerality of the method, since the bandwidth limitationof the analyzed tone is usually not known beforehand.Finally the noise burst was then multiplied by a Hanningwindow of 600 samples, scaled, and added to the attackpart of the excitation. The procedure was automated bydetecting the attack part of the excitation using a magni-tude criterion and then synchronizing both the attack andthe window maxima, as shown in Fig. 5.

It should be noted that the noise burst can be fully char-acterized by the coloring filter and the length of the win-


Fig. 4. Squared magnitude response of E(z) (– – –) and powerspectrum of attack part of a full-bandwidth excitation.


dow. While the latter was chosen according to an estimateof the duration of finger–string interaction, the formerwas designed by considering known features associatedwith the guitar body characteristics.

Based on informal listening tests, it was observed thatcoloring the noise burst has an important effect on thequality of the timbre of the resynthesized tone. The tim-bre of the resynthesized tone also varies depending onthe power of the noise burst, which can be adjusted toproduce a certain local SNR at the attack part of theexcitation.

For instance, in this experiment, if the SNR is set to 40dB, the resynthesized tone does not exhibit great percep-tual differences compared to the low-pass-filtered tone.Reducing the value of the SNR tends to increase the per-ceptual differences and emphasize the plucking event. AnSNR of about 20 dB was found to be a suitable value toachieve a resynthesized tone with perceptual quality closeto that of the original one. On the other hand, lower valuesof SNR, such as 10 dB, overemphasize the plucking event,compromising the perceptual quality of the tone.

The capability of the method to extend the bandwidth ofguitar tones is illustrated in Fig. 6, which shows time–fre-quency analysis plots of the original, the low-pass-filtered,and the resynthesized tones. In this case the noise burstwas generated as described before and scaled to producean SNR of 20 dB at the attack part of the excitation.

Additional tests were performed on the test guitar tonein which its bandwidth was limited to 500 Hz and 3000Hz. The results obtained were similar to those of the pre-vious case. However, the perceptual differences betweenthe original and the 3000-Hz band-limited tone are alreadyless prominent. Therefore the effects of the bandwidthextension procedure are more difficult to perceive.

4.2 DehissingIn the dehissing experiments a zero mean Gaussian

white noise was added to the test guitar tone signal, and itsvariance was adjusted to generate a global SNR of 20 dB.


Fig. 5. (a) Attack part of excitation. (b) Synchronization of noiseburst.

(a)

(b)

Fig. 6. Time–frequency analysis. (a) Original tone. (b) Low-pass-filtered tone. (c) Resynthesized tone.

(a)

(b)

(c)


The first step of the dehissing and postprocessingapproach consisted of dehissing the noisy signal through aWiener filtering scheme, as described in [1]. In this exper-iment signal frames of 256 samples were used with anoverlap of 50%. The noise variance was estimated in thefrequency domain by taking the mean value of the upperquarter of the power spectrum. In addition a gain wasassigned to the noise variance estimate. This gain, whichhereafter will be called noise floor gain, worked as a con-trol parameter for the amount of noise to be removed.

Since a single guitar tone is not a complex signal, itdoes not help in masking the residual noise effects in therestored tone, mainly after its attack part. However, theycan be reduced by overestimating the noise variancewithin the Wiener filtering scheme. Considering theWiener filter configuration and the test signal used in thisexperiment, it was found that a noise floor gain of 30 suf-fices to almost eliminate the residual noise effects in therestored signal despite its strongly smoothed characteristic.

The last step consisted of applying the bandwidth exten-sion procedure to the aggressively dehissed tone in order toreconstruct the lost harmonic frequencies. In this experi-ment good results were attained by employing the sameapproach and parameters that were used in Section 4.1.

In Fig. 7 time–frequency analysis plots of the noisy,aggressively dehissed, and bandwidth-extended signals areshown. As can be seen from Fig. 7(a), the noise masks thehigh–frequency harmonics, which together with the noiseare also removed after the aggressive dehissing procedure[Fig. 7(b)]. Nevertheless, the SSM-based bandwidth exten-sion scheme is able to reconstruct the missing harmonics inthe resynthesized tone [Fig. 7(c)]. In this case the timbre ofthe restored tone is similar to the original one, and theeffects of the residual noise are greatly reduced.

The experiment employing the integrated dehissing andreconstruction approach was performed on the artificiallycorrupted tone described before. The procedure was car-ried out by first estimating the string-model parametersand then obtaining the noisy excitation through inverse fil-tering. The noisy excitation was dehissed via the Wienerfiltering method, here adapted to account for the colorednoise at the excitation as well as for the application of avarying gain to the noise floor estimate (see Section 3.2).

As in the previous dehissing experiment, signal framesof 256 samples were used with an overlap of 50%. Theestimates for the noise floor within each signal frame wereobtained in the same way. In addition the noise floor esti-mates were multiplied by gain factors m, defined in Eq.(4), which were also computed for each frame.

The computation of m involved the following choices.The value of s2

attack was set as the power of the attack partof the noisy excitation. The value of mmax was chosen as50. The value of s2 was set as the power of the noisy exci-tation within a given frame; therefore it is the only param-eter that varies as a function of the frame index. Since thepresence of the noise prevents s2 to assume a null value,the value of e was chosen as zero.

In this case the sequence of values of m as a function ofthe frame index looks like the one shown in Fig. 8(a). Notethat outside the attack part of the excitation, which starts


Fig. 7. Time–frequency analysis. (a) Noisy tone. (b) Dehissedtone. (c) Bandwidth-extended tone.

(a)

(b)

(c)


at frame 30, m mmax. Thus an aggressive dehissing isperformed on the whole excitation except at its beginning.For the sake of simplicity, the values of m are shown forthe first 100 signal frames, which corresponds to approxi-mately 0.6 s in time.

The plot in Fig. 8(b) shows a time–frequency analysisof the resynthesized signal obtained from the previouslydehissed excitation. As can be seen, the procedure is capa-ble of removing the noise and reconstructing the lost har-monics. However, in this case the spectral tilt associatedwith the attack part of the excitation is determined by thatof the noisy excitation. Therefore an undesirable positivebias on the powers of the high-frequency harmonics isobserved. This reduces the perceptual quality of the attackpart of the resynthesized tone, which has a more syntheticquality than the restored tone obtained in the previousexperiment.1

5 DISCUSSION AND CONCLUSIONS

In this paper the enhancement of guitar tones was pre-sented within a sound source modeling framework. First itwas shown how the reconstruction of spectral informationin guitar tones can be attained by means of SSM tech-niques. Then that issue was taken into account in a dehiss-ing scheme, which mixed a traditional spectral-basedmethod with SSM. The results obtained for both the band-width extension and the dehissing experiments demon-strate that the proposed schemes are effective in improv-ing the perceptual quality of the restored tones.

Although showing some potential, the use of SSM foraudio enhancement purposes is still restricted to specialcases. For instance, when attempting to restore severelydegraded instrumental recordings addressing one singleinstrument, the most prominent music events, such as the

melodic lines, should be detected and isolated. Further,their features could be used to calibrate a synthesizer inorder to reconstruct another signal, which would soundsimilar to the original source but free of noise. The choiceof a synthesizer based on physical modeling would pro-vide more flexibility for adjusting the model parametersaccording to the extracted features of the music events aswell as the possibility of taking advantage of other avail-able information about high-quality sound sources.

It is important to note that the enhancement of audiosignals using SSM is still restricted to simple cases. Forinstance, restoration of solo guitar within the SSM frame-work would in itself imply challenging tasks due to thestructural representation required for the musical content.As a consequence, one needs effective and robust methodsto detect and locate the occurrence of musical events, toseparate notes or chords whose content overlaps both intime and frequency, and to extract musical features asso-ciated with the sound events from the observable soundwaveform.

Extensions to more general cases can be viewed as amultilayered problem in which even more demandingtasks would be required, such as recognition and separa-tion of more general musical elements in complex soundsource mixtures. In addition it is plausible to expect thatthe signal analysis performance decreases when dealingwith real recorded sounds. This is due to the possible pres-ence of spurious noises, nonlinear distortions, and evenstrong reverberation in the signal to be analyzed. On thesynthesis side of the chain, the requirements are related tothe development of model-based music synthesizers withmore realistic sounds, and capable of simulating the play-ing features of real performances.

SSM- and content-based audio processing is still in ayouthful stage of development. However, as long as itdevelops into better ways to represent and recreate soundsources, performing audio enhancement within the SSMframework can lead to better results compared to thoseattained by traditional techniques.


Fig. 8. (a) Values of noise floor gain as a function of frame index. (b) Time–frequency analysis of restored tone.

1Sound examples are available at URL: http://www.acoustics.hut.fi/publications/papers/jaes-ssm/.

(a) (b)


6 ACKNOWLEDGMENT

The work of P. Esquef has been supported by a scholar-ship from the Brazilian National Council for Scientificand Technological Development (CNPq-Brazil) and theSound Source Modeling Project of the Academy ofFinland. V. Välimäki has been financed by a postdoctoralresearch grant from the Academy of Finland.

7 REFERENCES

[1] S. J. Godsill and P. J. W. Rayner, Digital AudioRestoration––A Statistical Model Based Approach(Springer, London, 1998).

[2] O. Cappé, “Elimination of the Musical NoisePhenomenon with the Ephraim and Malah NoiseSuppressor,” IEEE Trans. Speech Audio Process., vol. 2,pp. 345–349 (1994 Apr.).

[3] P. J. Wolfe and S. J. Godsill, “The Application ofPsychoacoustic Criteria to the Restoration of MusicalRecordings,” presented at the 108th Convention of theAudio Engineering Society, J. Audio Eng. Soc.(Abstracts), vol. 48, p. 363 (2000 Apr.), preprint 5150.

[4] D. Tsoukalas, J. Mourjopoulos, and G. Kokkinakis,“Perceptual Filters for Audio Signal Enhancement,” J.Audio Eng. Soc., vol. 45, pp. 22–36 (1997 Jan./Feb.).

[5] B. L. Vercoe, W. G. Gardner, and E. D. Scheirer,“Structured Audio: Creation, Transmission, and Render-ing of Parametric Sound Representations,” Proc. IEEE,vol. 86, pp. 922–940 (1998 May).

[6] M. Karjalainen, “Immersion and Content––AFramework for Audio Research,” in Proc. IEEE Workshopon Applications of Signal Processing to Audio andAcoustics (New Paltz, NY, USA, 1999 Oct.), pp. 71–74.

[7] J. C. Brown, “Computer Identification of MusicalInstruments Using Pattern Recognition with CepstralCoefficients as Features,” J. Acoust. Soc. Am., vol. 105,pp. 1933–1941 (1999 Mar.).

[8] R. C. Maher, “Evaluation of a Method for Separ-ating Digitized Duet Signals,” J. Audio Eng. Soc., vol. 38,pp. 956–979 (1990 Dec.).

[9] A. Klapuri, “Automatic Transcription of Music,”Master’s thesis, Tampere University of Technology,Tampere, Finland (1998). Electronic version available atURL: http://www.cs.tut.fi/~klap/iiro/.

[10] T. Tolonen, “Object-Based Sound Source Model-ing for Musical Signals,” presented at the 109th Conven-tion of the Audio Engineering Society, J. Audio Eng. Soc.(Abstracts), vol. 48, p. 1094 (2000 Nov.), preprint 5174.

[11] T. Tolonen, V. Välimäki, and M. Karjalainen,“Evaluation of Modern Sound Synthesis Methods,” Tech.Rep. 48, Helsinki University of Technology, Lab. ofAcoustics and Audio Signal Processing, Espoo, Finland(1998). Postscript version available at URL: http://www.acoustics.hut.fi/publications/.

[12] T. F. Quatieri and R. J. McAulay, “Audio SignalProcessing Based on Sinusoidal Analysis and Synthesis,”in Applications of Digital Signal Processing to Audio andAcoustics, 1st ed., M. Kahrs and K. Brandenburg, Eds.(Kluwer, Boston, MA, 1998), pp. 314–416.

[13] J. O. Smith and X. Serra, “PARSHL: An Anal-ysis/Synthesis Program for Non-Harmonic Sounds Basedon a Sinusoidal Representation,” in Proc. Int. ComputerMusic Conf. (Champaign-Urbana, 1987).

[14] R. C. Maher, “Sinewave Additive Synthesis Re-visited,” presented at the 91st Convention of the AudioEngineering Society, J. Audio Eng. Soc. (Abstracts), vol.39, p. 996 (1991 Dec.), preprint 3128.

[15] J. O. Smith, “Physical Modeling Using DigitalWaveguides,” Computer Music J., vol. 16, no. 4, pp. 74–91 (1991).

[16] M. Laurson, C. Erkut, V. Välimäki, and M.Kuuskankare, “Methods for Modeling Realistic Playing inAcoustic Guitar Synthesis,” Computer Music J., vol. 25,no. 3, pp. 38–49 (2001).

[17] J. O. Smith, “Efficient Synthesis of StringedMusical Instruments,” in Proc. Int. Computer Music Conf.(Tokyo, Japan, 1993 Sept.), pp. 64–71.

[18] M. Karjalainen, V. Välimäki, and Z. Jánosy,“Towards High-Quality Sound Synthesis of the Guitar andString Instruments,” in Proc. Int. Computer Music Conf.(Tokyo, Japan, 1993 Sept.), pp. 56–63.

[19] V. Välimäki, J. Huopaniemi, M. Karjalainen, andZ. Jánosy, “Physical Modeling of Plucked String Instru-ments with Application to Real-Time Sound Synthesis,” J.Audio Eng. Soc., vol. 44, pp. 331–353 (1996 May).

[20] M. Karjalainen, V. Välimäki, and T. Tolonen,“Plucked String Models: From the Karplus–Strong Algo-rithm to Digital Waveguides and Beyond,” ComputerMusic J., vol. 22, no. 3, pp. 17–32 (1998).

[21] V. Välimäki and T. Tolonen, “Development andCalibration of a Guitar Synthesizer,” J. Audio Eng. Soc.,vol. 46, pp. 766–778 (1998 Sept.).

[22] T. I. Laakso, V. Välimäki, M. Karjalainen, andU. K. Laine, “Splitting the Unit Delay––Tools for Frac-tional Delay Filter Design” IEEE Signal Process. Mag.,vol. 13, pp. 30–60 (1996 Jan.).

[23] T. Tolonen, “Model-Based Analysis and Resyn-thesis of Acoustic Guitar Tones,” Master’s thesis, Rep. 46,Helsinki University of Technology, Lab. of Acoustics andAudio Signal Processing, Espoo, Finland (1998 Jan.).Postscript version available at URL: http://www.acoustics.hut.fi/publications/.

[24] H. Järveläinen and T. Tolonen, “Perceptual Toler-ances for Decay Parameters in Plucked String Synthesis,”J. Audio Eng. Soc., vol. 49, no. 11, pp. 1049–1059 (2001Nov.).

[25] O. Christensen and B. B. Vistisen, “Simple Modelfor Low-Frequency Guitar Function,” J. Acoust. Soc. Am.,vol. 68, pp. 758–766 (1980).

[26] M. Karjalainen, H. Penttinen, and V. Välimäki,“More Acoustic Sounding Timbre from Guitar Pickups,”in Proc. 2nd G-6 Workshop on Digital Audio Effects(DAFX’99) (Trondheim, Norway, 1999 Dec.), pp. 41–44.

[27] J. A. Moorer and M. Berger, “Linear-PhaseBandsplitting: Theory and Applications,” J. Audio Eng.Soc., vol. 34, pp. 143–152 (1986 Mar.).

[28] S. V. Vaseghi and R. Frayling-Cork, “Restorationof Old Gramophone Recordings,” J. Audio Eng. Soc., vol.40, pp. 791–801 (1992 Oct.).




THE AUTHORS

Paulo A. A. Esquef was born in Brazil in 1973. Hereceived an engineering degree in electrical engineeringfrom Polytechnic School of the Federal University of Riode Janeiro (UFRJ) in 1997, and an M.Sc. degree in elec-trical engineering from COPPE-UFRJ in 1999. His mas-ter's thesis addressed digital restoration of old recordings.From 1999 to 2000 he worked on research and develop-ment of a DSP system for analysis classification of sonarsignals as part of a cooperation project between the SignalProcessing Laboratory (COPPE-UFRJ) and the BrazilianNavy Research Center (IPqM). Since 2000 he has beenwith the Laboratory of Acoustics and Audio SignalProcessing at Helsinki University of Technology, wherehe is currently pursuing postgraduate studies. He is a grantholder from CNPq, a Brazilian governmental council forfunding research in science and technology. His researchinterests include digital audio restoration, computationalauditory scene analysis, sound synthesis, among others.

Mr. Esquef is an associate member of the IEEE. Helikes to play the piano in his spare time.

Vesa Välimäki was born in Kuorevesi, Finland, in 1968.He studied acoustics, audio signal processing, and infor-mation sciences at Helsinki University of Technology(HUT), Espoo, Finland, and received the master of sci-ence in technology, the licentiate of science in technology,and the doctor of science in technology degrees in electri-cal engineering in 1992, 1994, and 1995, respectively. Hisdoctoral thesis dealt with sound synthesis based on phys-ical modeling. Dr. Välimäki worked at the HUTLaboratory of Acoustics and Audio Signal Processingfrom 1990 until 2001. In 1996 he spent six months as apostdoctoral research fellow with the University ofWestminster, London, UK. He was appointed a docent inaudio signal processing at HUT in 1999.

Since 2001 Dr. Välimäki has been professor of signalprocessing at Pori School of Technology and Economics,Tampere University of Technology, Pori, Finland, where heteaches courses on digital audio, signal processors, and fun-damentals of signal processing. He also lectures on digitalaudio signal processing at HUT and at the Centre for Music

and Technology, Sibelius Academy, Helsinki, Finland. Hisresearch interests are in the fields of musical signal pro-cessing, active noise control, and digital filter design.

Dr. Välimäki is a senior member of the IEEE SignalProcessing Society and is a member of the AES, theInternational Computer Music Association, the AcousticalSociety of Finland, and the Finnish MusicologicalSociety. He is papers chair of the AES 22nd InternationalConference on "Virtual, Synthetic, and EntertainmentAudio," 2002 June, Espoo, Finland. His Web site ishttp://www.pori.tut.fi/~vpvm/. His e-mail address [email protected].

Matti Karjalainen was born in Hankasalmi, Finland, in1946. He received the M.Sc. and Dr. Tech. degrees in elec-trical engineering from the Tampere University ofTechnology in 1970 and 1978, respectively. His doctoralthesis dealt with speech synthesis by rule in Finnish.

From 1980 he was an associate professor and since1986 he has been a full professor in acoustics and audiosignal processing at the Helsinki University ofTechnology on the faculty of electrical engineering. Inaudio technology his interest is in audio signal processing,such as DSP for sound reproduction, perceptually basedsignal processing, as well as music DSP and sound syn-thesis. In addition to audio DSP his research activitiescover speech synthesis, analysis, and recognition, percep-tual auditory modeling, spatial hearing, DSP hardware,software, and programming environments, as well as var-ious branches of acoustics, including musical acousticsand modeling of musical instruments. He has written 250scientific and engineering articles and contributed toorganizing several conferences and workshops. In 1999 heserved as the papers chair of the AES 16th InternationalConference on Spatial Sound Reproduction.

Dr. Karjalainen is an AES fellow and a member of theInstitute of Electrical and Electronics Engineers, theAcoustical Society of America, the European AcousticsAssociation, the International Computer Music Assoc-iation, the European Speech Communication Association,and several Finnish scientific and engineering societies.

P. A. A. Esquef V. Välimäki M. Karjalainen

PAPERS

0 INTRODUCTION

In the field of speech quality and speech intelligibility alarge number of work has been carried out on the relationbetween objective and subjective measurements [1]–[11].In general these papers use the basics of human auditoryperception and speech production to find these relations.All of this work relates to the way we perceive the speechof another person talking. In daily life we also perceiveour own voice via the mouth-to-ear path and via reflec-tions from the environment. This provides a direct feed-back, which enables control over the speech productionprocess. From the moment we start to talk, as a child, weare accustomed to this effect, and normally will never giveit any thought. However, distortions in the perception ofone’s own voice can greatly influence the comfort withwhich we speak and the way we speak. The best knowneffect is the raising of one’s voice in the presence of loudbackground noise, the so-called Lombard effect [12], [13].The opposite effect also exists: when we are presentedwith a loud copy of our own voice over a headphone (loudsidetone), we lower the volume of our voice [13]. If thissidetone is delayed, we start to feel uncomfortable. Forsmall delays (10 ms) and high levels of sidetone thedirect signal that leaks to the ear entrance will interferewith the delayed version, leading to coloration in thesound of our own voice (comb filtering). For mediumdelays (10–30 ms) our voice sounds hollow, whereas for

larger delays (30 ms) we perceive a clear, distinct echo.When the delay is large (200 ms) and the level is high,we experience difficulty in talking. This effect is illus-trated in Fig. 1 when a normal telephone handset is usedin the feedback loop [14].

Fig. 1 shows the quality judgment of a telephone link,using a five-point scale (5 excellent, . . , 1 bad), whensubjects are asked to rate the conversational quality, whichincludes the quality with which we perceive our ownvoice, for different levels of echo as a function of delay.The level of the echo is expressed in the talker echo loud-


On the Quality of Hearing One’s Own Voice*

RONALD APPEL1 AND JOHN G. BEERENDS,2 AES Member

1Laboratory of Acoustical Imaging and Sound Control, Delft University of Technology, 2600 GA Delft, The Netherlands2KPN Research, 2260 AK Leidschendam, The Netherlands

The way we perceive our own voice is being studied. Contrary to classical speechlistening-quality experiments, where subjects judge the speech quality by listening,remaining silent themselves, in talking-quality experiments subjects judge the quality withwhich they perceive their own voices while actively speaking. In this way the self-listeningcomfort is measured. Six experiments were carried out. Five used a standard telephone setupwhere echo and distortions were introduced and judged by subjects on the disturbance. In oneexperiment subjects judged the talking quality of six different rooms. The subjective resultswere used to develop an objective perceptual talking-quality measure. The overall correlationbetween the subjectively perceived quality and the objectively measured quality was 0.97.

* Manuscript received 2000 November 20; revised 2001October 22.

Fig. 1. Example of how delay and echo loudness are related toperceived conversational quality using a standard telephone-band (300–3400-Hz) handset. A five-point mean opinion score(MOS) scale is used, representing the average opinion over alarge set of subjects. The level of the echo is expressed in thetalker echo loudness rating (TELR), a subjectively determinedlevel of attenuation in decibels, between the point 25 mm in frontof the talker’s mouth and the entrance of the talker’s ear.

APPEL AND BEERENDS PAPERS

ness rating (TELR), a subjectively determined level ofattenuation, in decibels, between a point 25 mm in front ofthe talker’s mouth and the entrance of the talker’s ear. Thisecho is heard superimposed on the natural sidetone, whichhas a loudness attenuation of approximately 15 dB. Forloud echo levels (low TELR) a delay of 10 ms alreadyleads to a lower quality score, although the echo cannotyet be perceived separately from the direct signal. For lowlevels of echo (high TELR) a delay up to 300 ms hardlygives any degradation.

Delay and echo play increasing roles in the quality oftelephone services because modern wireless and packet-based network techniques, such as GSM (Global Systemfor Mobile Communication), UMTS (Universal MobileTelecommunications System), and VoIP (Voice overInternet Protocol), inherently introduce more delay due tocoding and interleaving than the classical circuit switchingnetwork techniques, which do not use packetization.

Delay and echo together with the sidetone (the directpath from mouth to ear) are important factors in how weperceive our own voice in a telephone link [15]. Thispaper focuses on developing a general method for theobjective measurement of the quality with which we per-ceive our own voice, defined as the talking quality. Notethat here one cannot use the concept of intelligibilitybecause the speaker knows what he or she is saying; onecan only say something about the quality as perceivedby ourselves. It can be interpreted as the self-listeningcomfort.

In a set of earlier papers objective measurement meth-ods were developed for listening quality, the quality withwhich we perceive the voice of another person. Themethod for speech listening quality, called perceptualspeech quality measure (PSQM) [6], was benchmarked bythe International Telecommunications Union, where itshowed superior performance [16]. It was standardized bythe ITU [17] and revised [11] in order to cope with a widerset of distortions, such as dynamic time warping and lin-ear filtering. This revised method, called perceptual eval-uation of speech quality (PESQ), was developed jointlywith British Telecom [10] and accepted as an ITU recom-mendation [11]. In both PSQM and PESQ the listeningquality is predicted by calculating the difference in theinternal representation of the reference and the degradedspeech signal (see Fig. 2). In listening-quality experimentsthe signal inputted into the system under test can be usedas an ideal reference and the system’s output is used as thedegraded signal. In talking quality it is not clear what thereference signal is because we perceive our own voice dif-ferently from the voice of someone talking to us. In thispaper we will develop a universal method for assessingthis talking quality based on human perception. Thismethod is in principle capable of taking into account anytype of degradation in talking quality. The PSQM/PESQmodels are taken as the starting point in the developmentof this objective talking-quality model.

Most of the subjective talking-quality experiments inthis paper were conducted in a telephony context, where


(a)

(b)

Fig. 2. Overview of approach taken in development of objective listening-quality model. (a) Subjects normally compare the degradedoutput of a system with the input (reference) to the system. (b) Objective model takes the same approach by mapping input (reference)and output to internal representations and then using the difference as the basis for measuring the perceived listening quality.

PAPERS HEARING ONE’S OWN VOICE

subjects talked into a standard telephone handset and wereasked to judge the quality with which they perceived theirown voices. One experiment was carried out with subjectstalking normally in different acoustical environments, thusallowing to check whether the objective measurementmethod that is being developed can be used in a widerscope than only in telephony.

1 SUBJECTIVE TALKING-QUALITYEXPERIMENTS

1.1 IntroductionIn the development of the objective listening-quality

model a number of different subjectively scored speechdatabases were available, mainly from speech codec stan-dardization efforts for which the listening quality is of cru-cial importance. An overview of these databases is givenin [18]. However, for talking quality such databases werenot available. The main difficulty is that one has to talk tooneself in order to be able to judge the talking quality,making the subjective assessment and the recording of thesignals more cumbersome than is the case with listeningexperiments. In a listening-quality experiment one canmake recordings of the distortions and play back theserecordings to a large set of subjects. Databases for listen-ing quality are always constructed using this procedure[19]. If one would make recordings of talking-quality dis-tortions (such as echoes) and play these distortions back toa subject, the subject could only judge the listening qual-ity of these signals, and the masking effect that one’s ownvoice has on the perceived talking quality is lost. In thispaper all subjective experiments are thus carried out usingsubjects who are actively speaking.

In this study six different subjective talking-quality tests

were carried out using the guidelines of ITU-T P.800 [19].Eight untrained subjects were used. By averaging the indi-vidual opinion scores, a mean opinion score (MOS) wasobtained, which depends less on individual preferences.This subjective procedure results in a 95% confidenceinterval in the MOS scores of about 0.5 on a five-pointMOS scale.

Five experiments were carried out with a standard tele-phone handset using the degradation category rating(DCR) scale in the gathering of the subjective opinions.This scale has five quality categories, as given in Table 1.Subjects were asked to speak into a telephone handset in aquiet room with a background noise level below 30 dBA.The reverberation time in the room was below 0.4 s for allfrequencies above 250 Hz and below 0.8 s over the wholeaudio range (50–20 000 Hz). The room fulfills the IECnorm on listening tests for loudspeakers [20].

The first telephony experiment used a single echo. Thesecond one used sidetone distortion (you hear your ownvoice distorted) whereas in the third experiment subjectsjudged echo or sidetone within the same experiment. Thefourth experiment dealt with the effect of backgroundmasking noise on the perceived echo degradation. Thesefour experiments were used to develop and optimize thetalking-quality model. The fifth telephony experimentserved to validate the model. In this experiment subjectswere asked to judge the talking quality of telephone linksthat contained both echoes and sidetone distortion.

A final experiment was used to see whether the talking-quality model could be extended to predict quality outsidethe telephony context. In this experiment subjects wereasked to judge the talking quality of six different roomswhich had big differences in their reverberation times. Anabsolute category rating (ACR) scale, as given in Table 2,was used in the gathering of the subjective opinions. Thedifference in scale (ACR versus DCR) is due to the factthat for a telephony context the ideal case is no perceivedechoes, whereas for a room the ideal case contains smallamounts of reverberation. An anechoic room is not theideal surrounding for talking.

1.2 Single EchoThis first talking-quality experiment was carried out in

a telephony context. It used the most commonly experi-enced talking degradation in telephony, echo, where a sin-gle strong reflection of one’s own voice is returned toone’s own ear. A standard telephone handset, with a fre-quency range of between 300 and 3400 Hz as defined bythe ITU [21], was used. The echo was created by delayingand attenuating the talker’s voice and returning this signalto the loudspeaker in the handset.

Subjects were asked to give their opinions on a DCRscale, as given in Table 2. A DCR scale was used becausean explicit ideal can be defined as “inaudible echo,” and asingle discrete reflection will always lead to a degradation.The objective parameters in the test were the echo loud-ness (TELR between 19 and 43 dB) and the delay (be-tween 6 and 200 ms). This covers the subjective qualityrange from almost inaudible to loud and clearly audibleechoes. Table 3 gives the results of this talking-quality test.


Table 1. Degradation category rating (DCR)scale used in subjective tests. The averageover a set of subjects is called DCR mean

opinion score (DCR MOS).

Category DCR Score

No disturbance audible 5Audible, not annoying 4Slightly annoying 3Annoying 2Very annoying 1

From ITU-T P.800 [19].

Table 2. Absolute category rating (ACR)scale used in subjective tests. The averageover a set of subjects is called ACR mean

opinion score (ACR MOS).

Category ACR Score

Excellent 5Good 4Fair 3Poor 2Bad 1

From ITU-T P.800 [19].


1.3 Sidetone DistortionIn the second experiment sidetone via a handset tele-

phone was artificially distorted by amplitude clipping.When a pure sine is sent through the system, the percent-age of nonlinear distortion is the power introduced athigher harmonics relative to the power at the base fre-quency. The level of clipping is expressed as a percentageof nonlinear distortion and ranges from 0 to 26%. Thiscovers the subjective quality range from almost inaudibleto loud and clearly audible distortions.

Eight persons were asked to speak into the telephonehandset using the same experimental procedure as with thesingle-echo experiment. Table 4 gives the DCR MOS resultsas a function of the percentage of nonlinear distortion.

1.4 Single Echo or Sidetone DistortionThe two previous telephony experiments can be com-

bined into a single subjective test, where subjects judgeecho and sidetone distortion in a single experiment. Thisexperiment allows to check whether a talking model that

gives high correlations on either echo or sidetone distor-tion gives the correct weighting to both when they are pre-sented in a single experiment. Here 11 settings of increas-ing clipping (0–28%) and 12 settings for different echodistortions (delay between 0 and 200 ms and TELR be-tween 19 and 43 dB) were used. Subjects were asked torate the quality with respect to echo and sidetone distor-tions. Table 5 gives the DCR MOS values along with thesettings for TELR, delay, and the percentage of nonlineardistortion of the sidetone.

1.5 Single Echo Masked by Background NoiseA key issue in talking-quality measurements is that

noise at the other party’s side and noise in the systemunder test will have an influence on the talking quality.This noise will of course also have a significant influenceon the listening quality where the speech of the other partyis judged. If we want to assess the influence of the noise


Table 3. Results of single-echo test.

TELR Echo Delay Talking Quality(dB) (ms) (DCR MOS)

19 6 4.819 12 4.319 20 3.719 28 3.019 46 3.319 60 1.619 80 1.519 120 1.322 8 4.922 16 3.822 24 4.122 40 3.222 70 1.822 100 1.722 140 1.125 10 5.025 18 4.425 24 4.325 36 3.925 55 2.925 86 2.025 120 1.725 156 1.331 30 4.831 46 4.131 64 3.831 90 2.931 126 2.231 140 2.231 170 1.739 50 4.639 70 4.439 100 3.739 150 3.339 190 2.639 200 2.543 80 4.443 120 4.243 160 3.743 200 3.6

Table 4. Results of sidetonedistortion test.

% Nonlinear Talking QualityDistortion (DCR MOS)

3 4.85 4.88 4.5

13 4.114 4.015 3.616 3.317 2.818 2.520 1.823 1.426 1.2

Table 5. Results of test with either a single echo or a distortedsidetone.

TELR Echo Delay % Nonlinear Talking Quality(dB) (ms) Distortion (DCR MOS)

0 0 5.0 0 1 4.8 0 2 4.7 0 2 4.6 0 3 4.6 0 7 4.6 0 12 4.2 0 18 3.5 0 22 3.4 0 26 3.3 0 28 2.319 12 0 4.319 46 0 2.319 120 0 1.122 40 0 2.122 70 0 1.425 10 0 2.925 86 0 1.425 120 0 1.431 46 0 3.331 126 0 1.839 150 0 3.243 160 0 3.8


on the talking quality, subjects have to neglect all influ-ences that are related to listening quality. In fact the influ-ence of noise is opposite in listening- and talking-qualityassessment.

In listening-quality assessment an increase in noise willlead to a decrease in perceived quality. In the extreme casethe speech from the other party in the conversation willbecome unintelligible. In talking-quality assessment, how-ever, noise from the other side will mask the echo of one’sown voice. Thus louder noise can improve the talkingquality. Noise levels are kept at levels where the Lombardeffect [12], [13] plays no role. The range represents back-ground noise levels as found in normal telephonysituations.

To be able to check whether subjects can assess inde-pendently the effect of echo on the listening and talkingquality, two different subjective tests were carried out. Inthe first experiment subjects were asked to rate the listen-ing quality. They were offered a speech sample of two sen-tences distorted by Hoth noise [22] of variable levels, gen-erated in the telephone link. They were asked for theirlistening-quality opinions with respect to backgroundnoise using the DCR scale of Table 2.

In the second experiment subjects were asked to speaktwo sentences into the telephone link, distorted by echoand background Hoth noise of the same levels used in thelistening-quality experiment. The subjects were asked togive their opinion about the talking quality of the tele-phone link with respect to echo, ignoring the backgroundnoise. Table 6 presents the DCR MOS values for theexperiments, along with the settings for noise, TELR, anddelay.

To test the orthogonality of the talking and listeningscores a scatter plot was made (Fig. 3) and the correlationbetween the two calculated. The results show that subjectsare indeed able to judge talking and listening quality inde-pendently (the correlation is less than 0.1).

1.6 Single Echo and Sidetone DistortionThe last experiment used echo signals that were

returned to the talker’s ear with the same time distortion inthe sidetone signal. In Table 7 the DCR MOS values aregiven along with the settings for TELR, delay, and the per-centage of nonlinear distortion of the echo. This final testwas used to validate the model outside the scope of thedistortions used to optimize the model.


Table 6. Results of test with a single echo masked by background noise.

Noise TELR Echo Delay Listening Quality Talking Quality(dB SPL) (dB) (ms) (DCR MOS) (DCR MOS)

28 0 4.9 5.028 20 100 4.9 1.750 0 1.6 4.950 20 120 1.6 2.329 48 200 4.8 4.229 26 36 4.8 3.329 23 70 4.8 2.329 32 64 4.8 3.238 44 64 3.4 3.638 39 150 3.4 3.938 23 70 3.4 4.238 26 156 3.4 2.432 26 55 4.4 2.932 23 100 4.4 2.332 39 150 4.4 3.932 32 170 4.4 3.245 39 100 2.6 4.345 39 100 2.6 4.345 26 86 2.6 3.245 23 40 2.6 3.345 23 140 2.6 2.745 32 170 2.6 3.734 32 140 4.0 3.334 20 60 4.0 2.134 53 160 4.0 4.934 39 70 4.0 4.328 53 200 4.9 4.6

Fig. 3. Relation between talking and listening MOS values.Calculated correlation is 0.06, showing that talking quality andlistening quality vary independently if degradation is caused bybackground noise in return path.


1.7 Acoustical Reverberation in Rooms

The last experiment used the most natural type of talk-ing degradations, that is, reverberations in a normal room.Subjects were asked to judge the talking quality of six dif-ferent rooms ranging from extremely “dead” (anechoicroom) to extremely lively (reverberation room). The mid-band (500–1000-Hz octave band) reverberation timesranged from 0 to 7.5 s (see Table 3).

As an introduction each subject was guided through theanechoic room and the reverberation room, and the pur-pose of the test was explained. In each room the subjectwas asked to give an opinion about the talking quality onthe ACR scale given in Table 2. To do so the subject wasgiven two phoneme-rich sentences in Dutch and wasasked to speak them in a normal conversational voice. TheACR scale was chosen because the ideal talking room wasnot known and consequently a DCR scale could not beused. No restraints were placed on the evaluation time.After introduction to the two most extreme acoustic envi-ronments and an explanation of the test, the subject wasguided through each of the six rooms in random order. TheACR MOS results as a function of the average reverbera-tion time are given in Table 8.

Previous studies on listening quality [23], [24] haveshown that reverberation time is the most important factorin the perception of acoustical quality, but a number ofother acoustic criteria, such as initial time gap and otherfactors concerning spatialness, also have an influence.These criteria were not measured in this study.

Table 8 shows that a speaker finds it increasinglyannoying to speak in an acoustic environment in which thereverberation of the speaker’s own voice is noticeable fora longer period of time. But the total lack of reverberations

causes also a decrease in talking quality. This supports theidea that subjects base their quality evaluation of how theyperceive their own voice on the difference with an idealcondition. Even when subjects are not explicitly offeredthis ideal condition, they will base their judgment of thecondition under test on the difference in perception withan implicitly assumed ideal acoustic environment. Theresults suggest that this “ideal talking room” has a rever-beration time of between 0.1 and 1 s.

2 OBJECTIVE MODEL

2.1 IntroductionThe aim of the objective talking-quality model was to

predict the subjectively perceived talking quality. In thedevelopment of such a model it is therefore important touse recordings that are made under the same conditions asthe subjective experiment. A fundamental problem is,which signals have to be constructed as the correct repre-sentations of the reference and the degraded signal. This isnot a problem in a subjective listening experiment, wheretwo recordings are offered to a group of subjects. Onerecording is the reference, the other the distorted speechfragment. Every subject is offered the exact same record-ings. In a subjective talking experiment this is impossible,due to the fact that every subject has to use his or her ownvoice to evaluate the talking quality, and each individualvoice has different characteristics. For instance, for a loudvoice the quality perception can be different than for a softvoice. When a person speaks with a very staccato voice,using very short intervals of sound followed by a silence,echo will not be masked by this subject’s own voice.

One could make recordings of the live signals that areproduced by the subjects themselves. However, for a prac-tical objective measurement system it is not desirable tohave to rely on a subject uttering a sentence. Such anobjective method would, for example, produce nonrepeat-able outcomes because a subject never produces exactlythe same acoustical output twice. In order to get an objec-tive, reproducible talking-quality measure one would haveto use the same voice recording in measuring the differentsystems and compare the results with those obtained fromreal live speaking talkers. In order to achieve this goal, twodifferent approaches are taken.

In the first approach a speech file is coupled into the sys-tem under test on an electrical level, and the signal returnedfrom the system is recorded. This approach can of course


Table 7. Results of test with distorted sidetone and single echo.

TELR Echo Delay % Nonlinear Talking Quality(dB) (ms) Distortion (DCR MOS)

43 160 0 4.225 86 0 2.622 24 0 3.9 0 0 4.943 160 9 3.639 50 9 3.431 90 9 2.822 24 9 3.719 120 9 1.6 0 9 4.143 160 14 2.839 50 14 3.331 90 14 2.925 86 14 2.122 24 14 2.94 120 14 1.4

0 14 3.728 160 20 2.224 50 20 2.316 90 20 1.710 86 20 1.87 24 20 2.64 120 20 1.1

0 20 2.3

Table 8. Average reverberation times and MOS over500- and 1000-Hz octave bands of 6 rooms used in

subjective experiment.

Reverberation Time Talking Quality(s) (ACR MOS)

Anechoic room 0.0 3.1Dry listening room 0.3 4.5Empty lecture hall 0.7 4.4Concrete staircase 2.5 2.3Concrete small hall 3.4 2.1Reverberation room 7.5 1.1


not be used in the room reverberation experiment, but it isvery attractive for telephone systems because it makes themeasurement system easy to implement.

In the second approach a so-called head-and-torso sim-ulator (HATS) is used to simulate the characteristics of alive speaking talker. This HATS looks like the upper partof a human body and has a loudspeaker inside its mouthand a microphone inside its ear. By using a speech file thathas been recorded without reverberation, played over themouth from the HATS, one can produce a natural voice ina reproducible manner. By placing the telephone handsetat the artificial head, a realistic simulation of a personholding a telephone is achieved, and recordings can bemade of the ear signals containing the sidetone and echosignals (Fig. 4). In this way recordings from the acousticalsignals at the ear of the HATS are made under the sameconditions as if the subject were talking. This approach isalways possible because the signals are coupled into thesystem under test on an acoustical level.

2.2 Objective Measurement MethodThe PSQM model as developed and standardized ear-

lier [6], [7] is in essence a listening-only model of a sub-ject using a telephone link. The basics are simple: speechis recorded at the input and the output of the telephonelink, and both subject and model determine the listeningquality on the basis of differences in internal representa-tions (see Fig. 2). In most cases the input signal can bedefined as the reference signal (the ideal), and both thesubject and the PSQM model base their judgments on thedifference in the internal representations of the output andinput (the ideal) signals. An essential part in the objectivePSQM assessment is that the input and output signals aretime aligned. If the alignment is not correct, the PSQMvalues are not related to the perceived subjective quality.In ITU-T P.861, which specifies the PSQM method, timealignment is not part of the algorithm. In the follow-up

standard, P.862, it is included.If we take the same approach in the development of a

model for talking quality, we have a problem. If we makerecordings of the signals that go into the system under testand those that are returned, an alignment procedure wouldlead to a zero distortion in the case of a single echo. Thereason for this is clear. In the PSQM approach thedegraded signal is composed of the original plus a distor-tion component, whereas in the talking-quality test therecordings of the return channel only contain the distor-tion component. The simplest solution to this problem isto construct a degraded signal by adding the speech of thereturn channel and the send channel and use this speechfile as the degraded file.

If we take the approach with the HATS, the signal at theentrance of the ear is a mix of the acoustical sidetone witha delay close to zero and the returned speech signal fromthe telephone system under test, which may contain echoor sidetone. In this case, however, we cannot use thespeech signal that was used as the input to the artificialmouth of the HATS as a reference. Recordings made at theentrance of the ear sound completely different from theelectrical representation of this speech signal and alsofrom recordings made at the entrance of the mouth.Studies have shown that below 1000 Hz, sound pressurelevels at a talker’s ear appear to be about 15 dB higherthan those at 1.5 m, a normal listening distance, whereasabove 1000 Hz the head progressively shields the talker’sear and the level differences between talker’s and lis-tener’s ears diminish at a rate of about 6 dB per octave[25]. This, together with bone conduction, is the case whypeople hearing their own recorded voice played back tothem tend to think they sound strange. This sidetone fil-tering effect, which occurs between mouth and ear, can betaken into account by modifying the reference speech sig-nal or by recording an ideal reference signal at theentrance of the ear at an acoustical level.


Fig. 4. Basic measurement setup used in development of objective talking-quality measure. Subjective testing is simulated by HATS,which generates a natural voice input into the telephone link. Instead of comparing the distorted sidetone and echo signal with thespeech input file, it is compared to a recording of a “clean” reference sidetone, thus including filtering that occurs between mouth andear. A computer model of the subject is used to compare the signal at the ear of the HATS under clean conditions with that of the tele-phone link under test.


Although HATS recordings at the acoustical level arethe best way to assess talking quality, one should realizethat a recording at the entrance of the ear is still not givingthe correct sidetone as perceived by the subject himself.The last missing piece of the signal is the sidetone that isgenerated by the internal path from mouth to ear, via boneconduction. During the development of the objectivetalking-quality measure it turned out that this bone pathneed not be modeled in order to get high correlationsbetween subjective and objective scores.

The HATS approach is taken in all experiments descri-bed in Section 2 except for the experiment where echoesare used with background masking noise. For that experi-ment the return signal was recorded on the electrical inter-face just before the loudspeaker whereas speech was putinto the system just after the microphone. In this case theecho signal is recorded separately, and one has to mix thedirect speech input signal with the returned echo signal tosimulate the sidetone signal perceived by the subject. Thissetup was used to check whether the same model can beused to predict the talking quality without using a HATSin the measurement setup.

For the experiment that used the different acousticalenvironments as the experimental factor to be judged onehad to define an ideal reference signal in the same way asgiven in Fig. 4. It was found that by using the dry listen-

ing room as a starting point, the ideal reference could beconstructed by adding a little reverberation to the record-ing. Apparently subjects have maximum appreciation for aroom that has a certain amount of reverberation, and theyuse that as a reference on which to base their opinions forother environments.

In the next section an overview of the talking-qualityalgorithm is given. The basic elements from the psycho-acoustic model used are taken from PSQM [6], [17]. Itshould be noted that PSQM was designed for narrow-bandspeech signals (300–3400-Hz telephone band) while talk-ing quality analysis includes natural, wide-band voice sig-nals because of the acoustical coupling and bone conduc-tion. Therefore some of the ideas used in the objectivemeasurement of wide-band audio signals are also used inthis paper [26].

2.3 Talking Quality, the Perceptual Echo andSidetone Quality Measure (PESQM)

The final talking-quality model is called the perceptualecho and sidetone quality measure (PESQM). An over-view of the algorithm is given in Fig. 5. The starting pointis a time to time–frequency mapping implemented by ashort-term fast Fourier transform (FFT) with a windowsize of 32 ms. The overlap between successive time win-dows (frames) is 50%. Phase information within a single


Fig. 5. Schematic representation of PESQM talking-quality algorithm.


Hann window (index i) is discarded in PESQM, and allcalculations are based on only the power representationsper frame Pxi[k] and Pyi[k] of the input signals x[t] andcoded signal y[t].

The linear frequency scale (index k) is mapped to a psy-chophysical pitch-like scale (index j) by binning FFTbands and summing the corresponding powers of the FFTbands with a proper normalization of the summed parts.The resulting pitch power densities Pxi[ j ] and Pyi[ j ] arescaled to a fixed level calculated from the listening levelof 79 dB SPL equivalent [27]. Frequency smearing is thenapplied using an algorithm given in [26]. Only the upwardspread of masking (the target frequency is higher than themasker frequency), which is strongly level dependent, isimplemented:

with ft the target frequency, fm the masker frequency, bothin Hz, and Pi[k] the level of the excitation at frequency fm.S has a maximum of 32 dB per critical band. Implement-ing a downward spread of masking gave no improvementin the correlation between subjective and objective scoresand was thus not included into the PESQM algorithm. Theexcitation power at pitch j, PE[ j ], resulting from powerPi[m] at pitch m, is given by

, .m mj P jPEPE 1010 ( )/)/mi

S 1010 j7 7A A

Using a nonlinear addition of the individual excitationcomponents, this leads to the following excitation atpitch j:

( .E j jPEPE( , )

.

minminmi i

j

j

1 4040

11 2525

.0 8)!

R

T

SSS

7 7

V

X

WWW

A A

This calculation is performed for the input x and the out-put y, and results in the excitation representations Exi[ j ]and Eyi[ j ].

From the excitation representations Exi[ j ] and Eyi[ j ],the sampled loudness densities Lxi[ j ] and Lyi[ j ] are cal-culated using a compression function given in Zwickerand Feldtkeller [28]:

.. .L j S

P j

P j

E j

0 50 5 0 5 1

. .

ii

l0

0 2323

0

0 2323R

T

SSS

R

T

SSS

77

7

7V

X

WWW

V

X

WWW

AA

A

A

Z

[

\

]]

]]

_

`

a

bb

bb

with P0( f ) the absolute threshold and Sl a loudness scal-ing factor calculated from a calibration sinusoid of 40 dBSPL that maps to 1 sone. Negative values of the loudnessdensities Lxi[ j ] and Lyi[ j ] are set to zero.

The sampled noise disturbance density Ni[ j ] in pitchband j and frame i is computed as the absolute differencebetween Lxi[ j ] and Lyi[ j ]:

Ni[ j ] |Lyi[ j ] Lxi[ j ]| 0.01

where 0.01 sone represents the internal noise. If, becauseof the latter, Ni[ j ] becomes negative, than Ni[ j ] is set

equal to zero.The next step in the PESQM model deals with the influ-

ence of masking background noise. In listening-qualityassessment an increase in noise will lead to a decrease inperceived quality. In the extreme case the speech from theother party in the conversation will become unintelligible.In talking-quality assessment, however, noise from theother side will mask the echo of one’s own voice. Thuslouder noise improves the talking quality. It turned out thatdesigning an algorithm for the separation of echo andnoise was rather complicated, but that a simple pragmaticapproach can give excellent results. The key idea is thatbecause subjects will always have silent intervals in theirspeech, the minimum loudness of the distorted signal overtime is almost completely caused by the backgroundnoise. This minimum can then be used to set a thresholdfor which all frames that have a loudness below thisthreshold are set to zero. If this minimum is smaller than0.5 sone, the threshold value is set to 0.5 sone. The resultis a disturbance density function Nthri[ j ].

In order to get high correlations between subjective andobjective results, a local noise loudness density to signalloudness ratio had to be used instead of the plain noiseloudness Nthri[ j ]/Lyi. Furthermore it was noted that sub-jects put a larger weight on loud local disturbances, whichapparently determined the overall quality. Such an empha-sis on loud disturbances can be obtained by using an Lpnorm, equivalent to the nonlinear addition used in the cal-culation of the excitation. By not simply adding the noiseloudness in frames i, but first raising them to a certainpower ( p 1) and then taking the inverse power of thesum, the emphasis is shifted to the samples with larger dis-turbances [2]. The same approach is taken over the pitchscale:

.N LyLy

jNSRNSR

NthrNthr1

. .

ii

i

j

N

b

1 4

1

11 4

b

!J

L

KK

N

P

OO

R

T

SSSS

7

V

X

WWWW

A

To obtain the final PESQM score that has a high correla-tion with the subjectively perceived talking quality onehas to aggregate the NSRi over all frames using p 5,emphasizing loud parts:

.N

PESQMPESQM NSRNSR1

ii

N5

1

15

!J

L

KK

N

P

OO

To estimate the subjective quality on quality scales such asMOS, the PESQM value can be transformed using a non-linear regression.

3 RESULTS AND CONCLUSIONS

The two free PESQM model parameters, the powers Lpand Lq, with which the averaging over frequency and timeis carried out, were optimized on four of the six talking-quality databases. The optimum values Lp 1.4 and Lq 4 were used in the validation of the model on the com-bined echo/sidetone test and on the test that used naturalacoustic reverberation as the talking-quality degradation.


.,/ s

f

P jf f

HzHz

dBdBdBdB criticalcritical bandband2222

230230 0 2

i

mt m

7 A


The result for the echo/sidetone test is given in Fig. 6 andshows an excellent correlation between subjective andobjective results, r 0.93. With the acoustical reverbera-tion database results were even better (r 0.98).

Reoptimizing the model over all databases gave Lp 1.4 and Lq 5, only a slight improvement in the overallcorrelation. Results over all six databases after reopti-mization gave an average correlation of 0.97 (Fig. 7).

The most important conclusion is that for the six data-bases that contain different types of echo and sidetone dis-tortions a single objective talking-quality measure can beused to predict the subjectively perceived talking degrada-tions. Two different recording techniques can be used. If

the speech send signal is recorded separately from thedegraded return signal, a combination of these two canserve as the degraded signal for the measurement system.In this case the speech send signal can serve as a refer-ence. If recordings are made using a HATS, one can usethe undistorted sidetone signal as a reference to thedegraded return signal. In this case the combination ofsend speech signal and degraded return signal automati-cally takes place in the acoustic domain.

Using the PSQM P.861 listening-quality measure as astarting point, a number of modifications are needed totransform it into a talking-quality measure, of which themost important ones are the replacement of the linear


Fig. 6. Validation of model on combined echo and sidetone test.

Fig. 7. Final results of PESQM objective talker-quality measure for all six databases. Third-order regression functions are fitted toexperimental context in order to be able to deal with different use of MOS scale by subjects in each experiment.


time–frequency summation by an Lp norm over time andfrequency and the replacement of the calculation of thelocal disturbance by a disturbance to signal ratio.

4 ACKNOWLEDGMENT

The authors would like to thank two anonymousreviewers for their valuable comments and S. L. C. Lagen-dijk from the Delft University of Technology for coiningthe term self-listening comfort.

5 REFERENCES

[1] H. J. M. Steeneken and T. Houtgast, “A PhysicalMethod for Measuring Speech-Transmission Quality,” J.Acoust. Soc. Am., vol. 67, pp. 318–326 (1980 Jan.).

[2] S. R. Quackenbush, T. P. Barnwell III, and M. A.Clements, Objective Measures of Speech Quality(Prentice-Hall Advanced Reference Ser., EnglewoodCliffs, NJ, 1988).

[3] S. Wang, A. Sekey, and A. Gersho, “An ObjectiveMeasure for Predicting Subjective Quality of SpeechCoders,” IEEE J. Sel. Areas in Comm., vol. 10, pp. 819–829 (1992 June).

[4] S. Hayashi and N. Kitawaki, “An Objective QualityAssessment Method for Bit-Reduction Coding ofWideband Speech,” J. Acoust. Soc. Am., vol. 92, pp.106–113 (1992 July).

[5] O. Ghitza, “Auditory Models and Human Perfor-mance in Tasks Related to Speech Coding and SpeechRecognition,” IEEE Trans. Speech Audio Process., vol. 2,pp. 115–132 (1994 Jan.).

[6] J. G. Beerends and J. A. Stemerdink, “A PerceptualSpeech-Quality Measure Based on a PsychoacousticSound Representation,” J. Audio Eng. Soc., vol. 42, pp.115–123 (1994 Mar.).

[7] S. Voran, “Objective Estimation of PerceivedSpeech Quality––Part I: Development of the MeasuringNormalizing Block Technique,” IEEE Trans. SpeechAudio Process., vol. 7, pp. 371–382 (1999 July).

[8] S. Voran, “Objective Estimation of PerceivedSpeech Quality––Part II: Evaluation of the MeasuringNormalizing Block Technique,” IEEE Trans. SpeechAudio Process., vol. 7, pp. 383–390 (1999 July).

[9] M. Hansen and B. Kollmeier, “Objective Modelingof Speech Quality with a Psychoacoustically ValidatedAuditory Model,” J. Audio Eng. Soc., vol. 48, pp. 395–409 (2000 May).

[10] A. W. Rix, J. G. Beerends, M. P. Hollier, andA. P. Hekstra, “Perceptual Evaluation of SpeechQuality (PESQM)––A New Method for Speech QualityAssessment of Telephone Networks and Codecs,” pre-sented at the ICASSP, Salt Lake City, UT, USA, 2001May 7–11.

[11] ITU-T Rec. P.862, “Perceptual Evaluation ofSpeech Quality (PESQ), an Objective Method for End-to-End Speech Quality Assessment of Narrowband Tele-phone Networks and Speech Codecs,” InternationalTelecommunications Union, Geneva, Switzerland (2001Feb.).

[12] E. Lombard, “Le signe de l’élevation de la voix,”Ann. Maladies Oreille, Larynx, Nez, Pharynx, vol. 37, pp.101–119 (1911).

[13] H. Lane and B. Tranel, “The Lombard Sign and theRole of Hearing in Speech,” J. Speech Hearing Res., vol.14, pp. 677–709 (1971).

[14] ITU-T Rec. G.107, “The E-Model, a Computa-tional Model for Use in Transmission Planning,” Inter-national Telecommunications Union, Geneva, Switzerland(1998 Dec.).

[15] D. L. Richards, Telecommunication by Speech: TheTransmission Performance of Telephone Networks (But-terworths, London, 1973).

[16] ITU-T Study Group 12, “Review of ValidationTests for Objective Speech Quality Measures,” Doc. COM12-74, International Telecommunications Union, Geneva,Switzerland (1996 Mar.).

[17] ITU-T Rec. P.861, “Objective Quality Measure-ment of Telephoneband (300–3400 Hz) Speech Codes,”International Telecommunications Union, Geneva, Swit-zerland (1996 Aug.).

[18] ITU-T Study Group 12, “Report of the Question13/12 Rapporteur’s Meeting, Solothurn, Switzerland,”Doc. COM 12-117, International TelecommunicationsUnion, Geneva, Switzerland (2000 Mar.).

[19] ITU-T Rec. P.800, “Methods for SubjectiveDetermination of Transmission Quality,” InternationalTelecommunications Union, Geneva, Switzerland (1996Aug.).

[20] IEC 268-13, “Sound System Equipment, Part 13:Listening Tests on Loudspeakers,” International Electro-technical Commission, Geneva, Switzerland (1985).

[21] ITU-T Rec. P.48, “Specification for an Intermedi-ate Reference System,” International TelecommunicationsUnion, Geneva, Switzerland (1989).

[22] D. F. Hoth, “Room Noise Spectra at Subscribers’Telephone Locations,” J. Acoust. Soc. Am., vol. 12, pp.499–504 (1941 Apr.).

[23] P. Lehmann and H. Wilkens, “Zusammenhang sub-jectiver Beurteilungen von Konzertsälen mit raumakustis-chen Kriterien,” Acustica, vol. 45, pp. 256–268 (1980).

[24] O. Warusfel and J. P. Jullien, “Subjective Vali-dation of an Objective Model for the Characterization ofRoom Acoustic Quality,” presented at the 82nd Con-vention of the Audio Engineering Society, J. Audio Eng.Soc. (Abstracts), vol. 35, p. 388 (1987 May), preprint2457.

[25] J. M. Festen, “Speech Levels at the Ear of a Talkerand Their Implication for a Hearing-Impaired Person Tak-ing Part in Discussion,” J. Acoust. Soc. Am., vol. 87, suppl.1 (1990 Spring).

[26] J. G. Beerends and J. A. Stemerdink, “A PerceptualAudio Quality Measure Based on a Psychoacoustic SoundRepresentation,” J. Audio Eng. Soc., vol. 40, pp. 963–978(1992 Dec.).

[27] ITU-T, Handbook on Telephonometry (Interna-tional Telecommunication Union, Geneva, Switzerland,1992).

[28] E. Zwicker and R. Feldtkeller, Das Ohr als Nach-richtenempfänger (Hirzel, Stuttgart, Germany, 1967).




THE AUTHORS

Ronald Appel was born in Marken, The Netherlands, in1974. He studied applied physics at Delft University ofTechnology, where he completed his specialization at theLaboratory of Acoustical Imaging and Sound Control. Hisgraduation research was conducted at KPN Research inLeidschendam, where he worked on objective loudnessmeasurement and speech quality assessment. The mainresults of the work are presented in this paper, and theyhave led to a patent application by KPN. He received anM.Sc. degree in 2001.

John G. Beerends was born in Millicent, Australia, in1954. He received a degree in electrical engineering fromthe HTS (Polytechnic Institute), The Hague, TheNetherlands, in 1975. After working in industry for threeyears he studied physics and mathematics at theUniversity of Leiden, from which he received an M.Sc.degree in 1984. In 1983 he was awarded a prize of Dfl45000 by Job Creation for an innovative idea in the fieldof electroacoustics.

From 1984 to 1989 he worked at the Institute forPerception Research, where he received a Ph.D. from theTechnical University of Eindhoven in 1989. The main partof his doctoral work, which deals with pitch perception,was published in the Journal of the Acoustical Society ofAmerica. The results of this work led to a patent on a pitchmeter by the N. V. Philips Gloeilampenfabriek.

From 1986 to 1988 Dr. Beerends work on a psycho-acoustically optimized loudspeaker system for the Dutchloudspeaker manufacturer BNS. The system was intro-duced at the Dutch consumer exhibition FIRATO in 1988.

In 1989 he joined the KPN Research Laboratory inLeidschendam, where he worked on audio and video qual-

ity assessmernt, audio-visual interaction, and audio cod-ing (speech and music). The work on audio quality, car-ried out together with Jan Stemerdink, led to severalpatents and two measurement methods for objective, per-ceptual assessment of audio quality. The first methoddeals with the quality of telephone-band speech codecsand was standardized within ITU-T in 1996 asRecommendation P.861 (PSQM: perceptual speech qual-ity measure). The second method deals with the quality ofmusic codecs and was integrated into ITU-RRecommendation BS.1387 (1999, PEAQ: perceptualevaluation of audio quality). Most of the work on audioquality (speech, music, and audio-visual interaction) waspublished by the Audio Engineering Society and the ITU.

From 1996 to 2001 he worked with Andries Hekstra onthe objective measurement of the quality of video andspeech. The work on video quality led to several patentsand a measurement method for objective, perceptualassessment of video quality. This method was submittedto the ITU, and it showed highest overall correlations withsubjective measurements in a comparison of ten differentobjective video quality assessment methods. The work onspeech quality, partly carried out with researchers fromBritish Telecom, focused on improvements of the PSQMmethod to make it suitable for use in modern voice overIP, ATM, and mobile applications. The new method,called PESQ (perceptual evaluation of speech quality),has been standardized as ITU-T Recommendation P.862.

Dr. Beerends latest work focuses on the objectivemeasurement of speech quality in the acoustic domain andon the measurement of conversational quality in whichnot only listening quality plays a role but also talkingquality (self-listening comfort in terms of sidetone andecho).

R. Appel J. G. Beerends

PAPERS

0 INTRODUCTION

Under the assumption of source and receiver immobil-ity, the acoustical space in which they are placed can beconsidered a linear time-invariant system characterized byan impulse response h(t). In room acoustics the accuratemeasurement of the impulse response is very important,since many acoustical parameters can be derived from it.Moreover, in present-day audio applications (that is, vir-tual reality, auralization, spatialization of sounds) theimportance of measuring binaural room impulse respon-ses with a very high signal-to-noise ratio becomes moreand more evident. Once the impulse response has beenmeasured precisely, it can be integrated in a completeauralization process [1], [2]. In order to achieve the bestquality for this auralization process, the measured impulseresponse must reach a very good signal-to-noise ratio(more than 80 dB if possible).

A common method for measuring the impulse responseof such an acoustical system is to apply a known inputsignal and to measure the system output. The choice con-cerning the excitation signal and the deconvolution tech-nique that will permit obtaining the impulse response fromthe measured output is of essential importance:

• The emitted signal must be perfectly reproducible.• The excitation signal and the deconvolution technique

have to maximize the signal-to-noise ratio of the decon-volved impulse response.

• The excitation signal and the deconvolution techniquemust enable the elimination of nonlinear artifacts in thedeconvolved impulse response.

In general the signal-to-noise ratio is improved by tak-ing multiple averages of the measured output signal beforethe impulse response deconvolution process is started.

The most commonly used excitation signals are deter-ministic, wide-band signals known as:

• MLS (maximum-length sequence) and IRS (inverserepeated sequence), which use pseudorandom whitenoise

• Time-stretched pulses and SineSweep, which use time-varying frequency signals.

1 BRIEF DESCRIPTION OF MEASUREMENTTECHNIQUES

The acoustical impulse response measurements usingthe MLS technique were first proposed by Schroeder in1979 [3] and have been used for more than 20 years. Manypapers discussed the theoretical and practical advantagesand disadvantages of their technique [4]–[15]. Shortlyafter the publication of Schroeder, the IRS technique wasproposed as an alternative, allowing a theoretical reduc-tion of the distortion artifacts introduced by the MLS tech-nique [4], [16], [17].

Two years after the proposition of Schroeder, Aoshimaintroduced a new idea for the measurements of impulseresponses which led to the time-stretched pulses technique[18]. His idea was then pushed further by Suzuki et al.


Comparison of Different Impulse ResponseMeasurement Techniques*

GUY-BART STAN, JEAN-JACQUES EMBRECHTS, AES Member, AND DOMINIQUE ARCHAMBEAU

Sound and Image Department, University of Liège, Institut Montefiore B28, B-4000 Liège 1, Belgium

The impulse response of an acoustical space or transducer is one of its most importantcharacterizations. In order to perform the measurement of their impulse responses, four of themost suitable methods are compared: MLS (maximum-length sequence), IRS (inverserepeated sequence), time-stretched pulses, and SineSweep. These methods have already beendescribed in the literature. Nevertheless, the choice of one of them depending on the measure-ment conditions is critical. Therefore an extensive comparison has been realized. Thiscomparison was done through the implementation and realization of a complete, fast, reliable,and cheap measurement system. Finally, a conclusion for the use of each method accordingto the principal measurement conditions is presented. It is shown that in the presence ofnonwhite noise, the MLS and IRS techniques seem to be more accurate. On the contrary, inquiet environments the logarithmic SineSweep method seems to be the most appropriate.

* Manuscript received 2000 December 28; revised 2001October 19.

STAN ET AL. PAPERS

[19], who proposed what they called an “optimumcomputer-generated pulse signal.”

Recently Farina and Ugolotti introduced the logarith-mic SineSweep technique [20], [21] intended to overcomemost of the limitations encountered in the other tech-niques. The idea of using a sweep in order to deconvolvethe impulse response is not new [22], but the deconvolu-tion method used is different in Farina.

These techniques have already been described and dis-cussed in many papers. However, it is intended here tofocus on some important properties which are necessary tounderstand the comparison of the different methods.

1.1 MLS TechniqueThe MLS technique is based on the excitation of the

acoustical space by a periodic pseudorandom signal hav-ing almost the same stochastic properties as a pure whitenoise. The number of samples of one period of an mth-order MLS signal is L 2m 1.

More theoretical considerations about the MLS seq-uences can be found in [7], [10], [12], [13], [23] and in theexcellent book on shift-register theory [24].

With the MLS technique, the impulse response isobtained by circular crosscorrelation (as shown in [25])between the measured output and the determined input(MLS sequence). Because of the use of circular operationsto deconvolve the impulse response, the MLS techniquedelivers the periodic impulse response h[n], which isrelated to the linear impulse response as

[ ] [ ] .h n h n lLlL t

3

3

! (1)

Eq. (1) reflects the well-known problem of the MLS tech-nique: the time-aliasing error. This error is significant ifthe length L of one period is shorter than the length of theimpulse response to be measured. Therefore the order m ofthe MLS sequence must be high enough to overcome thetime-aliasing error. Our measurement system allows thegeneration of MLS sequences up to order 19 (which cor-responds to a period of 12 s if the sampling frequency is44.1 kHz).

1.1.1 MLS Immunity to Signals Not Correlatedwith the Excitation Signal

Each MLS sequence is characterized by a phase spectrumwhich is strongly erratic, with a uniform density of proba-bility in the [p, p] interval, as can be seen in Fig. 1.

According to this property, the MLS technique is ableto randomize the phase spectrum of any component of theoutput signal which is not correlated with the MLS inputsequence [5], [9]. As a consequence, any disturbing signal(that is, white or impulsive noise) will actually be phaserandomized, and this will lead to a uniform repartition ofthe disturbing effects along the deconvolved impulseresponse (Figs. 2 and 3) instead of localized noise contri-butions along the time axis. A postaveraging method canthen be used to reduce this uniformly distributed noiseappearing in the deconvolved impulse response.

1.1.2 Disadvantages of the MLS TechniqueThe major problem of the MLS method resides in the

appearance of distortion artifacts known as distortionpeaks [6]. These artifacts are more or less uniformly dis-tributed along the deconvolved impulse response. The ori-gin of the distortion peaks lies in the nonlinearities inher-ent in the measurement system and especially theloudspeaker.

These distortion artifacts introduce characteristic crack-ling noise when the measured impulse response is con-volved with some anechoic signal in order to realize theauralization process. These distortion peaks can be atten-uated by:

• The use of dedicated measurement methods (such as theinverse repeated sequence technique [4], [16].

• The optimization of some measurement parameters. Forexample, the amplitude of the excitation signal is, inpractice, a compromise between increasing distortions athigh levels and decreasing the signal-to-noise ratio at lowlevels. This optimization is very time consuming becauseof the practical difficulty of finding the optimum ampli-fication level. Moreover, this compromise level must bechosen carefully for each new measurement situation.


Fig. 1. Magnitude and phase spectra of MLS sequence. Phase spectrum was enlarged to clearly show its uniform random distribution.

PAPERS MEASUREMENT TECHNIQUES

Fig. 4 illustrates the quality of the results that can beobtained when particular care is taken in the optimizationof the parameters (mainly the output level) conditioningthe MLS (or IRS) input signal. It can be seen that the dis-tortion peaks are reduced significantly but not completelyremoved.

1.2 IRS TechniqueEach IRS sequence with a 2L sample period x[n] is

defined from the corresponding MLS sequence of periodL (MLS[n]) by the following relation:

[ ][ ], ,

[ ], ,

x n

n n n L

n n n L

MLSMLS eveneven

MLSMLS oddodd

0 2

0 2

* . (2)

The deconvolution process is exactly the same as for theMLS technique (circular correlation).

Fig. 5 shows the attenuation of the distortion peaks

when the IRS method is used. These impulse responseshave been obtained by performing the measurements in ananechoic room, leaving all measurement parametersunchanged from one measurement to the next.

1.3 Time-Stretched Pulses TechniqueThis method is based on a time expansion and com-

pression technique of an impulsive signal [18]. The aim ofusing an expansion process for the excitation signal is toincrease the amount of sound power emitted for a fixedmagnitude of the signal and therefore to increase thesignal-to-noise ratio without increasing the nonlinearitiesintroduced by the measurement system. Once the responseto this “stretched” signal has been measured, a compres-sion filter is used to compensate for the stretching effectsinduced and to obtain the deconvolved impulse response.

Fig. 6 shows the impulse response obtained with this


Fig. 3. (a) Impulse response obtained with single MLS sequence of order 16 in anechoic room when impulsive noise is simultaneouslypresent. (b) Zoom on magnitude scale.

Fig. 2. (a) Impulse response obtained with single MLS sequence of order 18 in classroom, when a white-noise generator is simultane-ously present. Noise level measured at microphone is 60 dB. (b) Zoom on end of impulse response.

(a) (b)

(a) (b)

STAN ET AL. PAPERS

technique. The magnitude scale has been enlarged toclearly illustrate the absence of distortion peaks. However,this does not mean that distortion artifacts are completelyremoved. They still appear in the impulse response (as aresidue of the deconvolution filter), as can be seen inFig. 7.

1.4 Logarithmic SineSweep TechniqueThe MLS, IRS, and time-stretched pulses methods rely

on the assumption of linear time-invariant (LTI) systemsand cause distortion artifacts to appear in the deconvolvedimpulse response when this condition is not fulfilled.

The SineSweep technique developed by Farina [21]overcomes such limitations. It is based on the followingidea: by using an exponential time-growing frequencysweep it is possible simultaneously to deconvolve the lin-

ear impulse response of the system and to selectively sep-arate each impulse response corresponding to the har-monic distortion orders considered. The harmonic distor-tions appear prior to the linear impulse response.Therefore the measured linear impulse response is assuredexempt from any nonlinearity, and at the same time, themeasurement of the harmonic distortion at various orderscan be performed.

Fig. 8 illustrates the black box modeling of the meas-urement process common to all four techniques discussed.In this model it is assumed that the measurement system isintrinsically nonlinear, but on the other hand, perfect lin-earity is considered regarding the acoustical space fromwhich the impulse response is to be derived.

As pointed out in Farina [21], the signal emitted by theloudspeaker is composed of harmonic distortions (consid-


Fig. 5. Zoom on impulse responses. (a) MLS. (b) IRS.(a) (b)

Fig. 4. Zoom on impulse response obtained after optimization ofmeasurement parameters (output sound level), MLS technique.

Fig. 6. Zoom on impulse response in classroom when usingtime-stretched pulse of about 6 s.


ered here without memory) and may thus be representedby the following equation (see Fig. 8):

( ) ( ) ( ) ( ) ( )

( ) ( ) ( ) ( )

w t x t k t x t k t

x t k t x t k t

7 7

7 7

NN

12

2

33 g (3)

where ki(t) represents the ith component of the Volterrakernel [21], which takes into account the nonlinearities ofthe measurement system.

In practice it is relatively difficult to separate the linearpart (the reverberation part in the impulse response) fromthe nonlinear part (distortions). In the following we willconsider the response of the global system (the output sig-nal from the system represented in Fig. 8) as being com-posed of an additive Gaussian white noise n(t) and a set ofimpulse responses hi(t), each being convolved by a differ-ent power of the input signal:

( ) ( ) ( ) ( ) ( ) ( )

( ) ( ) ( ) ( )

y t n t x t h t x t h t

x t h t x t h t

7 7

7 7

NN

12

2

33 g (4)

where hi(t) ki(t) 7 h(t). Eq. (4) underlines the exis-tence of the nonlinearities at the system output.

In the case of the logarithmic SineSweep technique, the

excitation signal is generated on the basis of the followingequation (see [21] for more theoretical information):

( )( / )

[ ]sinsinlnln w w

wx t

Te 1 ( ( / )lnln w wt

2 1

12 1/T)

* 4 (5)

where w1 is the initial radian frequency and w2 is the finalradian frequency of the sweep of duration T.

Fig. 9 shows the time and spectral representations of alogarithmic sweep with initial and final frequencies at 10Hz and 1000 Hz, respectively.

The impulse response deconvolution process is realizedby linear convolution of the measured output with the ana-lytical inverse filter preprocessed from the excitation sig-nal. Using linear convolution allows time-aliasing prob-lems to be avoided. In fact, even if the time analysiswindow has the same length as the emitted SineSweepsignal (and is shorter than the impulse response to bemeasured), the tail of the system response may be lost, butthis will not introduce time-aliasing. This is a first advan-tage over MLS and IRS methods.

In practice a silence of sufficient duration is added atthe end of the SineSweep signal in order to recover the tailof the impulse response.

The deconvolution of the impulse response requires thecreation of an inverse filter f (t) able to “transform” the ini-


Fig. 8. Modeling of global system, including loudspeaker (considered a nonlinear element) and acoustical space (considered a perfectlylinear system).

Fig. 7. Zoom on impulse response in anechoic room when using time-stretched pulse of about 1 s. In this case a poor-quality loud-speaker was used to emphasize the nonlinearity of the measurement system.

STAN ET AL. PAPERS

tial sweep into a delayed Dirac delta function:

x(t) 7 f(t) d(t K ) . (6)

The deconvolution of the impulse response is then real-ized by linearly convolving the output of the measuredsystem y(t) with this inverse filter f(t):

h(t) y(t) 7 f(t) . (7)

The inverse filter f (t) is generated in the followingmanner:

1) The logarithmic sweep (which is a causal and stablesignal) is temporally reversed and then delayed in order toobtain a causal signal (the reversed signal is pulled back inthe positive region of the time axis). This time reversalcauses a sign inversion in the phase spectrum. As such theconvolution of this reversed version of the excitation sig-nal with the initial SineSweep will lead to a signal charac-

terized by a perfectly linear phase (corresponding to apure delay) but will introduce a squaring of the magnitudespectrum.

2) The magnitude spectrum of the resulting signal isthen divided by the square of the magnitude spectrum ofthe initial SineSweep signal.

The time and spectral representations of the inverse fil-ter corresponding to the SineSweep (Fig. 9) are shown inFig. 10. To minimize the influence of the transients intro-duced by the measurement system and appearing at thebeginning and end of the emission of the excitation signal,the ends of the SineSweep signal are exponentially atten-uated (exponential growth at the beginning and exponen-tial decrease at the end).

In order to perform acoustical measurements over theentire audible range, the excitation signal must extendfrom 20 Hz to 20 000 Hz. As the transients outside thisrange have to be included, the choice of f1 10 Hz (ini-tial sweep frequency) and f2 22 000 Hz (final sweep fre-


(a) (b)Fig. 10. (a) Time representation of inverse filter corresponding to SineSweep signal in Fig. 9. (b) Corresponding magnitude spectrum.

(a) (b)

Fig. 9. (a) Time representation of SineSweep excitation signal (w1 2p10 rad/s, w2 2p1000 rad/s). (b) Corresponding magnitudespectrum.


quency) is realized.In practice the SineSweep deconvolution leads to the

apparition of a sequence of impulse responses clearly sep-arated along the time axis (Fig. 11). Fig. 11 shows that thedifferent harmonic distortion orders appear separatelyprior to the linear impulse response in increasing order,from right to left.

2 IMPLEMENTATION AND EXPERIMENTALSETUP

2.1 Measurement SystemA complete measurement system has been designed

and realized to enable fast, reliable, and simple compar-isons between the different methods. While several dedi-cated measurement systems already exist, they are gener-ally expensive, immutable (because of the hardwareimplementation of the algorithms), and bulky. Therefore agraphical, highly configurable, portable program written

under Matlab 5.3 has been developed for automaticallyobtaining the impulse response with common elementssuch as a microphone, a loudspeaker, and a computer. Thishas led to a global, cheap, and adaptable measurementsystem (Fig. 12), which allows fast and accurate measure-ments of the impulse response.

The Matlab program controls the generation of the dif-ferent excitation signals, their emission through the loud-speaker connected via a power amplifier to a full-duplexsoundcard, and the simultaneous recording of the signal atthe microphone. The deconvolution technique is then per-formed automatically. The time required for one measure-ment is very short: only a few seconds are necessary toobtain the impulse response of an acoustic space.

2.2 Room Impulse Response MeasurementTo measure room impulse responses accurately, the

measurement system characteristics must be taken intoconsideration. Calibration of the entire measurement


(a) (b)

Fig. 11. (a) Impulse response obtained in anechoic room with a logarithmic SineSweep of 1 s characterized by w1 2p10 rad/s andw2 2p22 000 rad/s). (b) Zoom on this response, showing extraordinary precision of achievable results.

Fig. 12. Schematic representation of measurement system.

STAN ET AL. PAPERS

chain requires an inverse filtering correction which, in thiscase, is realized by the application of the time reversalmirror filter technique [26]. This technique generates apreequalized excitation signal in three steps:

1) Determination of the measurement system impulseresponse in anechoic room (for example, by using one ofthe techniques described earlier).

2) Time reversal of the system impulse response afterappropriate truncation and addition of a time delay inorder to obtain a causal result.1 This “reversed” impulseresponse is linearly convolved with the excitation signal(MLS, sweep, . . .) that has to be preequalized.

3) Division of the spectrum magnitude of the signalobtained in step 2) by the square of the measurement sys-tem magnitude response [fast Fourier transform of theimpulse response obtained in step 1].

The results of the preequalization technique are pre-sented in Fig. 13. It can be seen that the phase spectrum isperfectly linear and the amplitude spectrum is almost con-stant (the residual oscillations around the mean value donot exceed 0.4 dB) in the range between 40 Hz and 18kHz.

3 COMPARISON OF METHODS

A comparison of the four impulse response measure-ment methods was first carried out in the anechoic roomin order to ensure individual control of the set of parame-ters conditioning the measurement. The characteristicparameters of each method were chosen in order to allowan objective comparison. For example, the duration of theexcitation signals and the number of averages used havebeen maintained constant during all measurements.

The program written under Matlab allows the followingparameters to be modified:

1) Parameters common to all methods

• Sampling frequency• Number of averages (number of times the emitted sig-

nal will be sent to the loudspeaker)• Recording mode: mono or stereo (for example, for head-

related impulse response measurements).

2) Parameters specific to each method

• MLS: Order of the MLS sequence (maximum order 19; number of samples contained in one period of anmth-order MLS sequence 2m 1).

• IRS: Order of the IRS sequence [maximum order 19;number of samples contained in one period of an mth-order IRS sequence 2 * (2m 1)].

• Time-stretched pulses: Total duration of pulse andstretching percentage (ratio between amount of timeduring which pulse has a nonnegligible amplitude andtotal duration of pulse).

• SineSweep: Initial frequency, final frequency, sweepduration, and duration of silences inserted after eachsweep.

Fig. 14 illustrates the arrangement of the transducers inthe anechoic room for comparing the different measure-ment techniques.

3.1 Optimal ParametersFig. 15 gives a general survey of the impulse responses

measured with the four methods presented when the meas-urement parameters (that is, the amplification sound level)were optimized individually for each measurement tech-nique. The magnitude scale was enlarged to focus on theresidual noise present.

The reverberation time and the ambient noise level inthe anechoic chamber being very low, the choice of anMLS sequence of order 16 seemed to be a reasonablecompromise between measurement time and good signal-to-noise ratio. The durations of the excitation signals used


(a) (b)

Fig. 13. (a) Impulse response obtained in anechoic room when excitation signal has been preequalized (logarithmic SineSweep tech-nique was used). (b) Corresponding magnitude and phase spectra.

1The aim of this step is to inverse the phase polarity of thesignal.


in the other measurement methods were then adjustedaccording to the duration of the MLS excitation signal(that is, 1.5 s for a sampling frequency of 44 100 Hz).

Thus in order to obtain comparable measurements, thefollowing parameters were used:

• Sampling frequency: 44 100 Hz

• MLS and IRS sequence orders: 16 (corresponding tosignals of 1.5- and 3.-s duration, respectively, accordingto the chosen sampling frequency)

• Output amplification level: optimized in accordancewith the method used by trial-and-error adjustment ofthe amplifier knob (the different levels used are listed inTable 1)


(c) (d)

Fig. 15. Zoom on impulse responses obtained in anechoic room. (a) MLS. (b) IRS. (c) Time-stretched pulses. (d) SineSweep.

(a) (b)

Fig. 14. Arrangement of measurement elements in anechoic room.

STAN ET AL. PAPERS

• Time-stretched pulse duration: 1.5 s• Time-stretched pulse stretch percentage: 80%• Initial and final SineSweep frequency: f1 10 Hz, f2

22 000 Hz• SineSweep duration: 1.5 s• No averaging• Noise level in the anechoic room when the computer is

present: 30 dB.

When the measurement parameters are optimized, thedifferences between the MLS and IRS methods tend tovanish [see Fig. 15(a) and (b)]. Furthermore the use of arelatively low output level and the timbre (a white noise isless disturbing than a sweep) of these methods is anadvantage if the measurements are to be made in occupiedrooms. The advantage of the IRS technique may still beconsidered. For example, at the extreme right of Fig. 15(b)the peak existing in Fig. 15(a) has disappeared. The dis-appearance of the distortion peaks when the time-stretched pulses method is used is clearly shown in Fig.15(c). Finally, the perfect separation of the harmonic dis-tortions from the linear impulse response with theSineSweep method is evident in Fig. 15(d).

The advantage of this last method lies in the fact that atedious optimization process of the measurement parame-ters is not needed to obtain optimal results since the limi-tation on the output amplification level to avoid significantdistortions no longer exists. (The nonsuperposition of thetail of the impulse response corresponding to the second-order distortion with the linear impulse response is theonly precaution that has to be taken by choosing a suffi-cient duration of the SineSweep signal [21].)

In Fig. 15 the time axes do not have the same origins inorder to focus on details particular to each method.

3.2 Signal-to-Noise RatioIn order to perform an objective comparison of the

impulse response qualities, the optimum signal-to-noiseratios achievable for each technique have been compared.In the following, the signal-to-noise ratio definition usedis the ratio expressed in dB between the average power ofthe signal recorded by the microphone and the averagepower of the noise and distortions present in the tail of thedeconvolved (linear) impulse response. Obviously, wemight expect a better signal-to-noise ratio for theSineSweep method since there are no distortion artifactspresent in the tail of the deconvolved (linear) impulse

response. This affirmation is confirmed by the results pre-sented in Table 2.

The maximum signal-to-noise ratio when a 16-bit quan-tization is used corresponds to 98 dB [27]. This upperlimit will of course never be reached in practice becauseof undesired contributions such as acoustical noise, elec-trical noise in the measurement system, quantizationerrors, and nonlinear distortions, principally due to theloudspeaker.

Table 2 shows the optimum signal-to-noise ratios (thatis, the signal-to-noise ratios obtained when the measure-ment parameters have been optimized) for each of the fourmethods presented.

Bleakley and Scaife [11] have shown that the signal-to-noise ratio for the MLS sequence increases by 3 dB whenthe period length of the MLS sequence is doubled. It isthus logical to obtain a 3-dB gain for the IRS technique incomparison with the MLS technique since the length of anIRS sequence is twice the length of the corresponding(same-order) MLS sequence.

The noticeable gain of 14 dB of the time-stretchedpulses method over the IRS method can be explained bythe use of an optimum output sound level far above theone used in the MLS or IRS cases, as well as by the dis-appearance of the spurious distortion peaks.

Finally the excellent signal-to-noise ratio (80 dB)obtained with the SineSweep technique is due to the totalrejection of the nonlinear distortion artifacts prior to thelinear impulse response. The output signal level is nolonger limited by the need to reduce the nonlinear influ-ence since all nonlinear distortions are measured sepa-rately. This leads to the optimum signal-to-noise ratio (20dB better than for the MLS method). This signal-to-noiseratio is only 3 dB above the one given by the time-stretched pulses method, but the nonsuperposition of thedistortion artifacts is guaranteed in this case.

All these signal-to-noise ratios have been obtainedthrough direct measurements (no averaging). Of course,better signal-to-noise ratios would have been obtained ifaveraging had been used.

3.3 Impulsive Noise ImmunityFig. 16 shows the impulse responses obtained when

measurements are performed in an environment whereimpulsive noise is simultaneously present. The IRSmethod gives results approximately identical to those ofthe MLS technique, and thus its corresponding impulseresponse is not shown. As announced previously, only thepseudorandom noise techniques (MLS and IRS) possessthe ability of randomizing the phase of any component inthe recorded signal that is not correlated to the input sig-nal emitted in the acoustical space. Thus any additionalnoise (white or even impulsive) will be distributed uni-formly along the deconvolved impulse response.Therefore additive impulsive noise (appearing as additivewhite noise in the deconvolved impulse response) is sub-ject to posterior attenuation by averaging techniques.

On the contrary, the presence of impulsive noise (Fig.16) when the time-stretched pulses or the SineSweep tech-niques are used, heavily compromises the impulse


Table 1. Optimum sound levels at the position of themicrophone for each method.

MLS IRS Time-Stretched Pulses SineSweep

75.5 dB 75.5 dB 83.9 dB 92.5 dB

Table 2. Optimum signal-to-noise ratios for each method.

MLS IRS Time-Stretched Pulses SineSweep

60.5 dB 63.2 dB 77.0 dB 80.1 dB


response deconvolution process through the presence ofexcitation signal residues in the deconvolved impulseresponse. These residues being strongly correlated withthe excitation signal will not be eliminated properly if pos-terior averaging techniques are used.

As a first conclusion we may say that in a (nonrandom)noisy environment, the MLS (IRS) method is subject togiving better results than the other two methods.

3.4 Impulse Response Measurements inClassical Rooms

The properties described so far were mostly illustratedby measurements performed in anechoic rooms. In thissection we will show the results obtained when the meas-urements are performed in more classical rooms such asauditoriums or lecture rooms.

As can be seen in Figs. 17 and 18, most of the proper-ties that were illustrated in the case of anechoic measure-ments are still visible. In particular we can focus on theseparation of the different harmonic distortions from the

linear impulse response in the case of the SineSweep tech-nique. In this case the impact of reverberation on eachimpulse response can be clearly seen and illustrates theneed for nonsuperposition of the second-order distortionwith the linear impulse response by using a sufficientlylong SineSweep.

4 CONCLUSIONS

A complete, inexpensive, and parameterizable measure-ment system has been realized in order to compare differ-ent impulse response measurement methods. This systemis based on a computer program written under Matlab 5.3,allowing an automatic and easy measurement to be real-ized. On the basis of this measurement system, four dif-ferent methods have been compared.

The comparison of the four methods leads to the fol-lowing conclusions.

1) The MLS (IRS) method seems the most interestingmethod when the measurements have to be performed in


(b) (c)

Fig. 16. Impulse responses obtained in anechoic room when intense impulsive noise is simultaneously present. (a) MLS. (b) Time-stretched pulses. (c) SineSweep. Note that amplitude scales are not identical.

(a)

STAN ET AL. PAPERS

an occupied room or in the exterior because of its strongimmunity to all kinds of noise (white, impulsive, or oth-ers), its weak optimum output sound level, and its timbre(white noise is more bearable and more easily masked outthan are sweep signals). However, its major drawback liesin the tedious calibration that has to be carried out toobtain optimum results and in the appearance of spuriouspeaks (distortion peaks) due to the inherent nonlinearitiesof the measurement system.

2) The time-stretched pulses method avoids the appear-ance of the distortion peaks. However, the remaining non-linear artifacts can possibly be superimposed with thedeconvolved “linear” impulse response. The presence of aresidue of the excitation signal in the deconvolved impulseresponse is a result of such superposition problems. Thisresidue can be almost completely eliminated if a precisecalibation of the measurement parameters (mainly the out-put level) is realized. However, its timbre and the highoptimum output signal level needed to mask out the ambi-

ent noise make it unusable in occupied rooms.3) The perfect and complete rejection of the harmonic

distortions prior to the “linear” impulse response, theirindividual measurement, and the excellent signal-to-noiseratio of the SineSweep method make it the best impulseresponse measurement technique in an unoccupied andquiet room. Moreover, unlike the preceding methods, itdoes not necessitate a tedious calibration in order to obtainvery good results (no compromise between the signal-to-noise ratio and the superposition of nonlinear artifacts inthe room impulse response). However, like the time-stretched pulses method, the SineSweep technique is notrecommended for measurements in occupied rooms.

5 REFERENCES

[1] M. Kleiner, B. I. Dalenbäck, and P. Svensson,“Auralization––An Overview,” J. Audio Eng. Soc., vol.41, pp. 861–875 (1993 Nov.).


(c) (d)

Fig. 17. Impulse responses obtained in auditorium 604, Europe Amphitheater, University of Liège (Belgium). (a) MLS. (b) IRS.(c) Time-stretched pulses. (d) SineSweep.

(a) (b)


[2] H. Lehnert and J. Blauert, “Principles of BinauralRoom Simulation,” Appl. Acoust., vol. 36, pp. 259–291(1992).

[3] M. R. Schroeder, “Integrated-Impulse Method forMeasuring Sound Decay without Using Impulses,” J.Acoust. Soc. Am., vol. 66, pp. 497–500 (1979).

[4] C. Dunn and M. O. Hawksford, “Distortion Immunityof MLS-Derived Impulse Response Measurements,” J.Audio Eng. Soc., vol. 41, pp. 314–335 (1993 May).

[5] D. D. Rife and J. Vanderkooy, “Transfer-FunctionMeasurement with Maximum-Length Sequences,” J.Audio Eng. Soc., vol. 37, pp. 419–443 (1989 June).

[6] J. Vanderkooy, “Aspects of MLS Measuring Systems,”J. Audio Eng. Soc., vol. 42, pp. 219–231 (1994 Apr.).

[7] H. Alrutz and M. R. Schroeder, “A Fast HadamardTransform Method for the Evaluation of MeasurementsUsing Pseudorandom Test Signals,” in Proc. 11th Int.Conf. on Acoustics (Paris, France, 1983), pp. 235–238.

[8] H. R. Simpson, “Statistical Properties of a Class of

Pseudorandom Sequence,” Proc. IEE (London), vol. 113,pp. 2075–2080 (1966).

[9] D. D. Rife, “Modulation Transfer FunctionMeasurement with Maximum-Length Sequence,” J. AudioEng. Soc., vol. 40, pp. 779–790 (1992 Oct.).

[10] J. Borish and J. B. Angell, “An Efficient Algorithmfor Measuring the Impulse Response Using Pseudoran-dom Noise,” J. Audio Eng. Soc., vol. 31, pp. 478–488(1983 July/Aug.).

[11] C. Bleakley and R. Scaife, “New Formulas forPredicting the Accuracy of Acoustical MeasurementsMade in Noisy Environments Using the Averaged m-Sequence Correlation Technique,” J. Acoust. Soc. Am.,vol. 97, pp. 1329– 1332 (1995).

[12] M. Cohn and A. Lempel, “On Fast m-SequenceTransforms,” IEEE Trans. Inform. Theory, vol. IT-23, pp.135–137 (1977).

[13] W. D. T. Davies, “Generation and Properties ofMaximum-Length Sequences,” Control, pp. 302–304,


(c) (d)

Fig. 18. Zoom on impulse responses obtained in auditorium 604, Europe Amphitheater, University of Liège (Belgium). (a) MLS.(b) IRS. (c) Time-stretched pulses. (d) SineSweep.

(a) (b)

STAN ET AL. PAPERS

364–365, 431–433 (1966).[14] M. Vorländer and M. Kob, “Practical Aspects of

MLS Measurements in Building Acoustics,” Appl.Acoust., vol. 52, pp. 239–258 (1997).

[15] R. Burkard, Y. Shi, and K. E. Hecox, “A Compari-son of Maximum Length and Legendre Sequences for theDerivation of Brain-Stem Auditory-Evoked Responses atRapid Rates of Stimulation,” J. Acoust. Soc. Am., vol. 87,pp. 1656–1664 (1990).

[16] N. Ream, “Nonlinear Identification Using Inverse-Repeat m Sequences,” Proc. IEE (London), vol. 117, pp.213–218 (1970).

[17] P. A. N. Briggs and K. R. Godfrey, “Pseudoran-dom, Signals for the Dynamic Analysis of MultivariableSystems,” Proc. IEE, vol. 113, pp. 1259–1267 (1966).

[18] N. Aoshima, “Computer-Generated Pulse SignalApplied for Sound Measurement,” J. Acoust. Soc. Am.,vol. 65, pp. 1484–1488 (1981).

[19] Y. Suzuki, F. Asano, H. Y. Kim, and T. Sone, “AnOptimum Computer-Generated Pulse Signal Suitable forthe Measurement of Very Long Impulse Responses,” J.Acoust. Soc. Am., vol. 97, pp. 1119–1123 (1995).

[20] A. Farina and E. Ugolotti, “Subjective Comparisonbetween Stereo Dipole and 3d Ambisonic SurroundSystems for Automotive Applications,” presented at the

AES 16th International Conference on Spatial SoundReproduction (1999). Available at url: HTTP://pcfarina.eng.unipr.it.

[21] A. Farina, “Simultaneous Measurement of ImpulseResponse and Distorsion with a Swept-Sine Technique,”presented at the 108th Convention of the Audio Engineer-ing Society, J. Audio Eng. Soc. (Abstracts), vol. 48, p. 350(2000 Apr.), preprint 5093.

[22] A. J. Berkhout, M. M. Boone, and C. Kesselman,“Acoustic Impulse Response Measurement: A NewTechnique,” J. Audio Eng. Soc., vol. 32, pp. 740–746(1984 Oct.).

[23] J. Borish, “Self-Contained Crosscorrelation Programfor Maximum-Length Sequences,” J. Audio Eng. Soc.(Engineering Reports), vol. 33, pp. 888–891 (1985 Nov.).

[24] S. W. Golomb, Shift-Register Sequences, rev. ed.(Aegan Park Press, Laguna Hills, CA, 1982).

[25] A. V. Oppenheim and R. W. Schafer, Discrete-TimeSignal Processing, 2nd ed. (Prentice-Hall SignalProcessing Ser., Englewood Cliffs, NJ, 1999).

[26] D. Preis, “Phase Distortion and Phase Equalizationin Audio Signal Processing––A Tutorial Review,” J.Audio Eng. Soc., vol. 30, pp. 774–794 (1982 Nov.).

[27] K. C. Pohlmann, Principles of Digital Audio, 3rded. (McGraw-Hill, New York, 1995).


THE AUTHORS

Guy-Bart Stan was born in Liège, Belgium, in 1977. Hereceived an electrical (electronics) engineering degree in2000 from the University of Liège, where he is currentlya Ph.D. student with F.N.R.S. (National Fund forScientific Research) support. His final study work dealtwith the measurement of room impulse responses andhead-related impulse responses. His current research is inthe area of nonlinear systems analysis and control.

Jean Jacques Embrechts studied electronics engineer-ing at the University of Liège, Belgium, where he grad-uated from in 1981. He then joined the team of audio,acoustics, and lighting engineering at the same univer-sity; and was involved in several research projects:sound ray tracing, computer-aided design for roomacoustics and lighting, modeling of light, and soundscattering by surfaces. He obtained a Ph.D. in appliedsciences in 1987.

Since 1999 Dr. Embrechts has been professor of soundand image techniques in the Department of ElectricalEngineering and Computer Science at the University ofLiège. His research interests in room acoustics includesound diffusion (with participation in the AES and ISOgroups working on that subject), auralization, and signalprocessing for measurement systems. He is also the teamleader of research projects in active noise control and pho-tometry. He is a member of the AES.

Dominique Archambeau was born in Verviers,Belgium, in 1973. He studied electrical engineering at theUniversity of Liège (ULg), Belgium, and received anM.Sc. degree in 1996.

Since 1997 he has been working as assistant lecturer atthe ULg Laboratory of Acoustics. His research interestsinclude digital signal processing, room acoustics simula-tion software, and presently automatic learning methods.

G.-B. Stan J. J. Embrechts D. Archambeau

ENGINEERING REPORTS

0 INTRODUCTION

Virtual auditory space can be generated by measuringthe location-dependent filtering effects of the torso, head,and pinnae [the head-related transfer functions (HRTFs)]and then presenting sound filtered by those HRTFs to alistener via headphones. Such sound appears to emanatefrom outside the listener’s head at the locations for whichthe HRTFs were measured [1], [2].

For the signals reaching a listener’s ears during head-phone presentation to accurately simulate those that reachthe ears during free-field presentation, a correction mustbe made to compensate for the headphone-to-ear-canaltransfer function (HpTF). The HpTF is comprised of twocomponents: the transfer function of the transducer andthe transfer function from the transducer to the ear canal.Kulkarni and Colburn [3] have shown that spectral fea-tures, for example, peaks and notches, in HpTFs can be ofsimilar magnitude and bandwidth as those in HRTFs.Spectral features in HRTFS are thought to provide impor-tant cues to sound-source location [4], [5]. In addition,HpTFs may incorporate group delays that vary betweenears and across placements on a single ear to an extent thatobscures the interaural time differences that provide lis-teners with information about the cone of confusion onwhich a sound source lies [6]. An appropriate correction

for the HpTF, therefore, is essential to the production ofhigh-fidelity virtual auditory space.

On observing high variability among the magnitudes ofthe HpTFs of a group of listeners, Møller et al. [7] andPralong and Carlile [8] have argued that correction for theHpTF needs to be listener specific. Kulkarni and Colburn[3] have claimed that making appropriate corrections willbe difficult, as the variability of HpTFs is also high acrossheadphone placements for a given listener. Presumablythis latter variability results from the variability of thetransfer function from the transducer to the ear canalacross headphone placements. Kulkarni and Colburn [3]measured the HpTFs associated with 20 headphone place-ments on an acoustic manikin and reported a standarddeviation of the magnitudes that approached 9 dB forsome frequencies between 9 and 14 kHz. They argued thatthis would make the stimulus waveform at a listener’s earsunpredictable and may reduce the perceptual adequacy ofa virtual audio display. The variability of HpTFs acrossheadphone placements has also been examined byWightman and Kistler [9] and Pralong and Carlile [8].They measured the HpTFs on humans associated with tenand six headphone placements, respectively [9], [8] andon an acoustic manikin [8], and reported standard devia-tions of the magnitudes that did not exceed 5 dB at anyfrequency from 0.2 to 14 kHz.

We have recently assessed the ability of listeners tolocalize sound presented via a virtual audio display thatincorporated listener-specific corrections for the HpTF


Variability in the Headphone-to-Ear-CanalTransfer Function*

KEN I. MCANALLY AND RUSSELL L. MARTIN

Air Operations Division, Defence Science and Technology Organisation, Melbourne 3001, Victoria, Australia

Headphone-to-ear-canal transfer functions (HpTFs) for 20 headphone placements weremeasured for each ear of three participants and an acoustic manikin. Head-related transferfunctions (HRTFs) were measured for nine sound-source locations within a 14.5º radius ofeach of eight representative locations. Noises were convolved with these functions and passedthrough a cochlear filter model to estimate cochlear excitation. The variability of the magni-tudes of the filtered HpTFs was much less than the variability of the magnitudes of the unfil-tered HpTFs. It was also considerably less than the variability of the magnitudes of thefiltered HRTFs. In addition, the variability of the group delays of the HpTFs for the threehuman participants was considerably less than the minimum discriminable interaural timedifference. It follows that much of the information in HRTFs that could provide a cue tosound-source location will not be masked by the variability of HpTFs across headphoneplacements. The spatial fidelity of an individualized virtual audio display, therefore, will notnecessarily be compromised by variability in HpTFs

* Manuscript received 2001 January 22; revised 2001 Decem-ber 27.

MCANALLY AND MARTIN ENGINEERING REPORTS

[2]. The correction for each listener was based on a singleHpTF measurement made at the time of HRTF measure-ments. Localization performance was assessed acrosseight sessions and, therefore, eight headphone placements.We found that listeners could localize virtual sound withfree-field equivalent accuracy, and that performanceacross the eight sessions in which virtual sound was local-ized was no more variable than that across the eight ses-sions in which free-field sound was localized. Such accu-rate and stable performance when localizing virtual soundis surprising in view of Kulkarni and Colburn’s [3] reportof high variability among the HpTFs associated with dif-ferent headphone placements.

We report here on the variability of the HpTFs associ-ated with 20 headphone placements on an acousticmanikin and the three listeners who participated in thestudy by Martin et al. [2]. We also describe the variabilityof the same HpTFs after they were passed through acochlear filter model [10], which estimates cochlear exci-tation. As fine spectral detail in unfiltered HpTFs wouldnot survive cochlear filtering (see [11] for a review), it isappropriate to remove that detail when considering thepotential impact of HpTF variability on audio-displayfidelity. We also describe the variability of the groupdelays of the HpTFs for the three listeners in Martin etal.’s [2] study.

For listeners to distinguish between sound-source loca-tions on the basis of HRTF spectral features, each locationmust be associated with a unique combination of features.Irrespective of their magnitudes, individual features thatdo not vary with location can contain no spatial informa-tion. Therefore one measure of the spectral signal avail-able to a listener for the purpose of localization is the vari-ation of the magnitudes of the HRTFs associated withclosely spaced sound-source locations. HpTF variabilityacross headphone placements is a measure of the noisethat would be introduced into a virtual audio display byemploying a fixed correction for the HpTF for each listener.

In this engineering report we compare the variability ofthe magnitudes of cochlear filtered HpTFs with that of themagnitudes of filtered HRTFs associated with nine sound-source locations within a 14.5º radius of each of eight rep-resentative locations. A 14.5º radius is appropriate becauseboth the absolute localization error for binaural listening[12], [2] and the minimum audible angle for monaural lis-tening [13] for sound presented in the free field are closeto 14.5º. Therefore the variability of the magnitudes of fil-tered HRTFs within a 14.5º radius should provide an esti-mate of the smallest spectral signal that listeners can useto determine sound-source location.

1 METHODS

Three adult listeners (two women and one man agedbetween 22 and 41 years) participated in this study. Asnoted earlier, they were the same three listeners who hadparticipated in the auditory localization study described inMartin et al. [2].

HpTFs and HRTFs were measured using Golay codes[14] and a “blocked-ear-canal” technique [7]. For a full

description of the procedure see Martin et al. [2]. Parti-cipants sat at the center of a sound-attenuating anechoicroom. Miniature microphones (Sennheiser KE4-211-2),embedded in foam ear protectors (Earlink, Cabot SafetyCorp.) or swimmer’s ear putty (Antinois, Panmedica),were placed in the participants’ ear canals. The micro-phone diaphragms were flush with the canal entrances.Golay codes (8192 points in length) were presented at 50kHz (Tucker Davis Technologies System II) and 75 dB(A-weighted) through a pair of headphones (SennheiserHD520 II). HpTFs were calculated from the responses ofthe microphones [14], which were low-pass filtered at 20kHz and sampled at 50 kHz (Tucker Davis TechnologiesSystem II). HpTFs were measured for 20 placements ofthe headphones on each listener and were not corrected forthe transfer function of the microphone (which was con-stant across all placements and therefore could not con-tribute to HpTF variability). As it is not clear whether theacoustic coupling of the headphone would be affected bythe blocked-ear-canal technique, HpTFs were also meas-ured via the microphones of an acoustic manikin (HeadAcoustics HMS II.3).

HRTFs were measured for 378 locations ranging from50º to 70º of elevation and covering 360º of azimuthat intervals of approximately 10º subtended at the partici-pant’s head. Golay codes were presented at 50 kHz and 75dB (A-weighted) through a loudspeaker (Bose FreeSpacetweeter) mounted 1 m from the participant. HRTFs werecalculated from the responses of the microphones andwere not corrected for the transfer functions of the loud-speaker or microphone, both of which were constantacross locations and therefore could not contribute toHRTF variability. (Note, however, that these correctionsare required for generating virtual sound that is accuratelylocalized [2].)

HRTFs and HpTFs were converted to impulse respon-ses, which were then truncated to 20.84 ms. Gaussiannoise (328 ms in duration, including 20-ms rise and falltimes) was convolved with the impulse response associ-ated with each HRTF and HpTF and passed through acochlear filter model ([10] as implemented by [15]). Themodel was comprised of a gammatone filter bank simulat-ing 80 frequency channels (two per equivalent rectangularbandwidth) with center frequencies distributed from 0.1 to25 kHz. The bandwidths of these filters increased withincreasing center frequency.

The group delay of each HpTF for each of the threehuman participants was determined from the slope of theHpTF phase from 0 to 8 kHz.

2 RESULTS

Standard deviations of the magnitudes of the unfilteredHpTFs associated with 20 headphone placements on eachof the six human ears are shown in Fig. 1(a) (solid lines)as a function of frequency. Standard deviations were gen-erally smaller than 2.5 dB for frequencies up to 10 kHz.The one exception was a peak of 6.3 dB at 8.5 kHz for oneear. Standard deviations were larger for frequencies above10 kHz and ranged up to about 9 dB.


ENGINEERING REPORTS VARIABILITY IN TRANSFER FUNCTION

Standard deviations of the magnitudes of the HpTFsassociated with 20 headphone placements on the acousticmanikin are also shown in Fig. 1(a) as a function of fre-quency (dashed lines). These standard deviations weresimilar in magnitude to those for the six human ears,except that they revealed slightly greater variability (up to1 dB) at frequencies below 500 Hz.

The HpTFs were convolved with noise and passedthrough a cochlear filter model to estimate the pattern ofcochlear excitation. Standard deviations of the magnitudesof the filtered HpTFs associated with 20 headphone place-ments on each of the six human ears (solid lines) and theacoustic manikin (dashed lines) are shown in Fig. 1(b).For the human ears, standard deviations again tended toincrease with increasing frequency but were generallysmaller than for the unfiltered HpTFs. For frequencies upto 10 kHz standard deviations were smaller than 1.4 dB.For frequencies above 10 kHz they ranged up to 2.3 dB forfive of the six ears and up to 3.6 dB for all ears. Standarddeviations for the manikin were again similar in magni-tude to those for the six human ears.

Standard deviations of the magnitudes of the filteredHRTFs associated with nine sound-source locations

within a 14.5º radius of each of eight representative loca-tions for one human ear (RM, right ear) are shown in Fig.2 as a function of frequency. All locations are in the righthemifield, as monaural spectral cues are thought to bemost useful for localizing sources in the ipsilateral field. Itcan be seen that these standard deviations were, in gen-eral, considerably larger than those for the filtered HpTFs.

The standard deviation of the group delays across head-phone placements varied between 1.04 and 2.13 ms for thesix human ears tested.

3 DISCUSSION

The variability of the magnitudes of unfiltered HpTFsdescribed in this engineering report is similar to that re-ported by Kulkarni and Colburn [3]. Substantial variabil-ity ranging up to about 9 dB was observed for frequenciesbetween 12 and 18 kHz for all ears. For one ear, variabil-ity was also substantial at 8.5 kHz. This variability is con-siderably greater than that reported by Wightman andKistler [9] and Pralong and Carlile [8] for a smaller num-ber of headphone placements (ten and six, respectively).As noted earlier, these authors reported variability that didnot exceed 5 dB at any frequency from 0.2 and 14 kHz.

The variability of the magnitudes of the unfilteredHpTFs was similar for the manikin and the humans forwhom measurements were made in this study. This sug-gests that the blocked-ear-canal technique used whenmaking the measurements in humans did not have a largeeffect on the variability of the HpTFs.

The variability of the magnitudes of HpTFs was greatlyreduced when they were passed through a cochlear filtermodel. This indicates that the unfiltered HpTFs varied pri-marily with respect to fine spectral detail at high frequen-cies. Consistent with this interpretation, inspection of Fig.1(a) reveals that peaks of variability for a given ear wereusually of particularly narrow bandwidth. As noted earlier,greatest variability was observed for frequencies above


Fig. 1. Standard deviations of HpTF magnitudes associated with20 headphone placements on each of six human ears (––) and anacoustic manikin (– – –), (a) Unfiltered. (b) Filtered. These latterstandard deviations relate to the intensity of modeled cochlearexcitation induced by noise-convolved HpTFs.

Fig. 2. Standard deviations of filtered HRTF magnitudes associ-ated with nine sound-source locations within a 14.5º radius ofeach of eight representative locations in ipsilateral hemifield forone human ear (RM, right ear). These standard deviations relateto the intensity of modeled cochlear excitation induced by noise-convolved HRTFs. Representative locations: 10, 0; 60, 0; 90, 0;60 30; 120, 30; 60, 30; 60, 60; 120, 60º of azimuth and ele-vation, respectively.

(b)

(a)

MCANALLY AND MARTIN ENGINEERING REPORTS

about 10 kHz. This variability is likely to have arisen fromdestructive interference. As cochlear filters at these fre-quencies are relatively broad, this variability would havebeen reduced by the cochlear filter model.

In agreement with Kulkarni and Colburn [3], inspectionof single HpTFs and HRTFs from our study (not shownhere) revealed that HpTF and HRTF spectral features canbe of similar magnitude. However, as argued earlier, it isthe variation of HRTFs across sound-source locations thatprovides the spectral signal from which a listener candetermine location.

We investigated the variability of the magnitudes of thefiltered HRTFs associated with nine sound-source loca-tions within a 14.5º radius of each of eight representativelocations. As 14.5º is similar to both the absolute localiza-tion error for binaural listening [12], [2] and the minimumaudible angle for monaural listening [13], the variabilityof the magnitudes of filtered HRTFs within a 14.5º radiusshould provide an estimate of the smallest spectral signalthat listeners can use to determine sound-source location.We have shown that the variability of the magnitudes offiltered HRTFs is, in general, considerably greater thanthat of the magnitudes of filtered HpTFs. This suggeststhat the spectral information used by listeners to localizesound is unlikely to be masked by the variability of theHpTF magnitude.

As the fidelity of a virtual audio display could also becompromised by inappropriate correction for the HpTFgroup delay, we also investigated the variability of thegroup delays of HpTFs across headphone placements. Thestandard deviation of HpTF group delays across place-ments was found to be less than 2.2 ms. As this is consid-erably less than the smallest discriminable interaural timedifference for tones (6 ms at 900 Hz [16]) or clicks (11 ms[17]), the use in a virtual audio display of a fixed HpTFcorrection for each listener is unlikely to have a significanteffect on the accuracy with which the cone of confusionon which a sound source lies is judged.

4 CONCLUSION

We conclude that the variability of HpTFs across head-phone placements is unlikely to have an adverse effect onthe fidelity of a virtual audio display. This conclusion isconsistent with our previous demonstration of free-fieldequivalent localization of sound presented via a virtualauditory display employing fixed HpTF corrections foreach listener [2].

5 ACKNOWLEDGMENT

The authors wish to thank Dr. Gavan Lintern and twoanonymous reviewers for providing comments on an ear-lier version of this manuscript.

6 REFERENCES

[1] F. L. Wightman and D. J. Kistler, “HeadphoneSimulation of Free-Field Listening: II. PsychophysicalValidation,” J. Acoust. Soc. Am., vol. 85, pp. 868–878

(1989 Feb.).[2] R. L. Martin, K. I. McAnally, and M. A. Senova,

“Free-Field Equivalent Localization of Virtual Audio,” J.Audio Eng. Soc., vol. 49, pp. 14–22 (2001 Jan./Feb.).

[3] A. Kulkarni and H. S. Colburn, “Variability in theCharacterization of the Headphone Transfer-Function,” J.Acoust. Soc. Am., vol. 107, pp. 1071–1074 (2000 Feb.).

[4] Z. M. Fuzessery, “Speculations on the Role of Freq-uency in Sound Localization,” Brain Behav. Evol., vol. 28,pp. 95–108 (1986).

[5] J. C. Middlebrooks and D. M. Green, “Sound Local-ization by Human Listeners,” Ann. Rev. Psychol., vol. 42,pp. 135–159 (1991).

[6] A. W. Mills, “Auditory Localization,” in Foun-dations of Modern Auditory Theory, J. V. Tobias, Ed.(Academic Press, New York, 1972), pp. 303–438.

[7] H. Møller, D. Hammershøi, C. B. Jensen, and M. F.Sørensen, “Transfer Characteristics of HeadphonesMeasured on Human Ears,” J. Audio Eng. Soc., vol. 43,pp. 203–217 (1995 Apr.).

[8] D. Pralong and S. Carlile, “The Role of Individual-ized Headphone Calibration for the Generation of HighFidelity Virtual Auditory Space,” J. Acoust. Soc. Am., vol.100, pp. 3785–3793 (1996 Dec.).

[9] F. L. Wightman and D. J. Kistler, “HeadphoneSimulation of Free-Field Listening: I. Stimulus Synthe-sis,” J. Acoust. Soc. Am., vol. 85, pp. 858–867 (1989 Feb.).

[10] R. D. Patterson, K. Robinson, J. Holdsworth, D.McKeown, C. Zhang, and M. Allerand, “Complex Soundsand Auditory Images,” in Auditory Physiology andPerception, Y. Cazals, L. Demany, and K. Horner, Eds.(Pergamon, Oxford, UK, 1992), pp. 429–446.

[11] B. C. J. Moore, “Frequency Analysis and Mask-ing,” in Hearing, B. C. J. Moore, Ed. (Academic Press,San Diego, CA, 1995), pp. 161–205.

[12] S. R. Oldfield and S. P. A. Parker, “Acuity of SoundLocalisation: A Topography of Auditory Space. I. NormalHearing Conditions,” Perception, vol. 13, pp. 581–600(1984).

[13] D. R. Perrott and K. Saberi, “Minimum AudibleAngle Thresholds for Sources Varying in Both Elevationand Azimuth,” J. Acoust. Soc. Am., vol. 87, pp. 1728–1731 (1990 Apr.).

[14] B. Zhou, D. M. Green, and J. C. Middlebrooks,“Characterization of External Ear Impulse ResponsesUsing Golay Codes,” J. Acoust. Soc. Am., vol. 92, pp.1169–1171 (1992 Aug.).

[15] M. Slaney, “Auditory Toolbox: A Matlab Toolboxfor Auditory Modeling Work, Version 2,” Tech. Rep.1998-010, Interval Corp., Palo Alto, CA (1998).

[16] W. A. Yost, “Discrimination of Interaural PhaseDifferences,” J. Acoust. Soc. Am., vol. 55, pp. 1299–1303(1974 June).

[17] R. G. Klumpp and H. R. Eady, “Some Measure-ments of Interaural Time Difference Thresholds,” J.Acoust. Soc. Am., vol. 28, pp. 859–860 (1956 Sept.).


Biographies for Ken I. McAnally and Russell L. Martin werepublished in the 2002 January/February issue of the Journal.

Report of the meeting of the SC-02-02Working Group on Digital Input/Output Interfacing of the SC-02 Subcommittee onDigital Audio, held in conjunction with theAES 111th Convention in New York, NewYork, US, 2001-11-28

Chair J. Dunn convened the meeting. The agenda and thereport from the previous meeting were approved as writtenexcept that a correction was made in the fourth paragraph of

discussion of project AES-X119 which should begin “Itshould not be . . . .”

Current development projects

AES3-R Revision of AES3-1992 (r1997), AES recommended practice for digital audio engineering—Serial transmission format for two-channel linearly represented digital audio dataA secretariat-prepared PWD for revision of AES3 had beenposted on the FTP site two months prior to the meeting and


COMMITTEE NEWSAES STANDARDS

Information regarding Standards Committee activi-ties including meetings, structure, procedures, re-ports, and membership may be obtained viahttp://www.aes.org/standards/. For its publisheddocuments and reports, including this column, theAESSC is guided by International ElectrotechnicalCommission (IEC) style as described in the ISO-IECDirectives, Part 3. IEC style differs in some respectsfrom the style of the AES as used elsewhere in thisJournal. For current project schedules, see the pro-ject-status document on the Web site. AESSC docu-ment stages referenced are proposed task-groupdraft (PTD), proposed working-group draft (PWD),proposed call for comment (PCFC), and call forcomment (CFC).

The AES Standards Committee was one of the firststandards bodies to embrace electronic communications as ameans of carrying out its work. Two meetings per year aresimply not sufficient to support the work demanded by theincreasingly rapid rate of change of technology in the fieldof professional audio engineering. The people who con-tribute their expertise to standards working groups are,almost by definition, very busy and it is difficult to arrangemore frequent meetings. Consequently, our ongoing workdepends on the use of effective communications other thanface-to-face meetings.

The benefits of electronic communication, now dominatedby the internet, have been proved by our pioneering work inthis field. The impact on the work of the StandardsCommittee has been highly successful over the last fiveyears and more. Replacing paper with internet communi-cations has brought us greatly increased speed and produc-tivity. It is now quite hard to remember a time when alldocuments needed to be duplicated on paper and sent byregular mail, when action could only happen at twomeetings per year, and when a formal decision could takesix months.

Access to the internet is now ubiquitous to the extent thatfew people in our field do not have regular access to e-mailand the internet. In the past, and in addition to publication in

the Journal, notification of Calls for Comment have beencirculated to AES members by physical mail. This is bothexpensive and decreasingly effective. To communicate moreeffectively, the AESSC Steering Committee has decided thatthe time is ripe to extend the rôle of the internet. In future,and in addition to publication in the Journal, the primarypublication of Calls for Comment and similar communi-cations will be by posting on the AES Standards Web site.

For those members who do not yet have internet access,and wish to receive these communications, we offer anoption to receive them by physical mail. This facility is in-tended to smooth a transitional period and will be subject toreview.

If you wish to register for notifications via paper mail,please mail the secretariat at the address below stating yourname, AES membership number, and full mailing addresstogether with telephone and fax number if appropriate:AESSC Secretariat, Mail Dept., Woodlands, GoodwoodRise, Marlow, Bucks SL7 3QE, UK.

Otherwise, please visit our Web site at www.aes.org/standards.

Mark YongeStandards Secretary e-mail [email protected]

Revision to Call for Comment Communications

there has been no resulting discussion on the e-mail re-flector. The draft is intended to conform to current AESSCstyle based on the IEC directives for the preparation ofstandards. All changes are editorial.

Three documents have been presented to the WGproposing new drawings for the example of a generalcircuit configuration (figure 6 of AES3-1992). They wereproduced by C. Travis, J. Brown, and S. Scott.

The figure is informative. The main issues raised were theinclusion of a receiver termination and the connection of thecable screen to the chassis (rather than some internal circuitground trace).

In addition, there was concern about implying that the useof transformers is required and the circuit changes neededwhen using very low cost transformers. Scott proposed twofigures illustrating implementations with and without trans-formers. In both cases there are d.c. blocking capacitors toensure no d.c. path between the signal conductors and nod.c. path from the signal conductors to anything else.

The Travis proposal was for one figure with minimalchanges from the current figure. It adds a termination re-sistor across the two signal conductors at the receiver sideand shows the cable screen’s connections only connecting tothe chassis at each end of the interface.

Brown proposed a figure and associated text. On it, trans-formers, capacitors, and resistors are replaced by a driving-network box at the driver end of the cable and atermination-isolation-network box at the receiving end.

The chair suggested that the conclusion of the group maybe to create an informative annex for this example circuit.This would make it more obvious that the circuit was notpart of the specification and would also allow more space fortext. It is also possible that the information could be movedto AES-2id.

Discussion of this circuit continues on the WG e-mail re-flector. It was noted that this issue affects similar work beingundertaken in IEC MT 60958-4.

AES-2id-R Review of AES-2id-1996, AES informationdocument for digital audio engineering—Guidelines forthe use of the AES3 interfaceIt was noted that some of the discussion on project AES3-Rmay have an impact on this project.

AES-3id-R Review of AES-3id-2001, AES informationdocument for digital audio engineering—Transmission ofAES3 formatted data by unbalanced coaxial cableNo action was required.

AES-10id-R Review of AES-10id-1995, AES informationdocument for digital audio engineering—Engineeringguidelines for the multichannel audio digital interface(MADI) AES10R. Caine will review the document and highlight any partsthat appear to need revision as a result of updating AES10.

AES10-R Review of AES10-1991 (r1997), AES recommended practice for digital audio engineering—Serial multichannel audio digital interface (MADI)It was decided at the last meeting to try to resolve the

comment by V. Recipon on the CFC without introducing asubstantive change. This had proven to be not possible, sothe draft was returned to the WG which assigned it to taskgroup SC-02-02-D. A PWD is to be reported by the taskgroup with a target date of 2002-02.

AES18-R Review of AES18-1996, AES recommendedpractice for digital audio engineering—Format for theuser data channel of the AES digital audio interfaceA CFC for reaffirmation is in process.

AES41-R Review of AES41-2000, AES standard fordigital audio engineering—Recoding data set for audio bit-rate reductionNo action is required.

AES-X50 Guidelines for Development of SpecificationsWhich Reference or Use AES3 Formatted DataThe group noted that IEC 61883-6 and the AES-X92 projectcan both carry AES3 formatted data so this project is topicalbut resources are not available to complete it. The WG is re-questing suspension of the project.

AES-X92 Digital Audio in Asynchronous Transfer Mode(ATM)The secretariat has recently produced a PCFC based on thetask group document which is now in the WG for consid-eration. The group decided to return the document to thetask group in two weeks, after the WG has been given timeto identify any items that need more work. It was noted thatthis standard is being driven by a user, the BBC, who needsit developed urgently. Many of the manufacturers known tobe developing audio over ATM do not appear to be activeparticipants in the project or, at least, have not joined thetask group.

AES-X94 Presto: Audio via Synchronous DigitalHierarchy (SDH)The group is requesting suspension of this project.

AES-X111 Transmission of the Unique MaterialIdentifier (UMID) on AES3C. Chambers initiated the project and posted a draft re-quirements specification and an updated project definitionon the working group e-mail reflector on 2001-08-17. Cainesuggested, in view of project AES-X114 assigned to SC-06-06, that project AES-X111 may result only in providing aflag in AES3 channel status so that, for example, the userdata channel could be identified as being used for carryingthe data defined in project AES-X114.

The group agreed to form a task group to prepare a draftamendment to AES3. The chair proposed that Chamberscontinue as convenor.

AES-X119 Connector for AES3 InterfacesThe convenor of the task group to which the project is as-signed has resigned.

Brown presented a block diagram to illustrate analogcabling infrastructure in a live performance venue. Thediagram illustrates a problem with mixing digital and analogsignals on the same connectors. Cable lengths of up to 500


AES STANDARDSCOMMITTEE NEWS

feet, and of mixed impedance, are used for the infrastructurebut will not work reliably with digital signals.

Caine suggested that the task group consider whether theconnector known as the XLD is to be incorporated intoAES3 as an option or as a replacement for the XLR, or notbe used at all. A. Eckart pointed out that AES42 describes anew connector (the XLD) that is not yet in production. Hesuggested that if this was only made an optional connector inAES3 then people would not bother to adopt the XLD andso it would not be produced.

Eckart considered that it was unfortunate that the AES42specification of a digital interface for microphones includesspecifics about the proposed XLD connector because SC-05-02 is to specify the connector. J. Nunn pointed out thatthe AES42 specification does not detail the connector butonly portrays its general appearance and use.

Brown advocated that the XLD be made a requirement forall newly designed products with AES3 interfaces. Cainecountered that the XLD connector should be only an alternativeand not a replacement for the XLR for digital applications.

Caine pointed out that connector specification details arenot in the scope of SC-02-02 but the group can discuss theuse of the proposed XLD. However, the group needs to seesome detail from SC-05-02.

Nunn reminded the group that the reason SC-02-02 is in-volved in the discussion of the proposed XLD connectorbefore it has been specified by SC-05-02, is to provide as-surance that the connector was needed and would be used.This assurance would provide encouragement for manufac-turers to develop it and the specification required by SC-05-02. For this to happen there would need to be some assurancethat the XLD would be permissible for use with AES3.

The chair advised the group that S. Harris had resigned asthe leader of the task group, SC-02-02-F, and that the taskgroup was looking for a new leader. Brown volunteered.

The discussion ended with a series of questions about theXLD being proposed to the task group. For equipment withAES3 interfaces, is it forbidden, permitted, preferred, op-tional, or required? Should it always, or optionally, besupplied with removable keys to allow it to become inter-mateable with the XLR? If XLDs with removable keys aresupplied should it be required that equipment is suppliedwith the keys present?

New BusinessIEC 60958-4 is the IEC document parallel to AES3. Its latestdraft, 100/396/CDV, is currently out for vote. AES canpropose a new diagram for figure 1 in that standard to besimilar to the latest draft from project AES3-R.

The IEC document is at the last stage, committee draft forvoting (CDV), during which new input can be provided butit can only be informative. It was agreed that a response ofWG experts would be discussed on the reflector.

[Secretariat note: The IEC was informed of the newdiagram.]

A. Mason asked that a copy of ITU-T V.11 be obtainedfor the WG. It is referenced in both the IEC and AESstandards. It may address some of the concerns with regard

to IEC 60958-4 and AES3 in connection with the examplecircuit.

IEC 60958-1 maintenance is being initiated. Thedocument MT60958-1(Dunn/Yoshio)20011220 contains adraft being prepared for circulation as a committee draft(CD) for amendment or revision of this document. This wasbased on the document presented to the previous meeting ofSC-02-02. Yoshio-san gave some explanation of the devel-opment of the text since that meeting.

Four new informative annexes are being proposed:a) to clarify the relationship between various application

standards and the state of the first two channel status bits;b) to show how the application standards relate to each

other;c) to indicate a potential problem with compact discs that

are encoded with non-linear pulse-code modulation data;d) to reference the other application standards using IEC

60958.This IEC project was to be discussed at a closed meeting

of the IEC maintenance team, MT 60958-1, on the followingday. In order to foster liaison, three members of the AESSCSC-02-02 were invited to attend that meeting as observers.

Nunn asked that a mechanism be developed to prepare aworking group response because some of the new annexescan be very controversial. M. Yonge reported on a meetinghe had with the secretary of IEC TC100/TA4 responsible forIEC 60958 maintenance. He gave an outline of how a liaisonmay be developed so that the liaison is between experts, andwhere expert opinion is required there should be joint mem-bership of groups.

New projectsNo project requests were received or introduced.

The next meeting is scheduled to be held in conjunctionwith the AES 112th Convention in Munich, Germany.

Report of the meeting of the SC-03-04Working Group on Storage and Handlingof Media of the SC-03 Subcommittee onthe Preservation and Restoration of AudioRecording, held in conjunction with theAES 111th Convention in New York, NewYork, US, 2001-11-30Chair T. Sheldon convened the meeting. The agenda and thereport from the previous meeting were approved as written.

Open projects

AES22-R Review of AES22-1997, AES recommendedpractice for audio preservation and restoration—Storageof polyester-based magnetic tapeThere were no proposals for change in this document.

AES28-R Review of AES28-1997, AES standard foraudio preservation and restoration—Method for estimating life expectancy of compact discs (CD-ROM),based on effects of temperature and relative humidityNo action was taken.

J. Audio Eng. Soc., Vol. 50, No. 1/2, 2002 January/February 269


AES35-R Review of AES35-2000, AES standard foraudio preservation and restoration—Method for esti-mating life expectancy of magneto-optical (M-O) disks,based on effects of temperature and relative humidityThere were no proposals for change in this review project.

Discussion again noted the need to add a section on lightsensitivity. Increasing evidence is accumulating to confirmthe negative influence of light on digital disks. D. Schüllerindicated that he would investigate putting together a plan toinvestigate the light sensitivity issue.

AES38-R Review of AES38-2000, AES standard foraudio preservation and restoration—Life expectancy of information stored in recordable compact disc systems—Method for estimating, based on effects of temperature and relative humidityWhile there has been no discussion thus far, the absence ofany statements on the light sensitivity issue was again noted.It was agreed that a section on this subject needs to be addedto this document.

AES-X51 Procedures for the Storage of Optical Discs,Including Read Only, Write-once, and Re-writeableThe group considered the publication of ISO 18925 by ISOTC 42. If no special professional needs—not covered by theISO document—are identified, the project may be retired.

AES-X54 Magnetic Tape Care and HandlingThe WG considered the single suggestion it had receivedfollowing the PCFC draft of the Care and Handlingstandard. Discussion occurred regarding slotless hubs. Itwas agreed that S. Winner will draft text to be added as newclause 4.6.5.

AES-X55 Projection of the Life Expectancy of MagneticTapeThe explorations of the Joint Technical Commission (JTC)(see liaison project AES-X80) were reviewed. Several possiblemethodologies were mentioned which may hold promise, al-though more work is needed. Schüller stated that in IASA anapproach will be made to improve relations with the tape in-dustry to encourage tape manufacturers to warn users aboutany problems arising with specific tape types and formulations.

AES-X80 Liaison with ANSI/PIMA IT9-5The liaison arrangement between SC-03-04 and the IT9-5group in the International Imaging Industry Association, anorganization maintaining the secretariats of several ISO TC42 working groups through the American NationalStandards Institute, was reviewed as it is practiced in theJTC. It was noted that some of the work writing the draftstandards for SC-03-04 is being done in the JTC in the roleof task group SC-03-04-L. The group also addressed the re-lationships within in the JTC and the relationships betweenparallel documents of ISO TC 42 and IEC TC 100. It wassuggested that the JTC meet under the auspices of theAESSC in the fall of 2002 with the hope of improving inter-national participation in JTC writing activities.


New businessThere was no new business.

The next meeting of SC-03-04 is scheduled to be held inconjunction with the AES 112th Convention in Munich,Germany.

Report of the meeting of the SC-03-06Working Group on Digital Library andArchive Systems of the SC-03 Subcom-mittee on the Preservation and Restora-tion of Audio Recording, held in con-junction with the AES 111th Conventionin New York, New York, US, 2001-11-30

The meeting was convened by T. Sheldon, chair of the AESSC-03-06-A Task Group on Metadata for Audio.

The agenda and the report of the previous meeting werereviewed and approved with one minor spelling correction.

Open projects

AES-X98 Review of Audio Metadata SC-03-06 Assignedto SC-03-06-AD. Schüller urged the importance for audio engineers to beinvolved in developing descriptive metadata. A discussionbrought out the need to harmonize the approaches of knowl-edgeable people in different fields (for example, librarians,broadcast archivists, audio specialists) and to take intoaccount different organizations and their needs (includingEBU, IASA, Scandinavian A/V Metadata Group). The groupaffirmed the goal of working toward a harmonized defi-nition of descriptive and administrative metadata for audio.

W. Sistrunk reviewed her draft of descriptive metadata foraudio, presenting a revised draft that sought to incorporateapproaches from the Scandinavian A/V Metadata Group andEBU. She will make minor revisions in the light ofcomments, and the draft will be placed on the WG FTP site.

D. Ackerman presented ideas on administrative metadata,specifically core audio metadata and processing historymetadata. He will prepare two PTDs as a result of themeeting discussion and place them on the WG FTP site.

AES-X99 Transfers to Digital StorageNo action was taken.

AES-X100 Asset ManagementNo action was taken.

AES-X120 Liaison with International Association ofSound and Audiovisual Archives IASANo action was taken.


New businessThere was no new business.

The next meeting of SC-03-06 is scheduled to be held inconjunction with the AES 112th Convention in Munich,Germany.




T he AES 21st International Conference, ArchitecturalAcoustics and Sound Reinforcement, will be the firstAES event ever to be held in Russia, thanks to the ef-

forts of Chair Nickolay Ivanov. Recent developments indigital signal processing have lead to new approaches for ar-chitectural acoustics and sound reinforcement systems.Solving the complex problems of sound image simulation,transmission, and reproduction requires familiarity with nu-merous audio techniques: simulation, auralization, adaptivedigital signal processing, analysis of subjective criteria ofaudio-visual images, perception, features of transferring andrecording three-dimensional images, equipment characteris-tics, and evaluation. The AES 21st International Conferencewill cover all of these topics.

The conference will take place June 1-3 in the gloriouscity of St. Petersburg. Founded by Peter the Great on a seriesof islands at the eastern end of the Gulf of Finland, St. Pe-tersburg rivals Amsterdam as the “Venice of the North.” Thecity is home to the Hermitage and more than 50 other muse-ums. Renowned for its ballet and classical music, St. Peters-burg is one of the cultural capitals of the world, giving us

such illustrious artists as Tchaikovsky, Shostakovich, Di-aghilev, Gogol, Dostoevsky, and Pushkin.

AMBITIOUS TECHNICAL PROGRAMScientific Committee Chair Irina Aldoshina and Papers ChairNatalia Tyurina have organized an impressive program of over70 papers and poster presentations. The conference opens onSaturday morning with a session of four invited papers: “Ar-chitectural Acoustics in Russia” by Michael Lannie; “SomeRules and Methods for Creation of Surround Sound” by An-drzej Czyzewski and Piotr Odya; “Mid/Side Boundary Micro-phone Technique for Live Staged Performances” by Ron Stre-icher; and “Cinematographic Audio Equipment Developmentin Russia” by K. G. Ershov.

The first session on Saturday afternoon, Sound Rein-forcement, Part 1, opens with an invited paper by Mar-shall Buck, “Dual-Range Horn with Acoustic Crossover.”Another invited paper, “Acoustic Design of the GreatPhilharmonic Hall in the Moscow International MusicDome (MIMD)” by Wolfgang Ahnert, Lev Bonsov, andChristofor Shirjetdki, leads the second session on Satur-

Architectural Acoustics and SoundReinforcement


day afternoon, Architectural Acoustics, Part 1.Sunday will be the busiest day of the conference, with

dual papers tracks throughout most of the day. In the morn-ing sessions 16 authors will be making poster presentationsin the posters area, at the same time as Sound Reinforce-ment, Part 2 and Room Auralization take place in the papersarea. The afternoon sessions will have Transducers in onetrack, and a second track with Psychoacoustics, Part 1, Bin-aural and Transaural Stereophony, and Wave-Field Synthe-sis. The invited paper “Loudspeaker Placement for En-hanced Monitor Sound Field and Increased PerformerSource Positioning” by Thomas Lagö will be the first paperin the session on sound reinforcement.

Monday morning will feature Architectural Acoustics,Part 2 and Psychoacoustics, Part 2. The afternoon sessionwill be Linear and Nonlinear Digital Processing. A calen-dar, complete list of papers with abstracts, and a registrationform follow on pages 276-289.

CENTER-CITY VENUEThe venue for the conference will be the 3-star Hotel

Moscow situated on the eastern end of Nevsky Prospekt onthe bank of the Neva River, next to the Alexander NevskySquare Metro station and other public transportation. Astroll on Nevesky Prospekt, the main avenue of the city,takes you from the Palace Embankment, with St. Isaac’sCathedral, the Winter Palace and Hermitage, and the Admi-ralty, on its western end past such architectural wonders asthe Stroganov Palace, Kazan Cathedral, and the shoppingarcade of Gostinyy Dvor to the Hotel Moscow and Alexan-der Nevsky Monastery on its eastern end. Other nearby cul-tural attractions include the Russian Museum and the world-renowned Mariinskiy (Kirov) Opera and Ballet.

Registration includes events in the evening to relax and so-cialize: a welcome reception on Friday (May 31), a boat tripon Saturday, and a banquet on Sunday. Attendees should try toadd extra days before or after the three days of the conferencein order to sample the cultural and architectural splendors ofSt. Petersburg. Technical tours are planned for the large filmstudio Lenfilm, the Ice Palace and its advanced sound-rein-forcement systems, as well as a tour devoted to the acousticsof old Russian palaces. For more details see www.aes.org.

St. Petersburg, RussiaJune 1–3, 2002

St. Petersburg sites, clockwise from left:Grand Cascade at Peterhof; Church onSpilled Blood; Alexander Column andGeneral Staff Building in Palace Square.

AES 21st International ConferenceArchitectural Acoustics and Sound Reinforcement

2002 June 1–3 • St. Petersburg, Russia

08:0008:3009:0009:3010:0010:3011:0011:3012:0012:3013:0013:3014:0014:3015:0015:3016:0016:3017:0017:3018:0018:3019:0019:3020:0020:3021:0022:0023:00

SATURDAY, JUNE 1 MONDAY, JUNE 3

Session 13Architectural Acoustics,

Part 2

Session 15Linear and Nonlinear

Digital Processing

Session 14Psychoacoustics,

Part 2

Closing Ceremony

Lunch

Lunch

Session 10Binaural and Transaural

Stereophony

Session 12Wave Field Synthesis

Session 11Transducers, Part 2

Boat Trip

Session 2Sound Reinforcement, Part 1

Registration

Registration

Session 6Room Auralization

Session 5Posters, Part 1

Lunch

Session 9Transducers, Part 1

Session 8Psychoacoustics,

Part 1

Opening Ceremonyand

Invited Papers

Session 4Sound Reinforcement,

Part 2

Welcome Reception

Session 3Architectural Acoustics,

Part 1

Banquet

FRIDAY, MAY 31 SUNDAY, JUNE 2

This schedule reflects accurate information as of press time.

Session 7Posters, Part 2


Technical Sessions

Information is accurate at press time.

Saturday, June 1 11:20 am–1:30 pm

SESSION 1: INVITED PAPERS

1-1 Architectural Acoustics in Russia—Michael Lannie,Research Institute for TV and Radio, Moscow, Russia

A historical review of the architectural acoustics in Russiais presented. Three periods are explored: 1930 to 1949,1950 to 1989, and 1990 to the present. Three main top-ics are reviewed for each period: scientific studies, mea-surement techniques, and practical works on acousticconsulting. Attention is primarily paid to various theaters,concert halls, studios, cinemas, and sport halls whichhave been designed by the Russian acousticians. A de-tailed bibliography on the subject is given as well.

1-2 Some Rules and Methods for Creation of SurroundSound—Andrzej Czyzewski and Piotr Odya, TechnicalUniversity of Gdansk, Gdansk, Poland

The problem of selecting adequate surround sound liverecording and reproduction methods still exists. Alterna-tive methods of organizing this process are discussed.Some experimental recording sessions employing the5.1 format were made with the use of various mikingtechniques and the convolution-based multichannel au-dio processing algorithm. The results were submitted tosome subjective assessments and then compared.Conclusions resulting from performed experiments arediscussed.

1-3 Mid/Side Boundary Microphone Technique for LiveStaged Performances—Ronald Streicher, Pacific Audio-Visual Enterprises, Pasadena, CA, USA

Given the inherent restrictions on microphone place-ment when producing stereophonic recordings and/orsound reinforcement of live staged performances, it isoften difficult to achieve an accurate aural image andsense of movement across the stage. For many years,the use of spaced floor-mounted microphones has beenthe norm for these performances, resulting in acousticinterference anomalies that compromise the quality ofthese pickups. This presentation will demonstrate thatby employing the mid/side microphone technique in afloor-mounted “boundary array,” an accurate and articu-

late stage image can be achieved without any of thecomb-filtering or other phasing problems inherent tospaced microphones.

1-4 Cinematographic Audio Equipment Developmentin Russia—K. G. Ershov, St. Petersburg State Universi-ty of Cinema and Television, St. Petersburg, Russia

This paper discusses some issues pertaining to cine-matographic sound engineering audio equipment devel-opment in Russia, particularly work being done at theSaint Petersburg State University of Cinema and Televi-sion. The domestic development of sound track record-ing and playing started in 1926 and was completed intwo years. The first cinema featuring sound track equip-ment was opened in Leningrad in 1929. Later, many cin-emas were equipped with sound equipment. This re-quired special research devoted to the recording,throughout the country, of sound tracks (photographic) atfilm studios, design work, and industrial production of allrequisite equipment (microphones, mixing panels,recorders, etc.), as well as sound playing systems in-stalled at cinemas (sound units in projectors, electronicamplifiers, cinema high-power loudspeakers). Finally,many architectural acoustics issues had to be resolved tomeet the required standards in cinema halls.

2:20 pm–4:00 pm

SESSION 2: SOUND REINFORCEMENT, PART 1

2-1 Dual-Range Horn with Acoustic Crossover—MarshallBuck, Gibson Labs, Los Angeles, CA, USA (Invited)

A new approach has been developed to combinemidrange and high-frequency sound into the throat of ahorn designed for sound reinforcement. An acoustic low-pass filter element is interposed between the lower fre-quency passage and the higher frequency passage, sothat a smooth combination of the two frequency bands isachieved at the entrance to the horn bell. Thus each fre-quency band has nearly identical dispersion, and the twosources have equal delay.

2-2 Modifying STI to Better Reflect Subjective Impres-sion—Peter Mapp, Peter Mapp Associates, Colchester,Essex, UK

The Speech Transmission Index (STI) is becoming theuniversally accepted method for measuring the potential

AES 21th InternationalConference Program

Architectural Acoustics and Sound Reinforcement

2002 June 1–3 • St. Petersburg, Russia

21st International Conference Program


intelligibility of a sound system. However, a number ofoperating conditions and sound system characteristicsseem not to be taken into account by current STI tech-niques. This paper highlights a number of these condi-tions and discusses possible modifications to the STI inorder to improve its potential use and accuracy.

2-3 Effect of Multiple Diffuse Sources on the Gain beforeFeedback Margin in Distributed Sound Systems—Peter Mapp, Peter Mapp Associates, Colchester, Essex, UK; and Henry Azima, New Transducers Ltd.,London, UK

Acoustic feedback stability is a fundamental limitation ofpublic address, sound reinforcement, and teleconferenc-ing systems. In a previous paper it was shown that dis-tributed mode loudspeakers (DML) are potentially lesssusceptible to feedback than conventional devices. Thispaper extends the work to multiple sources, such asthose found in typical distributed sound systems. The re-sults again show DMLs offer greater gain before feed-back margin (GBFM). It is shown that the in-room soundfield produced by a DML is significantly different from aconventional cone radiator. Extensive correlation andspatial uniformity measurements are used to help ex-plore and understand the underlying feedback mecha-nisms responsible for such improvements. Contrary tonormal practice, it was found that the available gain actu-ally increased by employing more rather than fewer dif-fuse sources.

2-4 Loudspeaker Array Simulator with Coordinated Posi-tioning of Elements—Arkady Gloukhov, Consultant, St. Petersburg, Russia

A high throughput simulator for loudspeaker array model-ing and optimization has been developed. The arraymodel simulates mechanical links between cabinets. Thearray baffle surface is approximated by two second-orderequations. Splaying of an array is performed by variationof curvature coefficients. Displacement of the entire arrayand reorientation are performed by moving and aiming asingle cabinet. The simulator automatically finds coordi-nates, splaying and aiming angles using direct soundcoverage parameters as a target function. Interferencepattern calculation is used in automatic optimization ofdelays for comb filtering reduction. The rigging simulationmodule calculates center of gravity location and mechan-ical loads.

2-5 Distributed Sound Reinforcement for Multiple TalkerLocations—Michael Pincus, Acentech Inc., Cambridge,MA, USA

This paper describes techniques used for designing andimplementing a distributed sound system for the historicrenovation of a chapel in Concord, New Hampshire. Theproject is unusual because of the chapel’s unique seatingarrangement and multiple talker locations, several ofwhich are used simultaneously during an event. The sys-tem is designed to help the audience localize the talkers.The paper compares the system with a typical distributedsystem and shows how today’s digital signal processorsallow complicated sound reinforcement techniques to beimplemented easily and cost effectively, even for small-to medium-sized projects.

4:20 pm–6:00 pm

SESSION 3: ARCHITECTURAL ACOUSTICS, PART 1

3-1 Acoustic Design of the Great Philharmonic Hall in the

Moscow International Music Dome (MIMD)—WolfgangAhnert, ADA Acoustic Design Ahnert, Berlin, Germany;and Lev Bonsov and Christofor Shirjetdki, Research Insti-tute of Building Physics, Moscow, Russia (Invited)

In Moscow, the construction of an International MusicDome incorporating three halls is nearing completion.The Great Philharmonic Hall of the Dome seats 1800 lis-teners and excels in originality and outstanding architec-tural design. It is mainly meant for the performance ofclassical symphonic concerts without electronic supportof the performers. A special feature of the hall is that itcan be used for three different concert forms (organ,symphonic, and chamber) requiring different perfor-mance qualities.

3-2 Comparative Analysis of Two Acoustic SimulationSoftwares—Miguel Arana, Ricardo Sanmartin, AntonioVela, and M. Luisa San Martin, Universidad Publica deNavarra, Pamplona, Spain

A comparative analysis of the acoustic simulation pro-grams RAYNOISE and ODEON was carried out. The in-fluence of the modeling accuracy of the convergencewas also analyzed.

3-3 A Review of Room Acoustic Indicators and TheirSubjective Attributes—Satya Pancharatnam, IITMadras, Chennai, India; and Ramachandraiah Alur

For decades the pioneering work of W. C. Sabine (1898)and his formula for computing reverberation time (RT)stood as the sole numerical indicator of acoustical qualityin concert halls. Over the years, acousticians have devel-oped other objective room acoustic indicators—such asclarity factor (C80), bass ratio (BR), and strength—in or-der to evaluate the acoustics of existing concert halls orhalls being designed. In this paper a review of these ob-jective parameters and their subjective attributes is madewith a brief discussion of the application of neural networkanalysis in the evaluation of some of these parameters.

3-4 The New Symphony Hall in Las Palmas, Gran Canaria—Jan Voetmann and Lise-Lotte Tjellesen,DELTA Acoustics & Vibration, Kgs. Lyngby, Denmark

One of the world’s most beautiful new symphony halls isthe Auditorio Alfredo Kraus in Las Palmas, Gran Ca-naria, Canary Islands, inaugurated in 1999. For seriousreasons the hall was completed without the collaborationof the original acoustical consultant. Shortly after theopening of the hall, the acoustics had problems primarilyin terms of too long a reverberation time. A team ofSpanish and Danish consultants were brought in to workin close collaboration with the architects to find a solution.The solution included a number of measures: roomacoustics, in order to bring the reverberation time downand increase the projection of sound to the audience;and special measures for the members of the orchestras,in order to make them feel more comfortable on the(large) stage. In the autumn of 2001, the alterations werecompleted, and the new acoustics were very favorablyreceived by the public, the owners, and the orchestras.New measures to make the hall function for modernelectronic amplified music are under preparation. Consid-erations and measurements are presented.

3-5 A New Criterion for Concert Hall Loudness Evalua-tion—Shuoxian Wu, South China University of Technol-ogy, Guangzhou, People’s Republic of China

Loudness is one of the most essential parameters for



assessing the acoustical quality of an auditorium. Be-cause of the lack of an authentic criterion, how to evalu-ate the loudness in concert halls remains unresolved. Inthis paper LpF, the mean forte sound-pressure level oftutti sound, is suggested as a criterion to describe theloudness in a hall. The prediction procedure of the LpFvalue distribution in a hall is described. Comparison be-tween predicted and measured LpF levels is given. Ten-tative optimum and allowable LpF values for concerthalls are also discussed.

Sunday, June 2 9:00 am–10:40 am

SESSION 4: SOUND REINFORCEMENT, PART 2

4-1 Loudspeaker Placement for Enhanced MonitorSound Field and Increased Performer Source Positioning—Thomas Lagö, Jönköping University,Jönköping, Sweden (Invited)

When handling the electroacoustics in a church, thereverberation time often is large enough to make thereturned and delayed sound field irritating and confus-ing to the performer (typically a singer or talker). It canbe compensated for by using monitor loudspeakersplaced on stage, facing the performer. The sound fieldwill be reflected in the wall behind the performer andwill decrease intelligibility for the audience because ofthese reflections. If the wall behind the performer issoft or absorbent, it is not a problem, but in manychurches and auditoriums the podium and the wall be-hind the performer are hard. By mounting loudspeak-ers on the wall facing the audience, the monitoring as-pect can be resolved and other advantages can beachieved at the same time. Since the performer’ssound field is not very loud, the position is often givenby the loudspeaker system, thus negatively affectingthe localization of the performer. This will result in alower intelligibility, especially for people with de-creased spatial hearing. By using loudspeakers be-hind the performer, a first and well-defined wave frontis created. The next set of loudspeakers are then ad-justed to stay within the 10-dB sound level that is stip-ulated by Haas (the so-called Haas Effect), whichstates that the second wave front will not contribute todirection given that the sound field stays within about25 ms and 10 dB, measured at the listeners’ positionrelative to the first wave front. This effect can beachieved because the sound level and time delaystays within certain boundaries at most positions inthe auditorium. This idea has resulted in an approvedpatent and implementation in the Bankeryd’s Mission-skyrka Church in Sweden. A large measurement se-ries, consisting of several hundred measurements,quantifying the effect of this innovative placement ofthe loudspeakers was performed. This paper givesbackground information about the sound challengesnormally found in many churches and auditoriums andhow they can be handled by a different loudspeakerplacement. The paper also describes the results ac-complished and other possible side effects that couldoccur.

4-2 Theoretical vs. Practical Considerations for Field Deployment of Modular Line Array Systems—DavidScheirman, JBL Professional, Northridge, CA, USA

This paper begins with a review of market trends leadingto the availability and proliferation of modular multiway line arrays. Variously referred to as line arrays,curved arrays, line-source arrays or vertical arrays, suchsystems present opportunities to reliably predict cover-

age patterns and average level in the intended audiencearea. They can also present unique challenges for fielddeployment which are influenced by the mechanical de-sign. Such systems provide relatively narrow vertical cov-erage patterns and increased apparent gain at distance.These acoustical characteristics can be used to greatbenefit when the system is properly configured. This pa-per reviews the various individual box design attributesthat influence array performance. It then uses a casestudy approach to examine the practical aspects of de-ploying temporary systems in performance spaces anddiscusses various design tradeoffs encountered whenusing such systems in different types of venues.

4-3 Speech Reinforcement Inside Vehicles—Alfonso Ortega, Eduardo Lleida, and Enrique Masgrau, Universi-ty of Zaragoza, Zaragoza, Spain

Improving oral communication inside vehicles is the goalof a cabin car communication system (CCCS). Commu-nication can be difficult because of the distance amongpassengers, lack of visual contact between speakers,high level of noise, and many other factors. To achievespeech reinforcement, CCCS makes use of a set of mi-crophones to pick up the speech of each passenger,then it amplifies these signals and plays them backthrough the car audio loudspeaker system. This systempresents two main problems: electroacoustic couplingand noise amplification. To overcome these problems,CCCS makes use of an acoustic echo-cancellation sys-tem and a noise reduction stage. A brief description ofthe system and some results are provided.

4-4 Initial Investigation of an Air/Air Interface as anAcoustic Boundary for Sound Dispersion Control atOutdoor Events—David Carugo, University of Limerick,Newbridge, Ireland

A basic ray model is used to examine refraction and re-flection of sound from an air/air interface. The results ofthis treatment are presented and examined for the possi-bility of using a heated air mass, forming an air/air inter-face, as an acoustic boundary at outdoor events to re-duce environmental noise pollution from sound leakagefrom the event.

4-5 Implementation of Intelligibility Algorithms intoEASE 4.0—Wolfgang Ahnert, Stefan Feistel, and OliverSchmitz, ADA Acoustic Design Ahnert, Berlin, Germany

In acoustic simulation programs very different algorithmsare used to calculate the intelligibility of speech and mu-sic. To get results there are postprocessed fixed energyratios as well as time-dependent impulse responses. InEASE 4.0 there now are all the usual intelligibility mea-sures, derived by means of simulated high-resolutiondata or by applying statistical estimations. This papercompares all these measures and methods, such as STI,ALcons, clarity, definition, etc., by means of the resultsobtained within a common computer model. Finally, rec-ommendations are given for applications which shouldbe distinguished.

9:00 am–10:40 am

SESSION 5: POSTERS, PART 1

5-1 A Loudspeaker Selection for the Low-FrequencyAcoustic Enclosure Given—Victor Mazin, V&L Ltd., St.Petersburg, Russia

The paper deals with loudspeaker selection for a par-ticular acoustic enclosure already in use. Such a task



appears when an acoustic system’s woofer breaksdown and there is no opportunity to replace it with anidentical woofer of the same brand. In this case thereexists a problem when choosing a suitable one fromthe list of available units. The resulting acoustic systemshould have the required shape of frequency responsein the low-frequency region, the given half-power fre-quency, and be efficiency-matched with the adjacentband. The paper introduces the expressions forelectromechanical parameters of the optimal loud-speaker. Application opportunities of closed-optimalloudspeakers are discussed.

5-2 Nonlinear Model of Condenser Microphone Capsule—Shakir Vakhitov, Mikrofon-M Ltd., St. Petersburg, Russia

Physical factors that cause nonlinear distortion at differ-ent parts of condenser microphones are analyzed. De-tailed mathematical models of these phenomena determine dependence of distortions on acoustical, me-chanical, constructive, and electrical parameters. Thenonlinear models allow one to calculate fairly accuratelyfrequency- and level-dependent sensitivity and harmonicdistortion for capsules of different condenser micro-phones. The systematic nonlinear model of a capsule isobtained by incorporation and approximations of themost important factors. Comparison of the calculationand measurement results is presented. Recommenda-tions for nonlinear distortion reduction in the process ofmicrophone design are given.

5-3 Analysis of Deformation on Coated Paperboard dur-ing a Scratch Test by AE—Shigekazu Suzuki, YasushiFukuzawa, and Shigeru Nagasawa, Nagaoka Universityof Technology, Nagaoka, Japan; Hideaki Sakayori,Koutou Carving Co.; and Isamu Katayama, KatayamaSteel Rule Die Co. Ltd., Shinjukuk, Tokyo, Japan

Coated paperboard, an anisotropic composite material,is used as packing material. It is important to machinetest the deformation and fracture behavior of paper-board for practical work efficiency with little loss. In thisstudy deformation and fracture behavior of coated pa-perboard during the scratch test were investigated withan acoustic emission (AE) test. Acoustic emission sig-nals occurred mainly at each layer of exfoliation. The de-formation and fracture behavior of several kinds of coat-ed papers could be observed by an AE sensor duringthe scratch test.

5-4 Analysis of Sound Radiated by Paperboard Die Cut-ting—Akira Sadamoto,1 Takashi Yamaguchi,2 ShigeruNagasawa,2 Yasushi Fukuzawa,2 Daishiro Yamaguchi,3and Isamu Katayama3

1 Tsukuba College of Technology, Tsukuba, Japan2 Nagaoka University of Technology, Nagaoka, Japan3 Katayama Steel Rule Die Co. Ltd., Shinjukuk, Tokyo,Japan

This paper reports on the radiated sound that occurs un-der paperboard die cutting. For increasing productivityand reducing the operator’s task, any kind of automatictechnique for detecting cutting conditions is required. Tosolve this problem, the sound radiated in the cuttingprocess was analyzed. Several sounds were measuredby varying several conditions: cutting force, paper thick-ness, paper direction, and blade tip width. Sound-pres-sure level (SPL) denoted obvious differences in eachcondition. It was confirmed that the cutting conditioncould be diagnosed by seeing the SPL.

5-5 Research on Active Structural Acoustic Control byRadiation Modes—Mao Qibo, Institute of Acoustics,Nanjing University, Nanjing, People’s Republic of China

In this paper an active control strategy based on radiationmodes is presented for sound radiation from elasticstructures with an example of a simply supported rectan-gular panel. The physical characteristics and mathemati-cal meaning of the radiation modes were analyzed. Theradiation efficiency of the radiation mode falls off veryrapidly with an increase of modes order at low frequency.A new control strategy has been developed: by cancelingthe joint coefficient of the first k radiation modes, thesound powers of the first k radiation modes is theoretical-ly zero. The numerical calculation is made by usingpoint-force actuators as control forces.

5-6 Contradiction in Calculation of Residential BuildingsElevation Against Traffic Noise—Zoltan Hunyadi,BUTE Department of Building Construction, Budapest,Hungary

Since the protection of permanent residential spacesagainst traffic noise became an important issue, noisecontrol has taken priority in the designing of windowsover the design of massive wall surfaces. In the lasttwenty years the sound insulation properties of the mostcommonly used windows has increased. Post and sashprofiles—as well as sash locks and wind fillings appliedin modern windows—were the result of developing andeffectively promoting increasing sound and heat-insulat-ing properties. Actually, the sound insulation index ofmodern windows is determined by the arrangement ofglazing in the Rw=32 to 42-dB range. Altogether the sortof rare gas filled up the glass pane’s space of heat-insu-lated glass, decreasing the sound insulation index ofLow-E quality windows, and inversely the filled up gas ofhigher quality sound-insulated glass increased the heatloss of the window.

The sound reduction of block walls made of blockbricks developed in the last twenty years and in wide useis considerably underdeveloped from the sound reduc-tion properties of traditional masonry. Block bricks weredeveloped for their energy-saving properties, and thesound reduction properties of walls are limited to the lev-el of the window’s properties.

The object of the paper is to call attention to the up-to-date block brick product assortment. It is almost impossi-ble to find a suitable wall type (block brick wall) for noisyareas, especially if ventilating equipment for continuouscirculation of residential apartments is required. The sub-sequent construction complement (coating) of up-to-dateblock brick walls causes decreased sound insulation;therefore, it is reasonable to begin developing blockbricks suitable for external walls of residential buildings innoisy areas.

11:00 am–1:20 pm

SESSION 6: ROOM AURALIZATION

6-1 Reflection Thresholds for the Design of Real Roomsand Room Auralization Systems—Durand Begault,NASA-Ames Research Center, Moffett Field, CA, USA

Reflection threshold data are useful in the context of de-signing critical listening environments, since path lengthattenuation and absorption can make potential reflectionsinaudible. The audibility of both early reflections and thediffuse sound field is shown to be equivalent betweenreal-world and headphone simulations involving spatialaudio cues. Results indicate that early reflections are

inaudible when 21 to 30 dB below the level of the directsound path, as a function of signal type and time of ar-rival. The diffuse sound field thresholds range from –19to –35 dB relative to the level of the direct sound path, asa function of reverberation time and relative energy withineach octave band. These threshold data are useful fordetermining engineering parameters for real-time simula-tion of virtual acoustic environments, thus effectively con-serving computational resources.

6-2 Measurements of Church Impulse Responses Usinga Circular Microphone Array for Natural Spatial Reproduction of a Choir Concert Recording—Diemerde Vries,1 Sandra Brix,2 and Edo Maria Hulsebos1

1 Delft University of Technology, Delft, The Netherlands2 Fraunhofer Institute IIS/AEMT, Erlangen, Germany

In a church in Weimar, Germany, a 12-track recordingwas made of a choir concert. Instead of trying to includethe acoustics of the church in the recording, the impulseresponses were recorded separately using a new mea-surement technique in which a microphone slowlymoves along a circle. The microphone measures thepressure as well as the velocity response, enabling dis-crimination between wave field components from differ-ent directions and extrapolation of the data to other virtu-al microphone positions. This way the responses can beestimated at all listening places of interest and con-volved with the “dry” recording of the singers’ voices forreproduction by wave field synthesis. The measure-ments were done within the framework of the CAR-ROUSO project.

6-3 Preliminary Theoretical Investigation on the Relationbetween Some Physical Room Acoustical Mea-sures—Djamel Ouis, School of Society and Technology,Malmö University, Malmö, Sweden

This work is a preliminary investigation into the theoreti-cal evaluation and study of the relationship betweensome room acoustical descriptors. The study concernsthe interaural cross-correlation coefficient (IACC), theearly lateral energy fraction (ELEF), and the initial timedelay gap (ITDG). To this end, the impulse response fora rectangular room is calculated. Furthermore, and inview of a more realistic determination of the IACC, theimpulse response is convolved with the head-relatedtransfer function (HRTF) for both ears as measured us-ing a dummy head. It was found that simple analyticalexpressions can be established between the three parameters.

6-4 The Realisation of Ambisonics and AmbiophonicsListening Room for Car Sound Systems Evaluation—Lamberto Tronchin, Valerio Tarabusi, and AlessandroGiusto, University of Bologna, Bologna, Italy

Ambisonics playback is a very promising technique insound field reconstruction. Realistic acoustical environ-ments can be reproduced supplying the listener with thefeeling of being part of the reproduced acoustical field.The listener’s brain can feel the illusion of the reproducedsound scene as if it is there. In this paper the techniqueis analyzed and the main issue concerning Ambiophon-ics and B format are revised. In order to test the fidelity ofmusic and noise reconstruction, some B format trackshave been created in ancient churches and notableacoustical places along with the transfer function of themeasured acoustical environments. A dedicated room,called Arlecchino, has been simulated, calibrated, de-signed, and finally realized for the purpose of Ambio-

phonics listening tests, especially for car audio systemsimprovements. Eight diffusers were implemented in theroom in order to reproduce the eight-channel B formatsignal. The B format tracks were then tested in the listen-ing room. Simulation of the measured acoustical fieldswas obtained through convolution of “dry” anechoic mu-sical pieces recorded in B format. The main steps for de-signing the listening room are illustrated, and the Am-bisonics–Ambiophonics listening test results arepresented.

6-5 An Efficient Auralization of Edge Diffraction—TapioLokki,1 Peter Svensson,2 and Lauri Savioja1

1 Helsinki University of Technology, Espoo, Finland2 Norwegian University of Science and Technology,Trondheim, Norway

Principles and implementation of efficient auralization ofedge diffraction are presented. The calculation principlefor the impulse response from an edge is reviewed. Thetechnique has been integrated into an acoustic modelingsystem which is based on the image-source method. Forauralization purposes a low-order digital filter for each dif-fracting edge was designed, which efficiently implementsthe diffraction phenomenon and is suitable for parametricauralization. Finally, a comparison of auralized impulseresponses with and without diffraction is presented. Thecase study was made in a simple room geometry con-taining occluders.

6-6 Digital Filtering Techniques for Room ResponseModeling—Tuomas Paatero and Matti Karjalainen,Helsinki University of Technology, Espoo, Finland

Computationally efficient modeling of room responsesis needed in many audio and acoustics applications,such as auralization, artificial reverberation, and equal-ization of loudspeaker–room responses for sound re-production. Digital filtering is an efficient means forsuch modeling, particularly in real-time implementa-tions. This paper discusses new DSP-based methodsto model measured room responses. One technique isKautz filtering, which is an attractive method, especial-ly at low frequencies where the modal density is rela-tively low. Another approach is modeling dense modalpatterns by filterbanks that approximate the responsein a perceptually meaningful way. The optimization offilter parameters of the models is discussed; achievedperformance is shown by example cases; and applica-tions are briefly reviewed.

6-7 Hearing Missing Walls in a Multi-Loudspeaker RoomReverberation Simulation—Kenji Suzuki and WilliamMartens, University of Aizu, Aizuwakamatsu-shi, Japan

The goal of this research project has been to determinewhether a simple 3-D model for multiloudspeaker simula-tion of room reverberation could produce identifiable differences in room geometry. This simple, image-model-based simulation was designed to produce distinctive-sounding results as the material was varied on each ofthe six walls of a modeled rectangular room. A realistic-sounding wall reflection simulation was developed andsubmitted to blind listening experiments designed to testwhether listeners could determine which one of five wallshad been eliminated from the simulation. Though listen-ers were not particularly good at this identification task,they were able to consistently distinguish between thespatial images associated with these five cases (fiveroom geometries).





11:00 am–1:20 pm

SESSION 7: POSTERS, PART 2

7-1 Algorithms of Digital Audio Data Compression: Standards, Problems, and Perspectives of Develop-ment—Dhammika Priyadarsana Yatagama Gamageand Yurii Kowalguin, St. Petersburg State University ofTelecommunications, St. Petersburg, Russia

The main specifications of the algorithms of digital audiodata compression in MPEG and ATSC standards, whichinclude the newest hybrid methods that combine the ad-vantages of parametric and subband coding, are consid-ered. On the basis of analysis, a generalized structuraldiagram of a coder with digital audio data compressionis given. The procedure of the audio data processing inthe blocks of time–frequency segmentation, entropycoding, and psychoacoustic analysis is discussed. Spe-cial attention is given to the procedure of audio data processing in the psychoacoustic analysis block and directions for implementation.

7-2 Dynamic Behaviour of the Nonholonomic Mechani-cal Systems in Changeable Working Regime—Miodrag Zlokolica, Bogdan Sovilj, and Vladimir Miskov,University of Novi Sad, Novi Sad, Yugoslavia

In many complex mechanical systems, the transmissionswith nonholonomic characteristics as transmitters withchangeable transmission ratios are found. This type oftransmission has the connection of a differential characterbetween transmission elements. A nonholonomic me-chanical system can be recognized as variable-speed dri-ves, contained in many complex systems of modern tech-niques. The aim of this paper is to give one approach tothe dynamical description of a general example of trans-missions with nonholonomic characteristics and to esti-mate the stability of the working system with a change-able working regime. For the dynamic description of themechanical nonholonomic system, Appell’s differentialequations are used. By numerically solving the differentialequations of movement, the answer concerning the work-ing stability as well as the dynamical and kinematics be-havior of the observed system is provided. The obtainedresults will serve as one of the constraints in choosing op-timal parameters in the synthesis of power transmissions.

7-3 Subjective Evaluation of Concert Hall Acoustics forCarnatic Music by Regular Concertgoers—SatyaPancharatnam, IIT Madras, Chennai, India; and Ramachandraiah Alur

A series of questionnaire-type surveys was conductedduring the music festival in Chennai, South India, to de-termine the subjective response of regular concertgoersto the acoustics of concert halls for carnatic music (orSouth Indian classical music). This paper presents theresults and analyses of both the pilot survey and theseven main surveys.

7-4 Efficiency of Noise Barriers with Non-Straight EdgeProfiles—Henrik Sandqvist, Royal Institute of Technolo-gy, Stockholm, Sweden

The straight edge of a noise barrier in some areas be-hind a screen causes noise levels to increase instead ofdecrease. Still, the noise barriers today are commonlybuilt with straight edges. An exact analytical solution de-scribing the sound field for straight-edge noise barriersas well as barriers with periodical-edge profiles has beenpreviously derived. It has also been shown that for a giv-

en frequency, there is an optimum length of the period ofthe edge profile. Using these solutions, how to constructan efficient-edge profile for a broadband signal with a giv-en spectrum was examined.

7-5 Measurements of Scattering Coefficient of Surfaces in a Reverberation Room—Jin Jeon, Byung Kwon Lee, and Sung Chan Lee, Hanyang University, Seoul, Republic of Korea

Scattering of surface materials is one of the most impor-tant aspects for evaluating the acoustics of concert halls.One of the methods that can reduce the errors in calcu-lating the reverberation time and other acoustic parame-ters through computer modeling is to calculate the scat-tering coefficient of surface materials. However, so far,no objective and reliable method for measuring scatter-ing coefficient has been suggested. In this situation, theISO has suggested a method of measuring the random-incidence scattering coefficient of surfaces in a diffusefield; whereas the AES has introduced a method of direc-tional incidence in a free field. In this study the scatteringcoefficients of different hemispheres were measured byusing the ISO method in a 1:10 reverberation chamber.

7-6 A FEM Method for Calculating Acoustic Transmis-sion Function and Impulse Response in a LightlyDamped Room—Yuezhe Zhao and Shuoxian Wu,South China University of Technology, Guangzhou, People’s Republic of China

A finite-element method is presented for studying theacoustic transmission function and acoustic impulse re-sponse of lightly damped rooms. It is shown that thecomputer model successfully predicts the effects of dif-ferent source–receiver locations on the amplitude spec-trum. Also, the model solution does capture the effects ofdirect sound and reflections. As an example, a fullythree-dimensional rectangular room has been modeledwith details.

7-7 Influences of Parameters in Determination of Im-pulse Response Truncation Point—Dejan Ciric andMiroslava Milosevic, Faculty of Electronic Engineering,Nis, Yugoslavia

Precise determination of a theoretically obtained optimaltruncation point of an impulse response is more difficultin practice. Namely, this determination is rather sensitiveto influences of various parameters, such as averaginginterval, decay range for slope evaluation, range fornoise-level evaluation, etc. The importance of these influ-ences was investigated, and several algorithms for trun-cation point determination, including one proposed by theauthor, were implemented. Simulated and measured re-sponses were used for the investigations. The resultsshow that the mentioned influences can be importantand care should be given to the setting of correspondingparameters.

7-8 Acoustical Design of an Electrical Emergency Plantto Reduce Outdoor Noise Level—EvgueniPodzharov,1 Francisco de la Mora Galvez,2 and Lioudmila Oleinikova1

1 University of Guadalajara, Guadalajara, Mexico2 University Panamericana, Guadalajara, Mexico

An analysis of noise transmission in an electrical emer-gency plant was done using the statistical energy analy-sis method. This analysis permitted evaluation of differ-ent measures and materials to reduce noise level. A



two-inch-thick layer of fiber glass was selected as coatingfor the walls and ceiling and silencers at the inlet of airand at the outlet of engine gases to reduce indoor andoutdoor noise levels. The noise measurement showedthat the noise level was considerably reduced after im-plementation of these measures. The reduction of noisewas 7 to 8 dB(A) inside the plant, 19 dB(A) at 10-m dis-tance from the plant, and 23 dB(A) at 15-m distance fromthe plant.

7-9 A Problem of Distortions at Electroacoustic Conver-sion—Alexander Gaidarov, Andreev Acoustical Institute,Moscow, Russia

The most general principle of not distorting transforma-tion of information signals of the arbitrary shape is thescale copying by an output signal input with possible de-lay of an output signal on constant time for all signalcomponents. However, a number of views and interpre-tations in basic concepts (fundamentals of physics ingeneral and electroacoustics in particular and sufficiencyused in the substantiation of amplitude-spectral repre-sentation of quality of conversion) demand major rethink-ing. In particular: a) Parameters of motion of any body—speed and acceleration are determined in mechanics, asa derivative from already given displacement. Substan-tially in the dynamics of Newton, the external force gen-erates acceleration of mass. Speed and displacementare determined by a series integrating the acceleration oftime. Thus, the constants of integration complementingan original signal are foregone. The values of constantsdepend on the condition of the oscillating system to theinitial moment of arrival of the next signal, which upsetsthe invariance of a system concerning time. b) The defin-ition of intermodulation distortions in a nomenclature ofIEC, omissions of physical features of parametric distor-tions at Doppler intermodulation, is groundless for nonlin-ear distortions (sometimes referred to as subspecies).Presented is a revision and refinement of views along thepath that requires veracity of sound reproduction.

7-10 New Method for Assessing the Sound Quality of CarAudio Systems: Fast Sound Quality (FSQ)—D. Svoboda, Acoustics Center of the Broadcasting andElectroacoustics Department, Moscow Technical Univer-sity of Communications and Informatics, Moscow, Russia

A new method for subjective-statistical expertise ofsound quality, dubbed fast sound quality (FSQ), is pri-marily intended for sound quality judging at car audiocompetitions, providing reliable scores in shorter timescompared with traditional IASCA-based techniques.Testing software is compiled into a CD with a totalsounding time of less than 15 minutes. FSQ was usedsuccessfully in 12 Russian competitions in 2001. Thisnew method can also be used for assessment of homehigh-fidelity and high-end sound systems. FSQ and itsaccompanying software were introduced at an AESMoscow Section meeting and then written up in Av-toZvuk, the leading Russian car audio magazine, andMetrology and Metering Techniques Communicationsmagazine. The new FSQ method was developed at theAcoustics Center, Moscow Technical University of Com-munications and Informatics Department.

2:20 pm–3:40 pm

SESSION 8: PSYCHOACOUSTICS, PART 1

8-1 Psychological Factors Influencing Judgments onMusical Instrument Quality—Alexander Galembo,Setchenov Institute of Evolutionary Physiology, St.

Petersburg, Russia; and Anders Askenfelt, Royal Institute of Technology, Stockholm, Sweden

Several factors mostly of a psychological origin, not di-rectly connected with the musical instrument objectiveparameters, but influencing the performer’s and listener’sperception and judgment of the instrument quality, arediscussed. These factors—such as room condition, mul-timodality of perception, individuality of experience,nonordered terminology—are able to considerably lowerthe reliability of the estimation procedures commonlyused in the musical instrument industry.

8-2 Improving Perceptual Coding of Wideband AudioSignal Taking into Consideration a Temporal Mask-ing—Alexander Zakharenko and Yurii Kowalguin, St.Petersburg State University of Communications, St. Petersburg, Russia

All existing coding systems do not take into considerationthe temporal masking phenomenon. However, it plays avital part when decreasing bit rate. There are two kinds oftemporal masking: backward and forward masking. Al-though many studies of backward masking have beenpublished, the phenomenon is poorly understood; thehighly practiced subjects often show little or no backwardmasking. Therefore in this paper forward masking ap-plied to bit-rate reduction is considered. This paper de-scribes a new high-quality audio coding system based onthe MPEG ISO/IEC 11172-3 layer 3 codec. This newcoding system makes use of forward masking phenome-non to raise the coding efficiency and requires 11 to 20percent less bits than the conventional MPEG layer 3.

8-3 Audio-Visual Perception of Video and MultimediaPrograms—Nina Dvorko, St. Petersburg University ofHumanities and Social Sciences, St. Petersburg, Russia;and Konstantin Ershov, St. Petersburg State Universityof Cinema and Television, St. Petersburg, Russia

This paper discusses the results of theoretical and exper-imental research of psychophysical and aesthetic as-pects of sound and picture interaction. Perceptual experi-ments examine: 1) the influence of visual factors onthreshold sensitivity of hearing, 2) the role of associativelinks in audio-visual perception, and 3) the correlation be-tween sound and picture images in the perception ofspatial localization in multichannel sound systems.

8-4 A Special Form of Noise Reduction—Peter Swarte,Ing., Eindhoven, The Netherlands

Pop music reproduction or reinforcement in the entertain-ment world, such as at dance clubs or poppodia on a veryhigh sound-pressure level, is highly appreciated by theso-called target group. However for the neighbors, it canbe very annoying, especially when these music sessionstake place during the night. Poor sound insulation createsan inadmissible sound emission level in, e.g., bedrooms.Noise reduction methods of a constructional nature are inmost cases very expensive. Two methods of active noisereduction were tried out in the sound system of a popplatform in the Netherlands: one by anti-sound and theother based on the phenomenon of the missing funda-mental. Both experiments and the results are discussed.The latter experiment is called dormant bass (DB).

2:20 pm–3:40 pm

SESSION 9: TRANSDUCERS (FUNDAMENTALS, DESIGN,AND MEASUREMENT), PART 1

9-1 An Inexpensive Precise Passive Crossover



System—Neville Thiele, University of Sydney, Epping,NSW, Australia

This paper describes a crossover system for a two-wayloudspeaker in which the drivers are fed through conventional second-order passive filters, but the para-meters of the high-pass filter take into account the pa-rameters of the associated closed-back tweeter, to re-alize a desired overall fourth-order high-pass filteredacoustic output. When that output is combined withthat of the second-order low-pass filtered woofer, thesummed response is flat. With no impedance correc-tion required, the system produces an inexpensive butprecise crossover.

9-2 Sophisticated Tube Headphones for Spatial SoundReproduction—Klaus Riederer and Risto Niska, Helsinki University of Technology, Espoo, Finland

Custom tube headphones, fulfilling the high requirementsof accurate spatial sound perception experiments, arepresented. The UD-ADU1b headphones demonstrate amaximum ±5-dB deviation in the frequency band from 30Hz to 9 kHz, one-third octave smoothed. The ear canalblocking attenuates background noise typically at 15 to20 dB and allows a precise positioning of the soundsource. The nonmagnetic tubes are used in neuro- andpsychophysiological research.

9-3 The Influence of Losses on the Frequency Response of the Band-Pass Loudspeaker Systems—Andrzej Dobrucki, Wroclaw University ofTechnology, Wroclaw, Poland

The influence of acoustical losses upon the frequency re-sponse of fourth-order band-pass loudspeaker systemsis examined. It has been proved that losses in enclo-sures and in the vent can usually be negated. However,the leakage losses between both chambers of the sys-tem very strongly influence the frequency response. Therules for the corrections avoiding the differences betweenfrequency responses obtained for lossless and actualsystems have been developed.

9-4 Measurement of Loudspeaker Large Signal Perfor-mance—Comparison of Different Testing Signals—Alexander Voishvillo, Eugene Czerwinski, and AlexanderTerekhov, Cerwin-Vega Inc., Simi Valley, CA, USA

In this paper nonlinear reaction of low-frequency loud-speaker, horn driver, and free propagation to severaldifferent short-term signals has been investigated andcompared. These signals are single tone burst, multi-tone burst, spectrally shaped pulse, and burst ofGaussian noise. The advantages and drawbacks ofthese signals are discussed. The relationship amongthe magnitude of the voice-coil excursion, distortionlevel, and variation of excursion-dependent parametersis discussed. Various situations with different behaviorof excursion-dependent parameters are discussed. Inmeasurement of maximum sound-pressure level pro-duced by horn drivers, the distortion produced by theair propagation may affect the accuracy of measure-ment. Some examples of this effect are demonstrated.The difference between such aggregated criteria as re-action to multitone stimulus, incoherence function, andTHD is discussed. Multitone burst and Gaussian noisebursts seem to be optimal signals to measure maxi-mum SPL in loudspeakers because of the ability ofthese signals to excite a large number of intermodula-tion products.

4:20 pm–5:40 pm

SESSION 10: BINAURAL AND TRANSAURALSTEREOPHONY

10-1 Realisation of an Adaptive Cross-Talk CancellationSystem for a Moving Listener—Tobias Lentz andOliver Schmitz, DEGA, Aachen, Germany

The starting point of this paper is static crosstalk cancel-lation. The main task for an adaptive system is to updatethe crosstalk cancellation filter, depending on the listen-er’s position. The required filter is calculated at run timeof the program. Depending on the head position, theHRTFs required for the filter calculation will be selectedfrom a database. The conclusion of the preliminary lis-tening test is that the dynamic crosstalk cancellation pro-duces impressive results. The listener can move in anarea of about 1 m2. Head rotation is possible within theangle spanned by the loudspeakers.

10-2 Observed Effects of HRTF Measurement Signal Level—Agnieszka Jost, AuSIM Inc., Scotts Valley, CA,USA; and Durand Begault, NASA-Ames Research Center, Moffett Field, CA, USA

The effect of varying the signal level on the magnituderesponse of a head-related transfer function measure-ment was investigated. Measurement signals with levelsranging between 50- and 86-dB SPL were presentedover a loudspeaker and recorded using blocked meatusmicrophones placed in a dummy head. Results indicatethat relative to a 74-dB reference level for measurementsignals below 62 dB and above 80-dB SPL: 1) the ipsilat-eral ear shows attenuated spectral notches; and 2) thecontralateral ear demonstrates a 6- to 8-dB attenuation inbandwidths at 11.5-kHz and 16.5-kHz center frequencies.

10-3 An Objective Model of Localisation—Alois Sontacchi,Piotr Majdak, Markus Noisternig, and Robert Höldrich,University of Music and Dramatic Arts, Graz, Austria

A mathematical model is presented to objectively derivesound localization performance using head-related im-pulse responses (HRIR) based on binaural reproductionsystems. Rendering a sound source via panning meth-ods causes artifacts that will lead to errors in localizationby human subjects. Studying the relationship betweenpanning and perceived directions using listening testsentails an enormous effort of time. In addition, the pre-sented mathematical model can be used to minimize thenumber of parameters to evaluate through listening tests.Furthermore, the localization performance of severalHRIR-based panning methods were evaluated.

10-4 Elevated Control Transducers for Virtual AcousticImaging—Takashi Takeuchi, Kajima Corporation, Cho-fu-shi, Japan; Philip Nelson, University of Southampton,Highfield, Southampton, UK; and Martin Teschi, AKGAcoustics GmbH, Vienna, Austria

In order to examine the characteristics of various eleva-tion positions of the control transducers for binaural reproduction over loudspeakers, an analysis was per-formed for both the spectral and dynamic cues that relate to localization. The frequency response of the plantthat relates transducer outputs to ear pressure signalssuggests that control transducer positions will be promis-ing at positions in the frontal plane above the listener’shead. The analysis of the dynamic cues induced by un-wanted head rotation also strongly supports the use oftransducer locations in the frontal plane. A subjective



experiment was performed, and the control transducerlocation above the head clearly shows an advantage withrespect to the robustness against the transmission offalse dynamic information.

4:20 pm–5:40 pm

SESSION 11: TRANSDUCERS (FUNDAMENTALS,DESIGN, AND MEASUREMENT), PART 2

11-1 Low-Frequency Room Excitation Using DistributedMode Loudspeakers—Mark Avis, Bruno Fzenda, and W. J. Davies, University of Salford, Salford, GreaterManchester, UK

Conventional pistonic loudspeakers excite the modes ofan enclosed sound field in such a way that introducemodal artifacts, which may be problematic for listeners ofhigh-quality reproduced sound. Their amelioration mayinvolve the use of highly space-consumptive passive ab-sorptive devices or active control techniques. Other ap-proaches have concentrated on the design of the driverused to excite the room. Distributed sources rangingfrom the dipole to more complex configurations can beexpected to interact with the room eigenvectors in a com-plicated manner, which may be optimized in terms of thespatial and frequency-domain variance of the sound field.Recent interest in distributed sources has centered onthe distributed mode loudspeaker (DML). This paper re-ports on an investigation into the interaction of DMLs withmodal sound fields. It is shown that large DMLs can beexpected to modify the low-frequency sound field andthat smaller panels may interact with the room in interest-ing ways at higher frequencies. Producing useful low-fre-quency control remains difficult but can be achieved insome circumstances.

11-2 Problems of Theory and Designing for Directional Interference Microphones—Shakir Vakhitov, Mikrofon-M Ltd., St. Petersburg, Russia

A mathematical model and theoretical analysis of a di-rectional microphone that consists of an interferencetube and a pressure-gradient capsule are presented.The analytical expressions for angular dependence ofthe geometrical path length and directional characteris-tic were received. Physical reasons of the differencesbetween polar patterns of such microphones and sep-arate capsules in the low-frequency range were stud-ied. Dependence of the required rear aperture acousticresistance on the acoustic antenna length is shown.The reasons of the polar pattern axis asymmetry wereanalyzed. Theoretical principles are illustrated with ex-perimental data. Practical recommendations for suchmicrophone designs are given.

11-3 Three-Way Loudspeaker Systems—Ekaterinoslav Sirakov and Georgi Evstatiev, Technical University ofVarna, Varna, Bulgaria

This paper presents three-way loudspeaker systems. Itreviews the possibility of using compensating circuitsand mathematical and computer analyses when design-ing filtered loudspeaker systems, where the impedanceof the loudspeaker and the system keeps near to thenominal resistance. Original ideas for circuits needing tocompensate the electrical impedance are included.

11-4 A Problem of Efficiency of Loudspeakers—AlexanderGaidarov, Andreev Acoustical Institute, Moscow, Russia

Loudspeaker efficiency in a given frequency band is the

major characteristic of any electroacoustic transducer.Until now, there was only an approximated analytical ex-pression of this parameter suitable for use. The targetedsynthesis and optimization of devices with given spectralproperties was difficult to ascertain because efficiencywas hindered by the absence of the conforming analyti-cal software. The uniqueness and manifestive way of in-tercoupling the Thiele–Small parameters of drivers forloudspeakers with a flat-amplitude frequency responseused in acoustic closed-box enclosures has allowed theuse of this analytical expression for loudspeaker efficien-cy by the way products of dimensionless factors are im-plemented with an ideal limit. The obtained expressionhas a pictorial form and can be easily interpreted in aphysical sense. It also allows analysis of the actual prob-lem of optimization energy efficiency of loudspeakers.The primary analysis of technological problems of in-creased efficiency of loudspeakers and the developmentof a compromise between the degree of approximation toan ideal limit and capability of practical implementationare discussed.

5:40 pm–6:40 pm

SESSION 12: WAVE FIELD SYNTHESIS

12-1 Distance Coding in 3D Sound Fields—Alois Sontacchiand Robert Höldrich, University of Music and DramaticArts, Graz, Austria

This investigation proposes a method to synthesize 3-Dsound fields over loudspeakers taking distance codinginto account. The system can be divided into two parts:combining the benefits by using both the wave field syn-thesis (WFS) and the Ambisonic approach. In order tocode the virtual source distances, the driving functionsusing a derivative of the WFS approach were primarilycalculated. In the second step, the apparent solid angleof the sources were coded.

12-2 Drawing Quality Maps of the Sweet Spot and Its Surroundings in Multichannel Reproduction andCoding—Aki Härmä, Tapio Lokki, and Ville Pulkki,Helsinki University of Technology, Espoo, Finland

The sweet spot, or the optimal listening area in a room, isa central concept in multichannel audio reproduction.However, it is a difficult attribute to characterize in an ob-jective way. Ways to measure the quality of sound withina wide listening area are discussed, and a map of thesweet spot and its surroundings in a simulated listeningroom setup is presented. The proposed technique canbe used to evaluate and compare multichannel reproduc-tion systems and audio coding algorithms.

12-3 Spatial Audio Reproduction Using Distributed ModeLoudspeaker Arrays—Ulrich Horbach, Studer Profes-sional Audio AG, Regensdorf, Switzerland; Diemer deVries, Delft University of Technology, Delft, The Nether-lands; and Etienne Corteel, IRCAM, Paris, France

True spatial reproduction of sound images over a largelistening area can only be achieved by wave field synthe-sis, which requires a high number of individual loud-speaker channels. This paper describes a novel methodto design such systems in a practical way using multiex-citer distributed mode panels and digital filtering. Ex-plained in detail are filter designs for the reproduction ofplane waves, which are required to efficiently transportand render a wave field in a perceptual sense, and filtersfor the creation of focused sound sources behind or infront of the panels. For MPEG-4 applications, the display



of moving sound objects requires special algorithms togenerate and interpolate long impulse responses.

Monday, June 3 9:00 am–11:00 am

SESSION 13: ARCHITECTURAL ACOUSTICS, PART 2

13-1 Urban Sound Diffusion: Experimental Validation of aGeometrical Scattering Model—Philippe Woloszyn,CNRS, Nantes, France

Because the dimensions of the irregularities are compa-rable to the sound wavelengths, the major type of reflec-tions on buildings produces an acoustic interference fieldin the neighborhood. In order to create a model for thisfield, a unique measurement method of a neoclassicalfrontage using mathematical morphology techniques wasdeveloped. To perform an experimental validation of thismethod, an MLS multisensor measurement of a neoclas-sical frontage was calculated in Nantes. Comparison be-tween experimental measurement and morphologicalcharacterization shows a good agreement between theresults. Developed for architectural design tools for urbanacoustics, this method allows a good evaluation of theacoustical reaction of an urban surface by using the mor-phological attributes of an architecture.

13-2 Casa da Música, a New Concert Hall for Porto, Portugal—Laurentius van Luxemburg, Constant Hak,and Heiko Martin, Eindhoven University of Technology,Eindhoven,The Netherlands; and Ben Kok, and KjellBijsterbosch, Dorsserblesgraaf, Eindhoven, TheNetherlands

In Porto a new concert hall is under construction. Thenew hall is shoebox-shaped with specific solutions to en-sure sufficient strong lateral reflections. Acoustically, themain challenge is the front and back walls being madeentirely of glass, giving the feeling that the rooms areopen to the city. As well as keeping out noise from exteri-or sources, these transparent walls have to be designedsuch that they contribute to the sound distribution withinthe hall. With these considerations in mind, the glasswalls of the Casa da Música’s concert hall will have hori-zontally waved structures. The acoustical quality of thehall has been studied by using a simulation model and ascale model.

13-3 Acoustical Measurements of Courtyard-Type Tradi-tional Chinese Theater in East China—YenKun Hsu,Weihwa Chiang, and Jinjaw Tsai, NTUST, Taipei, Taiwan; and Jiqing Wang, Tongji University, Shanghai,People’s Republic of China

Acoustical measurements were taken at six courtyard-type traditional Chinese theaters in east China. The the-aters were generally rectangular in shape and some ofthem were built inside Chinese gardens. All measure-ments were taken in unoccupied conditions and in somecases with no seats. The theaters consist of a pavilionlikestage attached to a courtyard surrounded by coveredcorridors or buildings. Preliminary analysis showed anaverage strength (G) of 3.1 dB, an average early decaytime (EDT) of 0.8 s, and an average early support (ST1)of –9.4 dB. This research was the first step in a three-year ongoing project about traditional Chinese theater.Subjective assessments will be undertaken later.

13-4 Acoustics Design of the Music Suite in Taipei National Architectural University of Arts—WeihwaChiang, Liangkuang Yang, and Wenling Jih, NTUST,Taipei, Taiwan

An acoustical study of the music suite in Taipei NationalUniversity of Arts based on computer modeling and1:20-scale modeling was conducted. This paper reportson the design progress of a 7000-m3 concert hall and a1500-m3 orchestra rehearsal room. The shoebox-shaped concert hall (600 seats) was designed mainly forthe university orchestra, and the high ceiling (13.1 mabove the stage floor) was kept to prevent the hall frombeing overpowered. The stage was restricted to 155 m2

to provide good support for the musicians and good bal-ance and blend for the audience. The midfrequency re-verberation time (RT) was set to 1.7 s with a fully occu-pied audience and a sixty-member orchestra. Withdisconnected overhead reflectors 9 m above the stagefloor, preliminary tests in the scale model yielded an ear-ly support (ST1) of –13.1 dB at the solo location. The or-chestra rehearsal room was designed to provide bothadequate reverberation and strength of co-players. Thisobjective was achieved by implementing a coupledroom effect partitioned by an array of overhead reflec-tors hanging 5 m above the floor. Preliminary measure-ments with 55 upholstered seats yielded a ST1 of –12.9dB and a RT of 1.6 s. Future study will be conducted onthe field tuning and measurements after the completionof the building.

13-5 Room Acoustic Quality of a Multipurpose Hall: ACase Study—Maria Ribeiro, FEUP, Porto, Portugal

The paper describes the acoustic solutions defined toadjust room acoustic quality to the aesthetic demandsof architecture and the reduced budget available forconstruction of a high-standard, multipurpose hall.The paper presents the values predicted by computersimulation and the measured values of acoustical pa-rameters—such as reverberation times (RTs), earlydecay times (EDTs), central time (Ts), loudness(G10), clarity (C80), and definition (D50)—for the mainuses defined for the hall, i.e., cinema, conferences,and music presentations of very different styles rang-ing from classical music, jazz, and ethnical music topercussion groups.

13-6 Operational Methods of Forming Sound Field in theRoom—Y. P. Schevyev, St. Petersburg State Universityof Cinema and Television, St. Petersburg, Russia

A frequency reverberation method in a room with si-lencers, where the given characteristic of the silencers isknown in advance, is discussed. Results of the analyticalinvestigation of the multilayer silencers are given. Amethod of synthesis construction, which has a heteroge-neous material basis, is described. The wave resistancechanges toward the sound as the sound wave expands.A sound absorption system was developed to provide acalculation method. This compact, lightweight system ishighly effective.

11:00 am–12:20 pm

SESSION 14: PSYCHOACOUSTICS, PART 2

14-1 Subjective Validation of an Objective LocalisationModel—Alois Sontacchi, Markus Noisternig, Piotr Majdak, and Robert Höldrich, University of Music andDramatic Arts, Graz, Austria

A subjective validation of a mathematical model forcharacterizing binaural head-related impulse response(HRIR)-based reproduction systems is presented. Theevaluated sound localization performance is vali-



dated by an informal listening test. The experimentalsetup is depicted, and the statistical evaluation of theresults is given.

14-2 An Evaluation of Audio Warning Signals through Localization Behavior of the Eyes—Grigori Evreinovand Darius Miniotas, University of Tampere, Tampere,Finland; and Alexander Agranovski, SpecvuzavtomatikaDesign Bureau, Rostov-na-Donu, Russia

This study focuses on establishing the type of envelopean auditory warning signal should have in order to mini-mize its distracting effect on attention. Listeners’ ability tolocalize square wave and ramp spatial sounds was in-vestigated. Their performance was evaluated using anSMI EyeLink Gaze Tracking system. Both of the auditorystimuli could be localized equally well. The reaction timewas shorter for the square wave, but not significantly dif-ferent from the ramp condition. The ramp stimulus, how-ever, was reported by the participants to be more accept-able. The approach of using spatial sounds with agradual onset may be a reasonable option to considerwhen selecting the most effective shape for a warningaudio signal.

14-3 Expert Evaluation of the Quality of Images—NickolayKolomensky, St. Petersburg State University of Cinemaand Television, St. Petersburg, Russia

New integral and differential criterion and algorithmsfor evaluation of the quality of the image and sound ofaudio-visual systems based on the discovered psy-chophysical laws of single-line and nonlinear stochas-tic differential images of physical and touch spaces (in-stead of the known psychophysical Weber–Fehner andStevens’ laws) were studied and approved. The plural-probabilistic approach for axiomatic consideration ofthe theory of subjective evaluation of quality of the im-age and sound (in audio and video systems) was de-veloped. This approach uses a ring of ensembles withó-algebras and multivariate Gilbert’s touch spacestructure from the probabilistic approach by Kol-mogorov and uses the measure of ensembles from theundefined approach by Zade. Designed and approveduniversal integral and differential criterion (factor) ofthe expert evaluation of the quality of the image andsound using the multivariate indicative function wastaken into account as deterministic and casual in na-ture of the perception touch–perceptional images ofsignals of the image and sound.

14-4 The Effects of Classroom Acoustics on Perfor-mance, Perception, and Well-Being of Pupils—Markus Meis, M. Klatte, A. Schick, C. Janott, A. Uygun, and C. Nocke, University of Oldenburg, Oldenburg, Germany

The literature on classroom acoustics shows thatpupils and teachers are often forced to work underpoor acoustic conditions. Teaching and learning innoisy classrooms with high reverberation times aredifficult. The aim of this paper is to show the method-ology and the main results of several studies of ourworking group in Oldenburg, Germany. In the firststudy nine- to ten-year-old students subjectively eval-uated the hearing impressions of classrooms withgood and bad acoustical conditions, defined by fre-quency-dependent reverberation times. Furthermore,the effects of active and passive room-acoustical in-terventions were evaluated. The second is a laborato-ry study in which the effects of a sound field systemby means of a battery of cognitive tests for auditory

discrimination, working memory, and executing com-plex instructions were evaluated.

1:40 pm–3:00 pm

SESSION 15: LINEAR AND NONLINEAR DIGITAL PROCESSING OF MUSICAL AND SPEECH SIGNALS

15-1 Echo Compensation by Equalizer with Precise Spectrum Estimation—Andrey Barabanov,1 KonstantinPutyakov,2 Sergey Salischev,1 and Vasilij Sitnikov1

1 St. Petersburg State University, St. Petersburg, Russia2 Children School of Art, St. Petersburg, Russia

This approach is based on measurement of the audiosignal by two microphones that are posed at different dis-tances from the source of the signal. A new adaptiveequalizer was developed for echo compensation and op-timal signal reinforcement. This technique uses the noiseattenuation algorithm developed for the linear filteringmodel. Identification of the model parameters becomesthe main problem of the approach.

15-2 Q Factor Modification for Low-Frequency RoomModes—Mark Avis, University of Salford, Salford,Greater Manchester, UK

Low-frequency normal modes of an enclosed sound fieldintroduce unwanted frequency, spatial, and temporal arti-facts to reproduced electroacoustic signals. A novel con-trol approach has been reported based on an analyticalmodal decomposition, using a low-frequency sound fieldmodel in a one-dimensional environment formed fromthe sum of a number of second-order IIR filter sections.In this paper these techniques are applied to the low-fre-quency resonances of a three-dimensional test room. Itis shown that significant reductions in modal Q and cor-responding reductions in modal decay times can beachieved, leading to smaller low-frequency sound fieldvariance and decreasing audibility of time-domain modalartifacts.

15-3 Further Developments of Methods for Searching Optimum Musical and Rhythmic Feature Vectors—Bozena Kostek, Marek Dziubinski, and Pawel Zwan,Technical University of Gdansk, Gdansk, Poland

The aim of this paper is first to review recent develop-ments in the domain of musical information retrieval andthen to present some methods developed at the Soundand Vision Engineering Department of the Technical Uni-versity of Gdansk, Poland. Especially important for musicretrieval systems is to find an optimum musical andrhythmic representation. This was done using both statis-tical evaluation and soft-computing methods. Results ofthe performed experiments are shown, and conclusionsto the content of the feature vectors are discussed.

15-4 Non-Stationary Filtering Methods for Audio Sig-nals—John Sarris, Sotirios Dalianis, and George Cambourakis, National Technical University of Athens,Athens, Greece

This paper deals with filtering methods for audio signalsusing time–frequency analysis. The concept of time–fre-quency filtering is vital for enhancement of nonstationarysignals and systems. Time–frequency filtering is per-formed as masking or convolution in the time–frequencydomain and is based on nonparametric modeling usingdirect convolution or multiplication. Applications of filteringin the time–frequency domain, including artificial reverber-ation and sound source motion simulation, are presented.


Registration: Fee includes attendance at the conference, aprinted copy and a CD-ROM of the Proceedings of theAES 21st International Conference, a welcome receptionFriday evening (May 31), a boat trip on Saturday evening,and a banquet on Sunday evening. Lunch and coffeebreaks are included Saturday through Monday. If you planto bring an accompanying person to the welcome recep-tion, boat trip, or the banquet, there is an additional fee(payable on site). All visitors need a visa to enter Russia,and you will need a personal invitation to get a visa fromthe Russian Embassy or consulate in your country. You will

Please return by mail or fax to: Monomax Ltd., P.O. Box 168, 195112, St. Petersburg, RussiaFax: +7 812 324 73 22 • Tel: +7 812 320 01 17Email: [email protected] and [email protected]

Conference registration is also available on the website at http://www.aes.org.No registration is final until the conference fee has been received.

Please check appropriate box and total payment in U.S. dollars.

MEMBERS 490 NONMEMBERS 590 AUTHORS (MEMBERS) 390 AUTHORS (NONMEMBERS) 490 AES STUDENT MEMBERS 250

First Name________________________________Last Name_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

Affiliation (Company/Institution)_________________________________________________________________________________

Street __________________________________________________________________________________________________________

City________________________State_______________Postal Code____________Country _________________

Telephone INT+______________________Fax INT+___________________________

Email______________________________________________ AES Membership No. __________________________

Registration and information on the AES 21st International Conference2002 June 1–3 • St. Petersburg, Russia

1

2

Payment Modes (check box) Total Amount $_________________________

Bank transfer. Bank: Nordea Bank Finland Plc, 1200 Helsinki - WTS branch, Aleksanterinkatu 17, 00100, Helsinki, FinlandAccount number: MRITFIHH 120042-1343; Account holder: Morena International Services Inc.

Amex Mastercard/Eurocard Visa

3

Card Number

Month Year/

Cardholder's Name (print)_______________________________________________________________________

Signature of Cardholder_________________________________________________________________________

Please print/type all information as you wish it to appear on your badge.

need to fill out the visa support form available online atwww.aes.org. The venue for the entire conference is the 3-star Hotel Moscow, centrally located on Nevsky Prospekt,the main avenue in St. Petersburg. Single rooms areUSD110 and doubles are USD140 per day. The hotelbooking form is available online at www.aes.org. Pleasebook as soon as possible since June is the busiest touristseason. If you do not have access to the Internet, phone(+7 812 320 01 17) or fax (+7 812 324 7322) a request fora copy of the hotel registration. Complete travel details areavailable on the AES website at www.aes.org.

ARRIVAL: Date Time FlightDEPARTURE: Date Time Flight

The organizing committee will provide participants and accompanying persons with transportation from the airportto the hotel and back. Please supply your arrival/departure information below or email it to [email protected]

as soon as you schedule your flights.

Expiration Date


EEssppoooo,, FFiinnllaannddJJuunnee 1155––1177,, 22000022

Pho

togr

aph

of F

inni

sh N

atio

nal O

pera

and

Too

lo B

ay b

y H

aral

d R

aebi

ger:

Cou

rtes

y H

elsi

nki C

ity T

ouri

st &

Con

vent

ion

Cen

ter

AESAES 2222ndnd

IInternternatnatiioonalnal ConferConferenceence

FFor the second time in three years, the AES FinnishSection is offering an exciting and action-packedconference experience for the audio engineeringcommunity. The beautiful days and bright nightsof Scandinavian summer, on the shores of the

Baltic Sea, will set the atmosphere for the AES 22nd Interna-tional Conference, Virtual, Synthetic, and EntertainmentAudio. The conference will be held on the picturesque campusof the Helsinki University of Technology, located in the city ofEspoo, about 10 kilometers west of Helsinki.

This three-day conference will bring together researchersand developers in the field of virtual and synthetic audio aswell as entertainment audio technologies and applications. 3-Daudio and virtual acoustics are an active area of research, andwill be a topic for several interesting sessions. Sound synthesisand modeling techniques will be discussed, with special em-phasis on physical modeling of musical instruments. Researchproblems in audio coding, delivery and presentation, as well asauditory scene analysis will be presented. Furthermore, percep-tual issues and subjective testing will be discussed.

HIGH-QUALITY TECHNICAL PROGRAMHIGH-QUALITY TECHNICAL PROGRAMPapers Chair Vesa Välimäki has assembled a program of 46high-caliber papers and poster presentations. Saturdaymorning begins with Virtual and Augmented Reality, and aninvited paper by Peter Svensson and Ulf. R. Kristiansen,“Computational Modeling and Simulation of AcousticSpaces.” The afternoon session is Sound Synthesis, startingoff with the invited paper “Present State and Future Chal-lenges of Synthesis and Processing of the Singing Voice,”by Xavier Rodet. The final, late-afternoon session on Satur-day will be 3-D Audio Technologies, Part 1.

3-D Audio Technologies, Part 2, opens Sunday morning.The next session, Audio Coding Techniques, will begin withJürgen Herre’s invited paper “Audio Coding—An All-Round Entertainment Technology.” A group of demonstra-tions and poster presentations will then run through thelunch break and for most of the rest of the afternoon. Thesession Physical Modeling will complete Sunday’s program.

Subjective and Objective Evaluation will run throughoutthe morning on Monday. The final session, ComputationalAuditory Scene Analysis, opens with a lunchtime invited pa-per by a pioneer in spatial-hearing research, Jens Blauert’s“Instrumental Analysis and Synthesis of Auditory Scenes:Communication Acoustics.” The calendar, complete pro-gram with abstracts, and conference registration form followon pages 292-301.

THE FINNISH EXPERIENCE AND OTHERTHE FINNISH EXPERIENCE AND OTHERSOCIAL DIVERSIONSSOCIAL DIVERSIONSTo complement the technical program, the conference com-mittee, headed by cochairs Jyri Huopaniemi and NickZacharov, has put together an exciting and eventful socialprogram. A kickoff dinner for the conference will be heldon Friday evening June 14 at the Nokia Research Centerhead office in downtown Helsinki. On Saturday evening at-tendees will be immersed in a traditional “Finnish Experi-ence,” which will surely leave no one cold (hint: there are1.2 million saunas in Finland). A cruise through the Helsin-ki Archipelago will precede a banquet on Sunday evening.

The conference venue will be on the campus of the Helsin-ki University of Technology in Otaniemi, which is part of thecity of Espoo. Espoo is the second largest city in Finland, af-ter nearby Helsinki. This area has rapidly become the leadingtechnology center in northern Europe. There are two hoteloptions: the SAS Radisson Royal in downtown Helsinki(with bus service to the conference site) and the SAS Radis-son Espoo close to the conference center at the university. Besure to book early as June is the busy tourist season.

The volunteers from the AES Finnish Section coordinat-ing the many details of the 22nd Conference are the samededicated crew who staged the highly successful 16th Con-ference in 1999 in Rovaniemi, Finland, above the Arctic Cir-cle in Lapland. They are warming to the task of hosting an-other AES conference during the long, sunny days of June.So join your colleagues in Finland at an intellectually stimu-lating conference, but expect to have some fun also. Formore details and online registration see www.aes.org.


VViirrttuuaall,, SSyynntthheettiicc,, aannddEEnntteerrttaaiinnmmeenntt AAuuddiioo

AESAES 2222ndnd InterInternatnatiional Conal CoonnffeerrenceenceVVVViiiirrrrttttuuuuaaaallll ,,,, SSSSyyyynnnntttthhhheeeettttiiiicccc,,,, aaaannnndddd EEEEnnnntttteeeerrrrttttaaaaiiiinnnnmmmmeeeennnntttt AAAAuuuuddddiiiioooo

2222000000002222 JJJJuuuunnnneeee 11115555–––– 11117777 •••• EEEEssssppppoooooooo,,,, FFFFiiiinnnnllll aaaannnndddd

Closing

Session 53-D Audio Technologies, Part 2

Session 6Audio Coding Techniques

Session 9Subjective and Objective Evaluation, Part 1

Session 10Subjective and Objective Evaluation, Part 2

Session 7Demonstrations and posters

Session 2Virtual and Augmented Reality, Part 2

Session 43-D Audio Technologies, Part 1

Session 8Physical Modeling

Session 3 Sound Synthesis

WelcomeReception Banquet

The Finnish Experience

This schedule reflects accurate information as of press time.

Session 1Virtual and Augmented Reality, Part 1

FRIDAY, JUNE 14 SATURDAY, JUNE 15 SUNDAY, JUNE 16 MONDAY, JUNE 1709:0009:3010:0010:3011:0011:3012:0012:3013:0013:3014:0014:3015:0015:3016:0016:3017:0017:3018:0018:3019:0019:3020:0020:3021:0022:0023:00

Session 11Computational Auditory Scene Analysis


Technical Sessions

Saturday, June 15 9:00 am–10:30 am

SESSION 1: VIRTUAL AND AUGMENTED REALITY,PART 1

1-1 Computational Modeling and Simulation ofAcoustic Spaces—Peter Svensson and Ulf. R. Kristiansen, NTNU, Trondheim, Norway (Invited)

The computational modeling of acoustic spaces is fun-damental to many applications in auralization/virtualacoustics. The demands vary widely, from real-timesimulation in multimedia and computer games to non-real-time situations with high-accuracy needs, such asprediction of room acoustic conditions in music perfor-mance spaces. Acoustic spaces include single roomor multiroom spaces, with simple or complex geome-tries and boundary conditions. Outdoor spaces canrange from city environments to open landscapes.Sound transmission through partitions is an importantissue in some cases. This paper gives an overview oftechniques used in the various auralization applica-tions. Aspects of accuracy and computational efficien-cy are discussed, as well as which acoustical phe-nomena can and can not be modeled with varioustechniques.

1-2 Better Presence and Performance in Virtual Environments by Improved Binaural Sound Rendering—Pontus Larsson, Daniel Västfjäll, andMendel Kleiner, Chalmers Room Acoustics Group,Göteborg, Sweden

The development of virtual reality (VR) technologyhas mainly focused on creating sensory cues for thevisual modality. It is hypothesized that adding aural

sensory cues to a virtual environment (VE) raises theexperience of presence perceived by the user. Twoexperiments were carried out in order to investigatepotential benefits of using high-quality, congruent au-ditory rendering in VEs. In Experiment 1, 40 subjectswere assigned either to a unimodal (vision-only) or abimodal (vision and hearing) VE. The subjects per-formed two memory and navigation tasks in succes-sion. After the completion of the tasks, subjects ratedpresence, awareness of external factors, enjoyment,and simulation sickness. Completion time for bothtasks was measured. Statistical analysis showed thatthe auditory information yielded a significant effect inthe second memory task. Ratings showed that sub-jects in the bimodal condition experienced significant-ly higher presence, were more focused on the situa-tion, and enjoyed the VE more than subjects whoreceived unimodal information. In Experiment 2, 40subjects were assigned to either a high-quality aural-ization VE or a low-quality auralization VE. The sub-jects performed the same tasks and ratings as in Ex-periment 1. In addition, three sound quality itemswere used. Statistical analysis showed that subjectsin the high-quality condition experienced a higher de-gree of presence, perceived the sound as addingmore to the overall experience, and could more easilylocalize the sound. No significant difference in memo-ry performance was found. The results indicate thathigh-quality auditory information may greatly improvethe overall performance of the VE and, in some cases, improve task performance.

1-3 Authoring of Virtual Sound Scenes in the Contextof the LISTEN Project—Olivier Delerue and OlivierWarusfel, IRCAM, Paris, France

AES 22AES 22 ndnd InterInternationalnationalConferConferenencc e Pre Prograogramm

Virtual, Synthetic, and Entertainment Audio

2002 June 15–17Espoo, Finland

22nd International Conference Program


This paper presents ListenSpace, a graphical author-ing tool designed especially for audio-augmented reali-ties in the context of LISTEN, a European project (ISTprogram). LISTEN is devoted to the design, develop-ment, and experimentation of interactive soundscapescreated for augmenting an everyday environment. Thisnew type of multisensory content requires the designof a specific authoring tool, allowing the user to de-scribe, control, and test the soundscape. The papercompares different approaches for describing virtualsound scenes according to the specific context and de-scribes how the authoring tool has been designed tofulfill this goal.

11:00 am–12:30 pm

SESSION 2: VIRTUAL AND AUGMENTED REALITY,PART 2

2-1 Audio Interpolation—Richard Radke, RensselaerPolytechnic Institute, Troy, NY, USA; and ScottRickard, Princeton University, Princeton, NJ, USA

Using only the audio signals from two real micro-phones and the distance separating them, the audiothat would have been heard at any point along theline connecting the two microphones was synthe-sized. The method is valid in anechoic environ-ments. The interpolated audio can be calculated di-rectly, without the need to estimate the number ofsources present in the environment or to separatethe sources from the received audio mixtures. How-ever, additionally estimating the mixing parametersis shown to dramatically improve results for speechmixtures. Experimental results are presented, andsample sound files can be found on the authors’Web site http://www.ecse.rpi.edu/homepages/rjradke/pages/ainterp/ainterp.html.

2-2 Hybrid Sound Reproduction in Audio-AugmentedReality—Christian Mueller-Tomfelde, FhG-IPSI,Darmstadt, Germany

In this paper a reproduction of sound is proposed,which uses a combination of loudspeaker and in-earmonitors. This hybrid sound reproduction integratesthe advantages of a loudspeaker sound reproductionwith those of headphones. Especially for collabora-tive group situations using audio-augmented reality,this approach can help the users to have directundisturbed interindividual communication within theteam and, at the same time, a personalized aug-mented sound environment. Possible applicationscenarios for augmented reality are described; andan example for a hybrid sound reproduction for theapplication of room acoustics is presented and dis-cussed in detail. The measured impulse response ofa room is separated in the time domain to form adual-channel filter. The resulting processed signalsare sent to a public loudspeaker and a personal in-ear monitor system.

2-3 Perceptually Similar Orthogonal Sounds andApplications to Multichannel Acoustic EchoCanceling—Yi-Wen Liu and Julius O. Smith, StanfordUniversity, Stanford, CA, USA

In recent years, people who are interested in full-duplexed, multichannel sounds for telepresence

services have realized that the acoustic echo can-cellation (AEC) problem is mathematically ill-con-ditioned. To regularize the problem, it is knownthat signals in multiple channels must be decorre-lated. This paper proposes the use of decorrelat-ed background sounds for solving the AEC prob-lem. Methods to genera te a rb i t ra r i l y manyorthogonal but perceptually similar copies from amono sound are presented, and their effective-ness in two-channel and five-channel AEC is eval-uated in simulations.

1:30 pm–3:30 pm

SESSION 3: SOUND SYNTHESIS

3-1 Present State and Future Challenges of Synthesisand Processing of the Singing Voice—XavierRodet, IRCAM, Paris, France (Invited)

The synthesis of the singing voice has been a topic ofstudy for more than 35 years. Recent work shows thatthe musical and natural quality of singing voice syn-thesis has evolved enough for high-fidelity commercialapplications to be realistically envisioned. This paperbegins by presenting a brief historical perspective ofsynthesis methods, control strategies, and research inthis field. The synthesis methods, i.e., the synthesiz-ers that compute the sound signal samples, are thenpresented. These range from models of the physics ofthe vocal apparatus to models of the signal producedby human singers and voice processing techniques.The next section of the paper presents control strate-gies, rules, and a variety of recorded data, which areemployed to compute the parameter values for syn-thesizers. Different levels of such rules are needed,from the low-level rules describing details of articula-tion to those of a higher level that implement variousaspects of musical interpretation. Some aspects ofchoir singing synthesis are also considered. Recentresearch and accomplishments are presented as well.Future challenges include synthesizer model improve-ments, automatic estimation of model parameter val-ues from recordings, learning techniques for automaticrule construction, and gaining a better understanding ofthe technical, acoustical, and interpretive aspects of thesinging voice. Sound examples will be played at theconference.

3-2 Modeling Bill’s Gait: Analysis and ParametricSynthesis of Walking Sounds—Perry Cook,Princeton University, Princeton, NJ, USA

This paper presents algorithms and systems for auto-matic analysis and parametric synthesis of walkingsounds. A recording of walking was analyzed by ex-tracting the gait (tempo and left/right asymmetries),heel–toe events, etc. Linear prediction was used toextract the basic resonances. Wavelet decompositionwas performed, and a high-frequency subband wasused to calculate particle statistics to calibrate a parti-cle resynthesis model. Control envelopes were ex-tracted from the original sound. A real-time synthesisprogram allows flexible resynthesis of walking soundsof any length, controlled by a score extracted from asound file, a graphical user interface, or parametersfrom game/animation/VR data. Results for the analy-sis algorithm are presented for handcrafted experi-mental sounds of gravel.



3-3 Using Voice Segments to Improve ArtistClassification of Music—Adam Berenzweig andDaniel Ellis, Columbia University, New York, NY, USA; and Steve Lawrence, NEC Research Institute,Princeton, NJ, USA

Is it easier to identify musicians by listening to theirvoices or their music? The authors show that for asmall set of pop and rock songs, automatically locatedsinging segments form a more reliable basis for clas-sification than using the entire track. This suggeststhat the singer’s voice is more stable across differentperformances, compositions, and transformations be-cause of audio engineering techniques rather than theinstrumental background. The accuracy of a systemtrained to distinguish among a set of 21 artists im-proves by about 15 percent (relative to the baseline)when based on segments containing a strong vocalcomponent, whereas the system suffers by about 35percent (relative) when music-only segments areused. In another experiment on a smaller set, howev-er, performance drops by about 35 percent (relative)when the training and test sets are selected from dif-ferent albums. This suggests that the system is learn-ing album-specific properties possibly related to audioproduction techniques, musical stylistic elements orinstrumentation, even when attention is directed to-ward the supposedly more stable vocal regions.

3-4 Perceptual Resonators for Interactive Worlds—Dylan Menzies, South Bend, IN, USA

The simulation of sounds in an interactive world requires efficient and flexible algorithms, so that a potentially large number of different sounds can begenerated simultaneously in real time. Diffuse res-onators, such as wooden doors, pose a particularchallenge because conventional delay–feedback-based methods of simulation are relatively computa-tionally expensive. A perceptual resonator is present-ed as an alternative. In this the perceptually relevantcharacteristics of a resonator are simulated ratherthan all the low-level acoustics, gaining efficiency andcontrol at the expense of some loss of detail as wellas strict linearity. As a byproduct, the resonator formsan interesting tool for perceptual investigation.

4:00 pm–5:30 pm

SESSION 4: 3-D AUDIO TECHNOLOGIES, PART 1

4-1 An Algebraic Theory of 3D Sound Synthesis withLoudspeakers—Paul Flikkema, Northern ArizonaUniversity, Flagstaff, AZ, USA

The problem of reproducing J signals with K loud-speakers is considered. An algebraic, time-domainsynthesis approach, which extends the multi-inputmultioutput inverse theorem (MINT) of Miyoshi andKaneda, was developed. The approach is general andcan encompass joint surround sound synthesis andloudspeaker–room correction. First, a discrete time-domain matrix description was developed, which cap-tures the effects of amplifiers, loudspeakers, and roomacoustics. Based on this model, it is shown that exactsynthesis is possible with practical reproduction appa-ratus only if K > J. Sufficient conditions that are basedon impulse responses are also presented. The resultsare specialized to the case of zero-crosstalk transaur-al sound reproduction, and the theoretical importance

of loudspeaker time alignment is illustrated. Finally,minimum-power exact synthesis is briefly described.

4-2 Interaural Time Difference Estimation Using theDifferential Pressure Synthesis Method—YufeiTao, Anthony Tew, and Stuart Porter, University ofYork, York, North Yorkshire, UK

A differential pressure synthesis (DPS) method is pre-sented. This method facilitates rapid estimation ofacoustic pressures on the surface of the human headfrom the geometry of the head at low and midfrequen-cies. Interaural time difference is calculated from thephase of the pressures at ear canal entrance positions.It is shown that DPS reduces the ITD errors calculatedusing spherical head models from 1 kHz to 3 kHz.

4-3 Efficient HRTF Interpolation in 3D Moving Sound—Fábio Pacheco Freeland, Luiz Wagner PereiraBiscainho, and Paulo Sergio Ramirez Diniz, FederalUniversity of Rio de Janeiro, Rio de Janeiro, Brazil

This paper addresses efficient interpolation of head-related transfer functions (HRTFs) for binaural gener-ation of moving sound. Aiming at computational com-plexity reduction, the authors recently proposed a newmethod for HRTF interpolation based on auxiliary in-ter-positional transfer functions (IPTFs). This paper re-views this IPTF-based method and focuses on IPTFdesign, which employs tools such as balanced modelreduction and spectral smoothing to compute a com-plete set of reduced-order IPTFs. The interpolationscheme based on these reduced-order IPTFs is com-putationally more efficient and yields results percep-tively comparable to those attainable by the bilinearmethod, which uses the HRTFs directly. Therefore,the proposed method is promising for real-time appli-cations of spatial sound generation.

Sunday, June 16 9:00 am–10:30 am

SESSION 5: 3-D AUDIO TECHNOLOGIES, PART 2

5-1 A Balanced Stereo Widening Network forHeadphones—Ole Kirkeby, Nokia Research Center,Helsinki, Finland

The purpose of a stereo widening network for head-phones is to compensate for the nonideal listeningconditions experienced when listening to materialmixed for playback over two widely spaced loud-speakers. The network is balanced in the sense thatthere is a constraint on the sum of the magnitude re-sponses of the crosstalk (left input to right output andright input to left output) and the direct paths (left inputto left output and right input to right output). In order toensure that no noticeable spectral coloration is addedto the original sound, only frequencies below 2 kHzare processed. This ensures a very natural character-istic of the reproduced sound, and it makes the stereowidening algorithm well suited for applications to high-quality digital source material. A structurally balancedimplementation makes it possible to run the stereowidening network very efficiently in fixed-point precision.

5-2 Frequency-Domain Techniques for Stereo toMultichannel Upmix—Carlos Avendano and Jean-Marc Jot, Creative Advanced Technology Center,Scotts Valley, CA, USA

This paper proposes a series of upmixing techniquesfor generating multichannel audio from stereo record-ings. The techniques use a common analysis frame-work based on the comparison between the short-timeFourier transforms of the left and right stereo signals.An interchannel coherence measure is used to identifytime–frequency regions consisting mostly of ambiencecomponents, which can then be weighed via a nonlin-ear mapping function and extracted to synthesize am-bience signals. A similarity measure is used to identifythe panning coefficients of the various sources in themix in the time–frequency plane, and different map-ping functions are applied to unmix (extract) one ormore sources and/or to re-pan the signals into an arbi-trary number of channels. The paper illustrates the ap-plication of the various techniques in the design of atwo- to five-channel upmix system.

5-3 Wavelet-Based Spectral Smoothing for Head-Related Transfer Function Filter Design—HuseyinHacihabiboglu, Banu Gunel, and Fionn Murtagh,Queens University Belfast, Belfast, Northern Ireland

Three wavelet-based spectral smoothing techniquesare presented in this paper as a preprocessing stagefor head-related transfer function (HRTF) filter design.These wavelet-based methods include wavelet de-noising, wavelet approximation, and redundantwavelet transform. These methods are used with time-domain parametric filter design methods to reduce theorder of IIR filters, which is useful for real-time imple-mentation of immersive audio systems. Results of asubjective listening test are then presented in order tojustify the perceptual validity of the investigatedsmoothing methods. Results show that wavelet-basedspectral smoothing methods are beneficial in reducingthe filter order and for increasing the perception of lo-calization without introducing any noticeable effect ontimbre.

11:00 am–12:30 pm

SESSION 6: AUDIO CODING TECHNIQUES

6-1 Audio Coding—An All-Round EntertainmentTechnology—Jürgen Herre, FhG-IIS A, Audio &Multimedia, Erlangen, Germany (Invited)

Audio compression technology has become a rapidlydeveloping core technology within the recent decade.Starting out from early basic research, its potential hasreceived increasing recognition and has led to the wide-spread deployment of this technology in numerous nov-el applications; and many more exciting applicationsare still to come in the near future. This paper providesan overview of the most relevant aspects of the audiocoding (r)evolution and illustrates how audio coding hasevolved over time to accommodate the needs of a widerange of entertainment applications.

6-2 Post-Processing and Computation in Parametricand Transform Audio Coders—Michael Goodwin,Martin Wolters, and Ramkumar Sridharan, CreativeAdvanced Technology Center, Scotts Valley, CA, USA

This paper considers the relationship of audio codingand postprocessing, both of which are important inmodern audio applications. The work primarily ex-plores computational issues in parametric and trans-

form coders with postprocessing at the decoder. Instandard subband/transform coders, the signal repre-sentation does not readily support general audio mod-ifications. This inflexibility leads to a computationalburden at the decoder if postprocessing is desired, es-pecially in multiple stream scenarios. This problem isexamined, focusing on the basic modification of linearfiltering and on showing that parametric coders pro-vide a computational advantage over traditional sub-band/transform coders.

6-3 Coding Principles for Virtual Acoustic Openings—Aki S. Härmä, Helsinki University of Technology,Espoo, Finland

Acoustic opening is a multichannel audio communica-tions system. It consists of an array of microphones inthe transmitting room and an array of loudspeakers inthe receiving room. The goal is to provide listeners inthe receiving room an impression that there is only anopening, or a window, on the wall between the tworooms. This can be done using a large number of audiochannels and performing rendering using wave fieldsynthesis techniques. At present, an acoustic openingcan be built using standard audio components. Howev-er, the problem of how to code and transmit a hugenumber of highly correlated but nonidentical audio signals is almost untouched. This coding problem isstudied in the paper. A general framework is formulated,and some alternative techniques are studied.

12:30 pm–3:30 pm

SESSION 7: DEMONSTRATIONS AND POSTERS

7-1 Perceptual Evaluation of Static Binaural SoundSynthesis—Jean-Marie Pernaux, Marc Emerit,Jérôme Daniel, and Rozenn Nicol, France TelecomR&D, Lannion, France

This paper presents the psychophysical validation of ahead-related transfer function (HRTF) database, using“static” binaural technique (head movements are nottaken into account). After measuring the HRTFs foreight subjects with a high-spatial resolution (965 direc-tions), the first goal was to evaluate the performancesof static binaural synthesis using individual HRTFs.The long-term goal of this study is to use individualHRTF data as a baseline to perceptually evaluate var-ious HRTF models and individualization strategiesand to assess the contribution of other localizationcues (such as dynamic cues provided by head move-ments and reverberation). In order to compare the re-sults to previous studies, an absolute localization pro-cedure across 56 sound source positions covering thedirectional sphere was used. Eight subjects reportedthat they perceived the direction of virtual soundsources on a 2-D graphical interface. The collecteddata was then analyzed in terms of front–back confu-sion rate, average angle of error, and dispersion of thejudgments. It can be highlighted that subjects who hadtraining with feedback prior to the experiment showedvery good localization performance compared withnaïve subjects.

7-2 W-Panning and O-Format: Tools for ObjectSpatialization—Dylan Menzies, South Bend, IN, USA

Real acoustic objects have spatial width and charac-



teristic radiation patterns. Techniques are describedfor efficiently synthesizing these qualities by encodingwith spherical harmonics. This approach naturallylends itself to ambisonic reproduction, although it re-mains applicable to other forms of reproduction.

7-3 Compensating Displacement of Amplitude-PannedVirtual Sources—Ville Pulkki, Helsinki University ofTechnology, Espoo, Finland

The localization of amplitude-panned virtual sources isbiased toward the median plane when loudspeakersare not positioned symmetrically with the medianplane. The bias is measured using listening tests andcomputational modeling of virtual source perception. Amodification to an existing pair-wise panning methodthat compensates the displacement of virtual sourcesis proposed. The proposed method is evaluated by in-terpreting conducted listening test results and by sim-ulating virtual sources with computational modeling.Evaluations suggest that the bias is nonexistent withthe proposed method.

7-4 Perception-Based Design of Virtual Rooms forSound Reproduction—Renato Pellegrini, StuderProfessional Audio AG, Regensdorf, Switzerland

A perception-based parametric model to design audito-ry virtual environments utilizing a limited small numberof reflections and a model of diffusion was used to sim-ulate virtual environments in a binaural (two-channel)and a wave field synthesis (more than 28 channels) re-production system. Different perceptual parameters,such as the room size, the distance to the source(s),the positioning of the sources(s), the apparent sourcewidth, and more, can be adjusted by simple control pa-rameters. Rather than recreating a real room as accu-rately as possible, the main goal was to design nice-sounding environments for speech and musicreproduction. The achieved quality of the resulting vir-tual rooms is compared to the quality of measured realrooms.

7-5 The Signal Presentation Time for Subjects toLocalize a 3D Sound via Headphones—WithGeneral HRTFs—Fang Chen, Linköping University,Linköping, Sweden

This paper presents an experimental study on subjects’reaction time for accurately localizing the 3-D soundcontinuously presented via headphones to the subjects.The results show that for most of the subjects, it takesless than 11 seconds to localize the sound. Large indi-vidual differences were found, which may be due to theuse of generic HRTFs in the experiment. Localizationadaptation was found in the experiment.

7-6 Least-Squares Theory and Design of OptimalNoise-Shaping Filters—Werner Verhelst, KULeuven, Leuven, Belgium; and Dreten De Koning,Vrije Universiteit Brussels, Brussels, Belgium

Signal requantization to reduce the word length of anaudio stream introduces distortions. Noise shapingcan be applied to make requantization distortions min-imally audible. The psychoacoustically optimal noise-shaping curve depends on the time-varying character-istics of the input signal. Therefore, the noise-shapingfilter coefficients need to be computed and updated ona regular basis. A new theory for optimal noise shap-ing of audio is presented. It provides more straightfor-

ward proof of properties of dithered and nonditherednoise shaping. In contrast with standard theory, itdoes show how optimum noise-shaping filters can bedesigned in practice. Experimental results illustratethe method. The link with two popular noise-shapingrequantization techniques (F-weighted noise shapingand super bit mapping) is also discussed.

7-7 The Modified Discrete Cosine Transform: ItsImplications for Audio Coding and ErrorConcealment—Ye Wang and Miikka Vilermo, NokiaResearch Center, Tampere, Finland

This paper presents a study of the modified discrete co-sine transform (MDCT) and its implications for audiocoding and error concealment from the perspective ofFourier frequency analysis. A relationship between theMDCT and DFT via shifted discrete Fourier transform(SDFT) is established, which provides a possible fastimplementation of MDCT employing an FFT routine.The concept of time-domain alias cancellation (TDAC)and the symmetric and nonorthogonal properties ofMDCT are analyzed and illustrated with intuitive exam-ples. This study gives some new insights into innova-tive solutions for audio codec design and MDCT do-main audio processing, such as error concealment.

7-8 Improved Harmonic+Noise Model for Vocal andMusical Instrument Sounds—Shlomo Dubnov, Ben-Gurion University, Beer-Sheva, Israel

The paper presents an improvement to harmonic+noise models (HNMs) for quality analysis/synthesisof vocal and monophonic musical instrument sounds.So far these models are considered to not offer suffi-cient quality for modeling of singing voices or musical instruments. The problems with the HNM are present-ed. It is shown that the HNM can not handle specifictypes of musical instruments as well as highly rever-berant recordings. This paper presents a new analysismethod that significantly improves the HNM’s perfor-mance. The main difficulty in the estimation of the sto-chastic part of the signal has been to overcome usingnonlinear (polyspectral) methods.

7-9 An Evaluation of Three Sound Mappings throughthe Localization Behavior of the Eyes—GrigoriEvreinov and Roope Raisamo, University of Tampere,Tampere, Finland

Sonification of objects is an important research fieldthat allows us to present graphical forms to people whohave low vision or who are blind. This paper presentssome preliminary results from a study of an approachto the problem of objective recording of subjectivesound images. Sonification of the same graphic ob-jects is based on an amplitude-panning methodthrough three different sound mappings. Visualizationof auditory localization strategies was carried outthrough the SMI EyeLink Gaze Tracking system. Theevaluation showed that audio stimuli cause a smallerirritation than cognitive control of eye behavior in theabsence of visual stimulus. Therefore, the authors sug-gest that eye tracking can not be directly used as a crite-rion of sound mapping efficiency.

7-10 Production of Virtual Acoustic Guitar Music—Mikael Laurson, Sibelius Academy, Helsinki, Finland;and Vesa Välimäki and Cumhur Erkut, HelsinkiUniversity of Technology, Espoo, Finland





This paper presents an overview of research resultsdealing with the simulation of natural sounding guitarsynthesis. The paper starts with the description of thesynthesis model. Next, an analysis and calibrationprocess is described, which results in excitation sig-nals and synthesis data for the model. The papershows how the guitar model is controlled with the helpof an extended notation package. The system allowssimulation of various playing techniques used typicallyby a classical guitarist. The paper also presents howthe standard 6-string classical guitar can be extendedto a 24-string virtual super guitar. With this novel instrument complex musical textures that can not beplayed by a human player using a standard acousticalinstrument have been realized.

7-11 An FD Finite-Width Excitation Modeling Approachin Violin Sound Synthesis—Jan-Markus Holm,University of Jyväskylä, Jyväskylä, Finland

In this paper a new approach for a bowed-string excita-tion model in violin sound synthesis is represented. Amajor improvement in physical modeling sound synthe-sis is achieved by estimating the bow-hair ribbon of theexcitation area by computation of fractional delay (FD)coefficients representing the contact locations. Morecomplex behavior of bowed strings is briefly added.This computationally efficient method gives feasible re-sults for musical instrument sound synthesis.

7-12 Stratigraphical Sound in 4D Space—AndrewPaterson, Tillicoultry, Clackmannanshire, UK

This paper describes a framework to facilitate the au-thoring of sound with spatio-temporal relativity in virtualor augmented environments. It is based on thepremise that the interactor with the environment cre-ates an individual interpretation of “narrative” as a re-sult of his movement through the space, accumulatingfragments of stimuli or content to create meaning. In-spiration is gained from the actions of the archaeolo-gist, using stratigraphy as a recording practice thatrecords spacio-temporal paradigms. As the author of aspatialized soundscape composes a database of nar-rative fragments and clues for the user to find, a dis-tinct design process that reverses the excavation pro-cedure and re-imagines stratigraphical layers asphases of sound in the present tense can be used.

7-13 Interactive Granular Synthesis of Haptic ContactSounds—Stephen Barrass and Matt Adcock, CSIRO,Canberra, Australia

The goal of this project is to automatically synthesizecontact sounds, such as scraping in continuous contactwith a surface. This continuous information is muchricher than simply triggering a sound sample. The au-thors investigated an ecological granular synthesis al-gorithm and implemented it on a Windows PC using theOpenAL API and hardware acceleration on the soundcard. The algorithm was integrated into the ReachinHaptic API with haptic-audio latency below the 2 msthreshold of perception. Using this system, informationcould be heard, such as force, rate and timing of scrap-ing gestures on the surface, as well as properties, suchas stiffness, friction, roughness, and bumpiness.

7-14 Transmitting Audio Content as Sound Objects—Xavier Amatriain and Perfecto Herrera, UniversitatPompeu Fabra, Barcelona, Spain

As audio and music applications tend toward a higherlevel of abstraction and to fill in the gap between thesignal processing world and the end user, we aremore and more interested in processing content andnot (only) signal. This change in point of view leads tothe redefinition of several classical concepts, and anew conceptual framework needs to be set to givesupport to these new trends. In 2001, a model for thetransmission of audio content was introduced by theauthor. The model is now extended to include the ideaof sound objects. With these thoughts in mind, exam-ples of design decisions and possible implications arealso given.

7-15 Encoding and Rendering of Perceptual SoundScenes in the CARROUSO Project—RiittaVäänänen, Helsinki University of Technology, Espoo, Finland; and Olivier Warusfel, IRCAM, Paris,France

This paper presents virtual sound scene rendering anduser interaction aspects in the European Union projectcalled CARROUSO. It is an IST project used for record-ing, transmission, and rendering 3-D sound scenes inthe MPEG-4 format and for using the wave field synthe-sis (WFS) method. In the encoder side the soundsources are recorded as monophonic dry signals, andparameters describing the room acoustics of therecording environment are encoded as a separateMPEG-4 stream. In the decoder, the transmitted roomacoustic description defines how the sounds should beperceived in the virtual sound environment created atthe renderer. This scheme ensures that the roomacoustic properties of the sound scenes can be modi-fied interactively, and together with the WFS renderingtechnique, makes it possible to have interactive 3-Dmusic performances that are simultaneously audible tomany users. The paper presents the CARROUSOframework and concentrates on user interface and ren-dering issues in a case where the acoustic environmentis represented with a set of perceptual room acousticparameters.

7-16 A Psychophysical Investigation of the Frequency-Warping Coefficient—Melis Senova, Ken McAnally,and Russell Martin, Defence Science and TechnologyOrganisation, Melbourne, Australia

The effect of varying the frequency-warping coeffi-cient (lambda) on localization and discrimination per-formance was investigated in two experiments.Cochlear modeling was also conducted to investigatethe difference between cochlear output for warpedand linear frequency stimuli. Both methods suggestthat lambda values greater than 0.65 are not optimalfor auditory applications. The optimal lambda valuesfound using these methods (0.47, 0.55, and 0.24 forlocalization, discrimination, and cochlear modeling,respectively) are lower than the value that best fitsthe Bark scale (0.77). It is recommended that optimallambda values be determined by psychophysicalevaluation.

4:00 pm–5:30 pm

SESSION 8: PHYSICAL MODELING

8-1 Combining Digital Waveguide and FunctionalTransformation Methods for Physical Modeling of


Musical Instruments—Lutz Trautmann,1 BalazsBank,2 Vesa Välimäki,3 and Rudolf Rabenstein1

1 University of Erlangen-Nüremberg, Erlangen, Germany

2 Budapest University of Technology and Economics, Budapest, Hungary

3 Helsinki University of Technology, Espoo, Finland

Digital sound synthesis based on physical models isrealized in real-time applications mostly with the well-known digital waveguide method (DWG). It approxi-mates the underlying physical behavior of a vibratingstructure in a computationally efficient way. Becauseof these computational efficient approximations, thewaveguide method loses the direct connection to theparameters of the underlying physical model. The re-cently introduced functional transformation method(FTM) solves analytically the underlying physical mod-el. Thus, the physical parameters are explicitly givenin the discrete realization of the FTM. But because ofthis physicality the computational cost of synthesis us-ing FTM is larger than using DWG. This paper com-pares the DWG with the FTM and shows that for lin-ear vibrating strings it is always possible to design anacoustically indistinguishable DWG approximationwith the parameters obtained from the FTM. In thatway, a computationally efficient and physically mean-ingful synthesis method is obtained. Furthermore, thispaper shows the limits of this new synthesis method.

8-2 Virtual Strings Based on a 1-D FDTD WaveguideModel: Stability, Losses, and Traveling Waves—Cumhur Erkut and Matti Karjalainen, HelsinkiUniversity of Technology, Espoo, Finland

The one-dimensional digital waveguide structures basedon finite difference time-domain (FDTD) formulationsprovide a flexible approach for real-time sound synthesisof simple one-dimensional (1-D) structures, such as a vi-brating string. This paper summarizes the basic 1-DFDTD waveguide theory, describes the stability analysisof the model, and presents a sufficient condition for thestability. The simulation of frequency-independent loss-es is also covered. The formation of the traveling wavesand initialization of the 1-D FDTD waveguides was in-vestigated. The methods shown in the paper may beused to interconnect the 1-D FDTD waveguides to othermodel-based sound synthesis structures, such as digitalwaveguides that are based on the traveling wave solu-tion of the wave equation.

Monday, June 17 9:00 am–10:30 am

SESSION 9: SUBJECTIVE AND OBJECTIVE EVALUATION, PART 1

9-1 Static and Dynamic Sound Source Localization in a Virtual Room—Matti Gröhn and Tapio Lokki,Helsinki University of Technology, Espoo, Finland;and Tapio Takala, Nokia Ventures Organization,Helsinki, Finland

An audio localization test comparing accuracy of staticand moving sound sources was carried out in a spa-tially immersive virtual environment using a loud-speaker array with vector-based amplitude panningfor virtual sound sources. Azimuth and elevation errorin localization were measured. Different sound signals

were compared as well. As expected, errors in az-imuth localization accuracy were smaller than errors inelevation accuracy. There was more localization blurwith virtual sound sources than sound sources reproduced directly from a single loudspeaker. Movingsound sources as well as static panorated soundsources were localized. Although the sound sourcesmoved steadily, the measurements indicated that sub-jects perceived the changes in sound source locationstepwise.

9-2 Subjective Audio Quality Trade-Offs in ConsumerMultichannel Audio-Visual Delivery Systems, PartII: Effects of Low-Frequency Limitation—SlawomirZielinski and Francis Rumsey, University of Surrey,Guildford, Surrey, UK; and Søren Bech, Bang &Olufsen, Struer, Denmark

The subjective effects of controlled low-frequency lim-itation of the audio bandwidth on assessment of au-dio quality were studied. The investigation focused onthe standard 5.1 multichannel audio setup (Rec. ITU-R BS.775-1) and was limited to the optimum listeningposition. The effect of video presence on audio quali-ty assessment was also investigated. The results ofthe formal subjective test indicate that it is possible tolimit the low-frequency content of the center or therear channels without significant deterioration of theaudio quality for most of the program material types.Video presence had a small effect on audio qualityassessment.

9-3 Round Robin Subjective Evaluation of StereoEnhancement System for Headphones—GaetanLorho, David Isherwood, Nick Zacharov, and JyriHuopaniemi, Nokia Research Center, Tampere,Finland

This paper presents an evaluation of stereo enhance-ment algorithms for the reproduction of stereo musicover headphones. Nine state-of-the-art systems werecollected and applied to six different stereo programs.A subjective evaluation was conducted to comparethe algorithms against each other and against the un-processed stereo material. The results show that un-processed stereo music was preferred in the majorityof cases over stereo enhancement systems. Further-more, there are significant differences in the perceivedquality of the tested enhancement systems.

11:00 am–12:00 noon

SESSION 10: SUBJECTIVE AND OBJECTIVE EVALUATION, PART 2

10-1 Evaluation of Geometry-Based ParametricAuralization—Tapio Lokki and Ville Pulkki, HelsinkiUniversity of Technology, Espoo, Finland

This paper presents evaluation results of a binauralauralization system which is based on image sourceand edge diffraction modeling. The auralizationmethod is briefly reviewed, and evaluation of thequality of auralization performed both objectively andsubjectively is described. As a case study the aural-ization of a lecture room was considered. The evalua-tion results show that the spatial properties of appliedauralization are almost identical to those obtainedwith real head recording. However, timbre of the



sound still differs between binaural auralization andreal head recording, especially with transientlike sig-nals and at frequencies above 6 to 8 kHz. The evalu-ation results prove that plausible and natural sound-ing binaural auralization is possible with currentauralization algorithms.

10-2 The Significance of Tonality Index and NonlinearPsychoacoustics Models for Masking ThresholdEstimation—Evelyn Kurniawati,1 Javed Absar,2Sapna George,2 Chiew Tong Lau,1 and BenjaminPremkumar1

1 Nanyang Technological University, Singapore, Singapore

2 ST Microelectronics Asia Pacific Pte. Ltd., Singapore, Singapore

This paper presents a method to improve the clas-sical transform-based psychoacoustics model tobetter suit the physiological ear model, which ismore accurate in approximat ing the maskingthreshold obtained from the psychoacoustical mea-surement. The linear additivity assumption for si-multaneous masking is identified as the main rea-son behind the lower threshold obtained from aclassical transform-based PAM. Different methodsto estimate the tonality index were investigated, andthe inclusion of its measure in the nonlinear psy-choacoustics model is proposed. A subjective lis-tening test proved that our nonlinear PAM performsbetter than its linear counterparts.

12:00 noon–3:30 pm

SESSION 11: COMPUTATIONAL AUDITORY SCENEANALYSIS

11-1 Instrumental Analysis and Synthesis of AuditoryScenes: Communication Acoustics—Jens Blauert,Institut für Kommunikationsakustik, Ruhr-UniverstitatBochum, Germany (Invited)

In this paper the author proposes to use the cover la-bel Communication Acoustics for those branches ofacoustics that are closely related to the informationtechnologies and computer sciences. After a short re-view of the relevant research fields of the Institute ofCommunication Acoustics at Bochum, Germany, twoareas are dealt with in more detail, namely, instrumen-tal analysis and instrumental synthesis of auditoryscenes. Both areas show that cognitive and multi-modal phenomena have to be taken into account.Thus, future communication–acoustical systems willprobably contain increased knowledge-based andmultimodal components and will be embedded assubsystems into more complex systems. This techno-logical development trend will coin the future of com-munication acoustics in the context of the informationtechnologies.

11-2 Pulse-Dependent Analyses of Percussive Music—Fabien Gouyon and Perfecto Herrera, IUA-UPF,Barcelona, Spain; and Pedro Cano, Pompeu FabraUniversity, Barcelona, Spain

With the increase of digital audio dissemination, gen-erated by the popularization of personal computersand worldwide low-latency networks, many entertain-

ment applications can easily be imagined for rhythmicanalyses of audio. This paper reports on a method ofautomatic extraction of a rhythmic attribute from per-cussive music audio signals: the smallest rhythmicpulse, called the tick. Evaluations of the proposedscheme yielded good results. The paper discussesthe relevant use of the tick as the basic feature ofrhythmic analyses.

11-3 Methods for Blind Computational Estimation ofPerceptual Attributes of Room Acoustics—AlexisBaskind and Olivier Warusfel, IRCAM, Paris, France

This paper presents a set of signal processing meth-ods that provide a low-level description of the room ef-fect related to a recorded audio scene in order to de-rive a perceptual characterization of its spatialfeatures. The goal has been to perform this analysisblindly, that is, without any other information than therecording itself. Considering the respective nature ofthe early and the late contributions of the room effect,two distinct tools, acting on a monophonic signal, areproposed to estimate the relevant information of eachcontribution. Then, the estimations of the early partsof the responses are simultaneously processed by acoincidence detector in order to derive the binauralcues of the sound scene.

11-4 Finding Repeating Patterns in Acoustic MusicalSignals: Applications for Audio Thumbnailing—Jean-Julien Aucouturier, SONY CSL, Paris, France;and Mark Sandler, Queen Mary University of London,London, UK

Finding structure and repetitions in a musical signal iscrucial to enable interactive browsing into large data-bases of music files. Notably, it is useful to produceshort summaries of musical pieces, or audio thumb-nails. This paper proposes an algorithm to find repeat-ing patterns in an acoustic musical signal. First thesignal was segmented into a meaningful successionof timbres. This gives a reduced string representationof the music, the texture score, which does not en-code any pitch information. The authors then lookedfor patterns in this representation, using two tech-niques from image processing: kernel convolution andHough transform. The resulting patterns are relevantto musical structure, which shows that pitch is not theonly useful representation for the structural analysis ofpolyphonic music.

11-5 Extracting Sound Objects by IndependentSubspace Analysis—Shlomo Dubnov, Ben-GurionUniversity, Beer-Sheva, Israel

This paper presents a scheme for unsupervised ex-traction of sound objects or sources from a singlerecording containing a mixture of sounds. The sepa-ration/extraction procedure is performed by orthogo-nal projection of the mixed sound onto subspacesthat are derived by the clustering of transform coeffi-cients, such as coefficients obtained by PCA or ICA.The clustering step reveals a residual nonlineargrouping structure of the signal that is omitted by thelinear transform. To achieve independence, a searchfor partitioning that maximizes the mutual informationbetween a component and a set to which it belongswas done. This information was obtained by consid-ering a pair-wise distance measure among all coeffi-cients. Source separation experiments are reported inthe paper.


AES 22nd INTERNATIONAL CONFERENCE REGISTRATION FORMRegistration is also available online at www.aes.org. Register by May 20.

Company/organization Position/title

Family name First name AES member no.

Street address/P.O. Box Mr. Ms.

Postal code City State Country

Tel. Fax E-mail

Name of accompanying person(s) Student no.


Payment

1. Bank transfer. Bank: Kredietbank (KBC) Arenbergstraat 14, BE-1000 Brussels, Belgium,

Account number: 439-7977883-17; Account holder: AES Europe Office; IBAN Number: BE61.4397.9778.8317

2. Please charge the total ________ USD to my credit card Visa MasterCard

Credit card number Expir. date

Cardholder’s name

Cardholder’s address

Date Signature

PLEASE BOOK YOUR FLIGHT AND ACCOMMODATION EARLY. THIS IS PEAK SEASON IN FINLAND!

Please return by mail or fax to: AES 22nd Conference, AES HQ, 60 East 42nd Street, Room 2520, New York NY 10165 USAFax: +1 212 682 0477.

Conference fees Member Nonmember Total (USD)

Delegate registration 650 USD 750 USD

Author registration 550 USD 550 USD

Student delegate registration 400 USD 450 USD

Conference Social Program Delegate Accompanying person(s)

Opening dinner (Nokia Research Center), Friday, June 14 Free 30 USD x ____ persons

The Finnish Experience (Sauna, Sea, Sun), Saturday, June 15 Free 30 USD x ____ persons

Cruise and banquet (Suomenlinna Island), Sunday, June 16 Free 50 USD x ____ persons

TOTAL

Accommodation Early registration is advised to ensure accommodation at these hotels.

SAS Radisson Royal (Helsinki) Single room (76 EUR) Arrival ___/___/2002 Departure ___/___/2002www.radisson.com/helsinki.fi Double room (92.5 EUR) Share with (Double)

SAS Radisson Espoo (HUT) Single room (68 EUR)Room preference ____ Smoking ____ Nonsmoking

www.radisson.com/espoo.fi Double room (76 EUR)

A deposit of 120 USD is required to secure your hotel booking Arrival ____ before 18:00 h ____ after 18:00 hThis will be deducted from your final hotel bill, payable directly to the hotel in Euro. Any other special needs (dietary, etc.)


Faculty Advisor:Ken PohlmannUniversity of MiamiSchool of MusicP.O. Box 248165Coral Gables, FL 33124-7610Tel. +1 305 284 6252Fax +1 305 284 4448E-mail [email protected]

Acting Chair:Richard CassidyTel. +1 202 885 1236E mail [email protected]

Secretary:John W. ReiserP.O. Box 169Mt. Vernon, VA 22121-0169Tel. +1 703 780 4824Fax +1 703 780 4214E mail [email protected]

Chair: Danel HarringtonTel. +1 615 423 5152E-mail [email protected]

Chair-Elect/Membership Coordinator:Aaron SeftonTel. +1 615 849 8113E-mail [email protected]

Secretary:Norman TealeTel. +1 615 494 9286E-mail [email protected]

Treasurer:Ryan Kyzar

Tel. +1 615 848 7081E-mail [email protected]

Webmaster:Phil SchulloTel. +1 615 217 3861E-mail [email protected]

Committee:Mike ParagoneTel. +1 615 848 1579E-mail [email protected]

Faculty AdvisorDoug MitchellAES Student SectionMiddle Tennessee State University301 E. Main St., Box 21Murfreesboro, TN 37132Tel. +1 615 898 2553E-mail [email protected]

Chair:Seth DockstaderTel. +1 480 969 2635E-mail seth [email protected]

Vice Chair:Troy Peery

Secretary:James ValenteTel. +1 480 688 7278E-mail [email protected]

Treasurer:Scott Kent

President:Chad HedrickTel. +1 916 979 0151E-mail [email protected]

Vice ChairRobert HightowerTel. +1 916 961 9212

Secretary/Treasurer:Glenn TietjenTel. +1 916 929 5323E-mail [email protected]

Faculty Advisor:Eric ChunAES Student SectionAmerican River College Chapter2401 River Plaza Drive, Suite 130Sacramento, CA 95833Tel. +1 530 888 9440E-mail [email protected]

Chair:Ronny Andersson

Secretary:Anders Eriksson

Treasurer:Björn Hansell

Vice Chair:Zdenek KesnerTel. +420 2 2272 6393

Committee:Jiri FolvarcnyJiri SchimmelMiroslav Lukes

Chair (Section Contact)Jozo TalajicE-mail [email protected]

Chair:Nandu BhendeOpp. Holy Cross ChurchJuhukoliwada, Mumbai IN 400 049, IndiaTel. +91 22 660 2818Fax: +91 22 660 0778E-mail [email protected]

Vice Chair:Manohar KunteMumbai University Music Dept.

B Rd. ChurchgateMumbai IN 400 020, IndiaTel. +91 22 2818995Fax +91 22 2042859E-mail [email protected]

Secretary:Avinash OakWestern Outdoor Media Technologies16, Mumbai Samachar Marg, Mumbai,IN 400 023, IndiaTel. +91 22 2046181Fax +91 22 2043038E-mail [email protected]

Treasurer:Uday ChitreE-mail [email protected]

India Section

INTERNATIONAL REGION

Bosnia-Herzegovina Section

SOUTHERN REGION, EUROPE

Czech Section

CENTRAL REGION, EUROPE

University of Luleå-Piteå Section(Student)

NORTHERN REGION, EUROPE

American River College Section (Student)

Conservatory of The RecordingArts and Sciences Section

(Student)

WESTERN REGION, USA/CANADA

Middle Tennessee StateUniversity Section (Student)

CENTRAL REGION, USA/CANADA

District of Columbia Section

University of Miami Section(Student)

EASTERN REGION, USA/CANADA

Updates and Corrections to the 2001/2002

AES INTERNATIONAL SECTIONS DIRECTORYThe following listing details updated and corrected information to the 2001/2002 AES International Sections Directory,as it appeared in the 2001 December issue of the Journal. Please note that only specific section information regardingthe particular office being updated or corrected is included here (e.g., Chairman, Vice Chairman, etc.).

OF THE

SECTIONSWe appreciate the assistance of thesection secretaries in providing theinformation for the following reports.

NEWS


Technology’s New WorldOn April 18, seven members and 12guests of the Singapore Section gath-ered at Para-Di (S) Pte. Ltd to attend aseminar by Raymond Ng on the impact of information technology (IT)in the audio industry. Ng is owner ofPara-Di and a veteran provider of pro-ject studios and digital audio worksta-tion solutions. Section chair RobertSoo opened the meeting with an announcement of future section activi-ties and upcoming elections. He thenintroduced Ng as someone who per-sonally inspired him to investigateelectronic music and try composing.

Ng began his talk by describing whata typical day in the studio was like tenyears ago when analog recording wasprevalent. He compared an analog ses-sion with present methods; i.e., the pro-gression from multitrack tape recordersto computer-based systems with unlim-ited editing features, backing up andtransfer of audio by network. However,Ng stated that although the technologyhas changed, the rules of production re-main the same. As the local supplier ofDigidesign ProTools, he is well awarethat technology cannot replace creativi-ty. Creativity, command of the audiotechnicalities, a good audio mix and, ofcourse, a good relationship betweenproducer and engineer are still required.

Ng discussed many of the chal-lenges faced by studio owners, wholike it or not, must keep abreast of thedirection in which IT is progressingand make a commitment to stay cur-rent with the technology. Engineersand maintenance teams must alsomaintain a high level of computer lit-eracy in order to master and servicethe latest digital audio techniques.

“It is a brave new world out there,”said Ng. Traditional analog and tape

equipment will diminish in availabilityand importance and the digital studioand random access audio will becomea way of life. Ng concluded the meet-ing with a short tour of his studio. Ashe pointed out, all of the rooms werenetworked and had IT-related productsfor the distribution of audio and video.

Jibby JacobOn November 30, 12 members and

13 guests of the Singapore Sectionmet at A.C.E. Daikin Pte Ltd. for aseminar on wireless digital audio tech-nology. Guest speaker Chan Chee Oei,director of engineering at FreeSystemsPte Ltd., said that one of the missionsof his company is to develop wirelesssystems and solutions for digital audiothat take advantage of modern tech-nologies and replace older technolo-gies now in use, such as FM radio.According to Oei, some of the attribut-es of wireless technology include lowidle noise, at least CD quality audio,immunity to interference, and goodmobility.

Oei then provided a short history of

wireless headphone technology. Thefirst wireless headphones were intro-duced in the 1970s and exhibited anSNR in the range of only 30-40 dB.From 1995-1997, slight improvementsin wireless technologies were able toincrease the SNR of wireless head-phones to the 50-60 dB range. Howev-er, the quality was still well below the96 dB that CDs — the primary sourcemedium of the time — could deliver.

Oei described the various phases ofdevelopment of the new FreeSystemswireless headphone technology. Sig-nals had to be coded and modulatedbefore transmission, then demodulat-ed, decoded and corrected for errorsupon reception. For example, theFreeSystems FS1901/2 transmitter/receiver has an isochronous data rateof 2 Mbps with a stereo signal of 48kHz/24-bit. The modulation methodused was based on pulse positionmodulation, known for its power-effi-cient coding. For error correction cod-ing, the company employed a form ofReed-Solomon coding using a

Singapore members gathered after a meeting on “Wireless Digital AudioTechnology” in November. Chan Chee Oei, guest speaker, in front row, fifth from the left, holds AES logo. photo by Roland Tan

dios. Other features of the complex include a large narration booth, sophis-ticated machine layout for high pro-duction efficiency and specific bassmanagement using a subwoofer to han-dle bass components of all the chan-nels. Someya concluded with a sounddemonstration that showed off the stu-dios’ acoustic excellence.

On January 24, 36 members andguests gathered at the Pioneer listeningroom in Tokyo to experience firsthandthe company’s recently developed“Automatic Sound Field EqualizationSystem for Multichannel Sound Lis-tening.” Shinji Koyano of Pioneer discussed some of the problems associ-ated with conventional systems, whichattempt to simulate the acoustics of existing concert halls. In such systems,he said, the listener’s own room rever-beration is inevitably added to that included in the sound source. This con-fuses the results.

It is difficult to meet the profession-al ITU requirements for multichannelreproduction. Koyano’s group tried todevelop a system that automaticallyequalized frequency response, leveldifferences and delay times of the reproduced multichannel sound asclosely as possible to ITU require-ments.

Yoshiki Ota then explained how thesignals were processed based on thesound detected by a microphoneplaced at the listener’s ear. Frequencyresponse was adjusted according to theoutput of a 9-band-pass filter using agraphic equalizer. Level differences

During his lecture, Moses discussedthe salient issues related to the distrib-ution of audio on the Internet, whichinclude related technologies, the man-agement of artist rights, and new busi-ness models. As Moses pointed out,audio searches and downloads havematched or exceeded almost all otheruses for the Internet in recent years. Hesaid that this could no longer be ignored or mismanaged by the record-ing industry. He also covered issuesrelevant to recording engineering prac-tice, such as mastering for the formatand the audio practitioner’s role intranslating recordings for new media.

At the end of the program, the audi-ence had the opportunity to participatein a single-blind listening test to compare multiple audio players at dif-ferent data rates. After the meeting,several officers discussed section busi-ness and preparations for the AESConference on multichannel surroundin June 2003.

John Sorensen

Two from Japan Eighteeen members and guests of theJapan Section visited the Sony PCLVideo Center in Tokyo on October 19.The center’s studios 405 and 408 wererecently renovated for multichannelsound productions for DVD, digitalTV broadcasting and commercialfilms. Sony sound designer KazutakaSomeya talked to the group about howthese studios had to conform to theTHXpm3 requirements of Lucas Stu-

hardwired DSP engine to perform thecomputationally intensive algorithm.Soft muting, using a gentle entry and recovery in the muting profile, helpedto overcome unpleasant clippingsounds whenever a noncorrectable error occurred.

Regarding the transmission medium,Oei spoke about the implementation ofwireless audio using analog and digitalIR, as well as analog and digital RF.He compared their range, ability towork in enclosed areas, EMI issuesand error correction capabilities. Clear-ly, digital is superior to analog, yield-ing better figures for frequency response and total harmonic distortion(THD+N). The digital IR system maintained the high SNR with respectto distance, compared to its analogequivalent.

After a brief discussion of the compa-ny’s product line, Oei talked about thecurrent and future direction thatFreeSystems is taking. The company’sgoal, he said, is to make a wirelesstransmission system that is as good as awired system. He touched on the chal-lenges in using RF, such as a highertransmission bit rate requirement.Demonstrating the FreeSystems digitalIR system, Oei compared it to a com-mercially available analog IR system.The hissing noise was very obvious inthe analog system. Using a dozen setsof wireless headphone receivers andone transmitter, the audience enjoyed ashort clip of a recent action movie andexperienced the fine audio quality of thenew technology.

Christopher Yap

Moses Visits AlbertaOn February 9, the Alberta Sectionmet in the Chamber Music Studio atthe Banff Centre for a session on Internet audio with Bob Moses.

Moses is president and chief techni-cal officer of Island Digital MediaGroup and was the founder and chieftechnical officer of Digital HarmonyTechnologies. A wide range of corpo-rations, including Rane, Symetrix,JBL, Microsoft, Harman, Peavey, and Denon, have used many of his techno-logical developments to build audiostreaming technologies for LANs.

OF THE

NEWS

SECTIONS


At Japan Section’s January meeting on sound field equalization for multichannelsound listening, standing from the left: K. Koguchi, S. Koyano, and Y. Ota.

OF THE

NEWS

SECTIONS

receiver to microphone is the externalair path (but this is rarely the case).

Phone housings and PC boards actas ducts, directing energy down thephone toward the microphone. Thisducted energy can be reduced by cre-ating partitions and using grommetsaround microphone and loudspeaker.Alternate acoustic paths, such as leak-ing through jacks on the side or bot-tom of the phone, also need to besealed. Separately rotating housingseliminate both the directly ducted energy and alternate acoustic pathsand also decrease the microphone-to-mouth distance.

Audible Alerts or ringers have histor-ically been able to produce only a fewtones at a very high level for their sizeand input voltage. Armature alerts aretypically used to produce these tones.However, consumers are now askingfor programmable musical tones andaudio for games from these same devices. Since space is very limited forthese devices, new system configura-tions are being implemented to get asmuch SPL out of as little volume aspossible. Increasingly microloudspeak-ers are replacing the armature alerts toprovide more musical alert tones alongwith speakerphone operation. Since vol-ume and board space are at a premiumon cellular systems, components arenow being used for multiple uses. Sev-eral companies are now marketing mul-tifunction transducers, which typicallyinclude vibrator (silent alert), alert, andreceiver.

New techniques are constantly being developed to overcome the chal-lenges being presented by new trendsin portable communications devices.Coupled with consumer demands forgreater integration of entertainmentdevices, new devices may eventuallychange our lives in the same way thefirst cellular phones have in the 30years since their inception.

Giles Davis

Studio Designer in India Members of the India Section met atthe Empire Audio Centre on January 8for the first technical meeting of 2002.The guest speaker was studio designerSam Toyoshima, who began his

present were impressed with Neve’s incredible knowledge. They said theylearned a great deal from the discussion.

Seth Dockstader

Wireless Phones“Wireless Telephone Design Consid-erations” was the topic of ChicagoSection’s December 19 meeting at theHoliday Inn North Shore in Skokie,IL. Bob Zurek of Motorola was thefeatured speaker.

As wireless phones have matured, theexpectations of performance and sizehave led to new acoustic challenges.The desire to reduce the size of mod-ern cell phones has made the incorporation of an earcup in the device impractical. Without an earcup,a decent seal usually assumed with alandline phone is not possible. The environment that a cell phone is used incontributes to the lack of seal in the sys-tem. Users will often set the volume toits maximum setting and vary the amplitude by moving the phone towards or away from their ear. A sys-tem that is designed to accommodate

this varying imped-ance is often referredto as leak-tolerant.

The push to reduce the overallsize of the unit hascreated system echoproblems that do notexist in full sizelandline systems.Shortening thelength of the phoneincreases the m i c r o p h o n e - t o -mouth distance, affecting signal-to-noise ratio and microphone sensi-tivity. A reduction inthe internal volumecoupled with an

increased distance between the mouthand the microphone tend to increasereceiver to microphone feedback. Thisfeedback coupled with digital systemdelay (as much as 200 ms) creates sig-nificant echo problem for the far enduser. To reduce the problem, considerthat the best case coupling from

between loudspeakers were compen-sated for based on the power levels derived from pink noise weighted forloudness. An exponential pulse wasused to obtain delay time differencesof less than 0.1 ms accuracy betweenchannels and minimal backgroundnoise disturbance. Using these para-meters, Ota’s team was able to employreal time signal processing for tuningfilters and delay lines to set up the tar-get listening conditions using DSPsdeveloped for just that purpose.

To conclude, Kiichiro Koguchidemonstrated a system built into theModel VSA-AX10 multichannel amplifier. He compared results byplaying DVD discs with and withoutequalization.

Vic Goh

Neve Visits ConservatoryRupert Neve was a guest speaker at ameeting of the Conservatory ofRecording Arts and Sciences StudentSection on November 7. Present werestudents, faculty, and many AESmembers from the Greater Phoenix

area, who gathered to hear this audioengineering legend.

Neve talked about his history and involvement with equipment designover the years and shared his wisdomon various sound recording topics. Hestressed the importance of referencequality and extended bandwidth. All


Students at the Conservatory of Recording Arts andSciences Section hosted guest speaker Rupert Neve.From the left: Eric Weaver, J.D. Pinkus, R. Neve, JoshHamilton, and Terrisa Frame (in front). photo by Sonny Wong

NEWS

thickness results in varying resonances.The two-hour meeting was hailed as

a grand success. The section expressedits thanks to Michael Music and theEmpire Audio Center for making themeeting possible.

Avinash Oak

his philosophy in relation to each indi-vidual case. He stressed the impor-tance of isolation, and talked about theinstances in which he strived toachieve the NC15 criteria in his designs. He also talked about the“floating studio” and “variableacoustics” concepts. He used detaileddrawings and an array of photographsto help illustrate and explain the con-cept of rotating panels.

Toyoshima later led a lively ques-tion-and-answer session. He talked atlength about the materials he uses inacoustic treatments, and said that heonly uses those listed in the Japan orUK standards list. One guest wantedto know if it would be possible to include a material of Indian origin inthese standards lists. Toyoshima offered to test and evaluate the materi-al. In answer to another question onthe design of the observation windowbetween the studio area and controlroom, Toyoshima explained that theuse of three glass panes of differing

talk by describing his early days at theJVC Research Center.

Using his laptop to organize his presentation, Toyoshima guided theaudience through the various stepping-stones of his career: his designs of someof the premier studios in Japan, his rolein the establishment of the Acoustic Design Group in the United Kingdomand many subsequent design projects,which took him all over the world.

Toyoshima then showed some of hisfavorite project designs and explained


Sam Toyoshima shows his favoritestudio designs to India Section.

SUBMISSION OF PHOTOGRAPHS

Sections may send photographselectronically for publication in“News of the Sections.” Forgood reproduction in the Jour-nal, the files should be saved asan uncompressed TIFF or EPSformat with a resolution of 300dpi. They may be sent on a zipdisk to headquarters or over theInternet to: [email protected]. Sec-tions that submit black andwhite photographs may continueto mail them to Abbie Cohen,Senior Editor, at the headquar-ters address.

For price and ordering information send e-mail to [email protected], visit the

AES Web site at www.aes.org, or call any AESoffice at +1 212 661 8528 (USA); +44 1628 663725 (UK);

+33 1 4881 4632 (Europe).

9000Journal technical articles,convention preprints, andconference papers at your

fingertipsThe Audio Engineering Society has published a 17-diskelectronic library containing most of the Journal technicalarticles, convention preprints, and conference papers pub-lished since its inception through the year 2000. Theapproximately 9000 papers and articles are stored in PDFformat, preserving the original documents to the highestfidelity possible while permitting full-text and field search-ing. The library can be viewed on Windows, Mac, and UNIXplatforms.

You may purchase the entire 17-disk library or disk1 alone. Disk 1 contains the program and installa-tion files that are linked to the PDF collections onthe other 16 disks. For reference and citation con-venience, disk 1 also contains a full index of alldocuments within the library, permitting you toretrieve titles, author names, original publicationname, publication date, page numbers, andabstract text without ever having to swap disks.

OF THE

SECTIONS

ABOUT PEOPLE…

AES member Mark Gander, vicepresident of marketing for JBL Profes-

sional, has received the JBL GoldenSpeaker Award for his 25 years of ded-ication to the company. PresidentMichael MacDonald presented theaward to Gander during JBL’s annualholiday party.

“Mark is the first and last stop foraudio information at JBL Professional.We are privileged to have him on ourteam,” said MacDonald.

Gander received his M.S.E.E. degree from the Georgia Institute ofTechnology in 1976, specializing inaudio electronics and instrumentation.That year he joined James B. LansingSound, Inc., where he was responsiblefor various consumer and professionalloudspeaker products including the E-Series and Cabaret Series musical instrument loudspeakers. At JBL, hehas held positions as transducer engi-neer, applications engineer, productmanager, vice president marketing,vice president engineering and vicepresident of strategic development.


TRACK

SOUND

Gander became an honorary memberof the AES in 1975.

In 1994, Gander received an AESfellowship for significant develop-ments in low-frequency transducer design and measurement and loud-speaker systems for large-scale musicreinforcement. That year he was elect-ed an AES governor. He is also amember of the Acoustical Society ofAmerica, the Institute of Electrical andElectronics Engineers, and the Societyof Motion Picture and Television Engineers.

AES member Daniel J. Field hasbeen elected to the Board of Directorsof the Association of LoudspeakerManufacturing and Acoustics Interna-tional (ALMA International), a globalnon-profit trade association for theloudspeaker industry, for the 2002-2005 term.

Field is an electronics industry veter-an with over 23 years experience in engineering and product developmentfor consumer audio equipment. Aftergraduating from Southern Illinois Uni-versity in 1977, Field held positions incommercial sound and broadcasting before signing on with Cletron/Har-man-Motive in 1979. In 1991, he became senior member of the technicalstaff for Thomson RCA, working onadvanced television sound. From 1997to 2002, he was vice president of Klipsch, LLC.

Field has been an ALMA memberrepresentative since 1989, havingserved as standards chair from 1991until 1999 and on the board of direc-tors from 1993 until 1999. In addition to the Audio Engineering Society, he is also a member of theAcoustical Society of America.

Field holds two patents and haswritten six technical papers dealing

with loudspeakers and sound systems.His interests include room acoustics,loudspeaker design, linear program-ming and powerful appliances.

MEETINGS, CONFERENCES…

The Ninth International Congresson Sound & Vibration, ICSV9, willtake place at the University of CentralFlorida, UCF, in Orlando, Florida, July8-11, 2002. It is co-sponsored byNASA, IIAV and UCF.

A detailed list of topics of interest isshown on the ICSV9 Web site togetherwith information on the 40 special ses-sions being planned with contact information about their organizers.

As a travel destination, Orlando offers many local attractions. In addi-tion to a first-rate scientific program,attendees can also experience DisneyWorld, EPCOT, Universal Studios andSea World. NASA’s Kennedy SpaceCenter (KSC), will be open for techni-cal visits. The banquet will be held inthe building housing a full scale SaturnV Moon launch vehicle.

The International Conference onConsumer Electronics (ICCE), has issued its annual call for technical papers. The conference, which attracts authors and attendees from around theworld, will be held in Los Angeles,CA, on June 18-20, 2002, with tutori-al sessions on June 16-17. The dead-line for technical paper submissionsfor both oral and poster presentationsis January 14. Topics includeaudio/video devices, digital camerasand camcorders, home networking,wireless communication and comput-ing, PC/TV convergence and mobileentertainment. Details and submissionforms may be found online at:www.icce.org.

Mark Gander (left) receives awardfrom JBL’s Michael MacDonald.

A E S S U S T A I N I N G M E M B E R

BODYPACK TRANSMITTER mea-sures 61 mm x 53 mm x 17 mm anddelivers 30 mW of wireless audio. Thisis about half the size of the SennheiserSK 50. Model SK 5012 features lowself-noise, multichannel capabilities,and an easy-to-program interface. TwoAAA batteries power the metal unit forseven hours of continuous operation.Signal strength is independent of bat-tery status. Sennheiser ElectronicCorporation, 1 Enterprise Drive, OldLyme, CT 06371, USA; tel. +1 860434 9190; fax +1 860 434 1759; Website www.sennheiserusa.com.


CLIP-ON INSTRUMENT MICRO-PHONE is a miniature condenser thatcan be securely mounted to almost anyinstrument as a self-contained system.The Beta 98 H/C utilizes a clip-onclamp that allows a quick change

between instruments. The unit is lowprofile in design to minimize stagepresence and has the capacity for han-dling high SPLs and optimal gainbefore feedback. The unit comes witha flexible gooseneck, locking wind-screen for outdoor use, and an integrat-ed isolation shock mount to reducetransmission of instrument “key noise”and other mechanical noise. Themicrophone is available in a wirelessconfiguration terminated to a minifour-pin connector (WB98H/C) or as a wired version with an in-line preamplifier (Beta 98H/C). Shure Inc.,222 Hartrey Avenue, Evanston, IL 60202, USA; tel. +1 847 866 2200; fax +1 847 866 2279; [email protected] (MichelleKohler); Web site www.shure.com.

SUBWOOFER is an active 18-in unitdesigned for mobile entertainers andpermanent installation applications.The B-52 SP-18 has over 1200 W ofcontinuous program power handling, afrequency response down to 30 Hz,and a peak SPL of over 133 dB. Thepatent pending 1200-W rms digitalamplifier is 90+ percent efficient withbuilt-in signal processing, 24dB/octave filters, compression, andVCA limiter circuitry. Additional fea-tures include XLR high-pass and full-range outputs, a phase switch, and aNeutrik Speakon output. The unit canalso power the SP-18S, a passive sub-woofer version of the SP-18. B-52 Pro Audio, 3383 Gage Avenue,

Huntington Park, CA 90255, USA; tel.+1 323 277 4100 or 800 344 4384(toll-free); fax +1 323 277 4108; e-mail [email protected]; Web site www.B-52PRO.com.

IN-WALL LOUDSPEAKER achievesinvisibility by utilizing a frameless,one-piece grille comprising a whitepowder coating over a zinc-fortifiedepoxy base that prevents corrosion inhigh-humidity areas. Neodymiummagnets mount the grille to the bafflewith more than enough strength forceiling mounts. The SW-150 uses theidentical components of M&K’s S-150 satellite and MPS-2510 profes-sional and consumer monitors. Otherfeatures include phase-focusedcrossover design and an acousticalfoam kit to enhance the sound bybreaking up the backwave produced bythe woofer driver. M&K Sound, 9351Deering Street, Chatsworth, CA91311, USA; tel. +1 818 701 7010; fax +1 818 701 0369; Web sitewww.mksound.com.

MODULAR SINGLE-CHANNELAMPLIFERS are designed for on ornear loudspeaker mounting. ThePowerPac line comprises three ampli-fiers that deliver 60 W, 120 W or 250W into 8 Ω. All units come with five-way banana jack outputs, RCA (unbal-anced) and XLR (balanced) inputs, anda ground lift switch to mitigate anysystem hum. The PowerPac 120 and

AND

DEVELOPMENTSProduct information is provided as aservice to our readers. Contact manu-facturers directly for additional infor-mation and please refer to the Journalof the Audio Engineering Society.

NEW PRODUCTS


250 have 12-V triggers for remote con-trol. The amplifiers can be bolted ontothe back of any loudspeaker and arecompatible with the company’s line ofProfessional Monitor Company transmission line loudspeakers, whichare PowerPac-ready. Bryston Ltd., 677Neal Drive, Peterborough, Ontario K9J 7Y4, Canada; tel. +1 705 7425325 or 800 849 2912 (toll-free); e-mail [email protected] (Karen Richardson); Web sitewww.bryston.ca.


POWER AMPLIFIERS use a con-ventional power supply with a toroidaltransformer in conjunction with semi-conductor technology and proprietarycopper cooling to offer excellent sonicperformance in a compact 2-U pack-age. Models MA900 and MA200Qfeature a multiple position gain switch,removable front panel dust filter cover, multiple input connectors, three-pinPhoenix terminal block for permanentinstallations, and LED indicators which show the status of the amplifier.The MA900 includes 2 x 450 W into 4Ω. The MA200Q includes 4 x 200 Winto 4 Ω. Martin Audio, Ltd., P.O. Box44019, Kitchener, Ontario N2N 3G7,Canada; tel. +1 519 747 5853; fax +1519 747 3576; e-mail [email protected] (Rob Hofkamp); Web sitewww.martin-audio.com.

DIGITAL AUDIO WORKSTATIONis a 24-bit/48-kHz digital unit designedpredominantly for field use. The com-pact system integrates recording soft-ware, an audio interface, and a SonyVAIO Picturebook C1VN laptop with alarge-capacity hard drive and is opti-mized Windows build. The Sonicorderis offered in three turnkey configurationsor as a custom-designed system on arecording professional’s laptop ofchoice. Accessories available for usewith each model include external 60-and 80-GB 3.5-in external hard drivesand CD recorders, 30-GB 2.5-inFirewire dc/bus-powered hard drives,and lightweight power options to extendrecording time. Sonic Sense, Inc., 2755Gilpin Street, Denver, CO 80210, USA;tel. +1 303 753 0201 or 877 324 4463(toll-free); fax +1 303 753 0817; Website www.sonicsense.com.

MICROPHONES for broadcasting,film, television, and home recordingstudio applications are now available.Model StudioPro M2 is a large con-denser, dual-diaphragm capsule withthree directional patterns: onmidirec-tional, cardioid, and figure eight. Userscan switch between patterns using abutton below the head grille. TheStudioPro M1 offers only the cardioidpattern. Both models have a 10-dBattenuation switch which enables themicrophones to handle sound-pressurelevels of up to 140 dB without distor-tion. The frequency response of thecardioid and figure-eight patterns arevery flat for frontal sound incidence,even in the upper frequency range. In

addition, the microphones can be usedvery close to a sound source withoutthe sound becoming unnaturally harsh.Likewise, a high-pass filter helpsreduce the interference of subsonic andlow frequencies. Peavey ElectronicsCorporation, 711 A Street, Meridian,MS 39301, USA; tel +1 601 483 5365;fax +1 601 486 1278; [email protected] (Ladd Temple);Web site www.peavey.com.

LOUDSPEAKER MANAGEMENTSYSTEM provides flexibility toATK/Audiotek’s large public addresssystems. The new XTA DP224 proces-sor adds stereo capability to thosefunctions handled by the XTA DP200processors, which are already heavilyused by the company. These proces-sors allow the parameter curves ofeach specific configuration to be stored and instantly recalled. XTAElectronics, Group One Ltd., 200 SeaLane, Farmingdale, NY 11735, USA;tel. +1 631 249 1399; fax +1 631 7531020; e-mail [email protected] (SueAdamson); Web site www.g1ltd.com.


AND

NEW PRODUCTS

DEVELOPMENTS

-praxis-Liberty Instruments’ new audio measurement system

visit www.libinst.com for free software download

Liberty Instruments, Inc.P.O. Box 1454, West Chester, OH 45071 USA Phone/Fax 513 755 0252

Powerful

Compatible

Portable

Intelligent

Versatile

Affordable

ment, first order digital filters, audiotone control and equalizers, productspecifications, product usability, financial models, and digital reverber-ation. For information and to obtaincopies of these notes, contact him at617 489 6292 or at Seven Woods Audio, Inc., 44 Oak Ave, Belmont,MA, 02478 USA.

Practical Recording Techniques,Third Edition, The Step-by-Step Ap-proach to Professional Audio Record-ing by Bruce and Jenny Bartlett (FocalPress) is a hands on guide for begin-ners and intermediate recording engin-ers, producers, musicians and audioenthusiasts.

The guide offers advice on equip-ping a home studio—from low budgetto advanced—and includes sugges-tions for set-up, acoustics, choosingmonitor loudspeakers and preventinghum. The 576-page paperback also ex-plains how to judge recordings anduse the equipment available to im-prove them. Filled with tips and short-cuts, the guide includes situations andtopics not covered in similar texts, in-cluding documenting the session,recording spoken word, troubleshoot-ing bad sound and more.

New to this edition are chapters onsurround sound techniques, impedanceand audio for the Internet, as well asexpanded information on digitalrecording, MIDI sequencing and auto-mated mixing. Price is $34.99. FocalPress, 225 Wildwood Avenue,Woburn, MA 01801-2040, USA; tel:800-366-2665 or 781-904-2500, fax:800-446-6520 or 781-904-2640, Inter-net: www.focalpress.com, e-mail:[email protected],

LITERATUREThe opinions expressed are those ofthe individual reviewers and are notnecessarily endorsed by the Editors ofthe Journal.

AVAILABLE

IN BRIEF AND OF INTEREST…

TV Operator’s Certification Hand-book, Fifth Edition, by Fred Baum-gartner and Doug Garlinger (Societyof Broadcast Engineers) provides aconcise resource for TV operators.The new edition includes the latesttechnology and broadcast practices,including centralized broadcasting(centralcasting) and new informationregarding the Children’s TelevisionAct.

Once operators read the book, theyhave the opportunity to take a test todemonstrate their knowledge. Applicants with passing scores areawarded the SBE’s Certified Televi-sion Operator (CTO) designation,which is valid for five years.

The book’s authors are experiencedTV broadcast engineers with a com-bined 63 years of professional experi-ence. Baumgartner is director of NewProduct Development for the AT&TBroadband and Internet Services Divi-sion in Denver, Colorado. Garlinger isdirector of engineering of LeSEABroadcasting Corporation and worksfrom his office in Noblesville, IN. Toorder the book, contact: Society ofBroadcast Engineers, 9247 NorthMeridian Street, Suite 305, Indianapo-lis, IN 46260, USA; tel: 317-846-9000, fax: 317-846-9120, Internet:www.sbe.org.

Christopher Moore, AES memberand a consultant in the development ofnew audio products, has written adozen application notes covering var-ious aspects of new product develop-ment. Topics include level diagrams,customer interviews, noise measure-


2002 June 1-3: 21st AES International Conference,“Architectural Acoustics andSound Reinforcement,” St.Petersburg, Russia, HotelMoscow. For informationsee p. 328.

•

2002 June 3-7: 143rd Meetingof the Acoustical Society ofAmerica, Pittsburgh, PA.For information contact tel:516-576-2360, fax: 516-576-2377 or e-mai l :[email protected].

•

2002 June 4-6: 6th Interna-tional Symposium on Trans-port Noise and Vibration,St. Petersburg, Russia. Formore information, [email protected].

•

2002 June 9-12: Seminar onAudio Engineering, FederalUniversity of Mines, BeloHorizonte, Brazil. For infor-mation, tel: +55 (31) 3499-4848, fax: +55 (31) 3499-4850 or e-mai l :[email protected] .

•

2002 June 15-17: 22nd AESInternational Conference, Espoo, Finland, “Vir tualSynthetic, and Entertain-ment Audio,” Helsinki Uni-versity of Technology. Seep. 328 for more information.

•

2002 June 18-20: 21st IEEEInternational Conference on Consumer Electronics(ICCE), Los Angeles, CA,USA. Contact DianeWilliams, conference coor-dinator, tel: 585-392-3862,fax: 585-392-4397.

•

2002 October 5-8: 113th AESConvention, Los Angeles,CA, USA, Los Angeles Con-vention, Los Angeles, CA.See p. 328 for information.

Upcoming Meetings


Section symbols are: Adelaide (ADE), Alberta (AB), All-Russian State Institute of Cinematography (ARSIC), American RiverCollege (ARC), American University (AMU), Argentina (RA), Atlanta (AT), Austrian (AU), Ball State University (BSU),Belarus (BLS), Belgian (BEL), Belmont University (BU), Berklee College of Music (BCM), Berlin Student (BNS), Bosnia-Herzegovina (BA), Boston (BOS), Brazil (BZ), Brigham Young University (BYU), Brisbane (BRI), British (BR), Bulgarian(BG), Cal Poly San Luis Obispo State University (CPSLO), California State University–Chico (CSU), Carnegie MellonUniversity (CMU), Central German (CG), Central Indiana (CI), Chicago (CH), Chile (RCH), Citrus College (CTC), CogswellPolytechnical College (CPC), Colombia (COL), Colorado (CO), Columbia College (CC), Conservatoire de Paris Student(CPS), Conservatory of Recording Arts and Sciences (CRAS), Croatian (HR), Croatian Student (HRS), Czech (CR), CzechRepublic Student (CRS), Danish (DA), Danish Student (DAS), Darmstadt (DMS), Denver/student (DEN/S), Detmold Student(DS), Detroit (DET), District of Columbia (DC), Duquesne University (DU), Düsseldorf (DF), Ex’pression Center for NewMedia (ECNM), Finnish (FIN), Fredonia (FRE), French (FR), Full Sail Real World Education (FS), Graz (GZ), Greek (GR),Hampton University (HPTU), Hong Kong (HK), Hungarian (HU), Ilmenau (IM), India (IND), Institute of Audio Research(IAR), Israel (IS), Italian (IT), Italian Student (ITS), Japan (JA), Kansas City (KC), Korea (RK), Lithuanian (LT), LongBeach/student (LB/S), Los Angeles (LA), Louis Lumière (LL), Malaysia (MY), McGill University (MGU), Melbourne (MEL),Mexican (MEX), Michigan Technological University (MTU), Middle Tennessee State University (MTSU), Moscow (MOS),Music Tech (MT), Nashville (NA), Netherlands (NE), Netherlands Student (NES), New Orleans (NO), New York (NY), NorthGerman (NG), Northeast Community College (NCC), Norwegian (NOR), Ohio University (OU), Pacific Northwest (PNW),Pennsylvania State University (PSU), Philadelphia (PHIL), Philippines (RP), Polish (POL), Portland (POR), Portugal (PT),Ridgewater College, Hutchinson Campus (RC), Romanian (ROM), SAE Nashville (SAENA), St. Louis (STL), St. Petersburg(STP), St. Petersburg Student (STPS), San Diego (SD), San Diego State University (SDSU), San Francisco (SF), SanFrancisco State University (SFU), Singapore (SGP), Slovakian Republic (SR), Slovenian (SL), South German (SG), SouthwestTexas State University (STSU), Spanish (SPA), Stanford University (SU), Strasbourg Student (SBS), Swedish (SWE), Swiss(SWI), Sydney (SYD), Taller de Arte Sonoro, Caracas (TAS), Technical University of Gdansk (TUG), The Art Institute of Seattle(TAIS), Toronto (TOR), Turkey (TR), Ukrainian (UKR), University of Cincinnati (UC), University of Hartford (UH), Universityof Javeriana, Bogota (UJ), University of Lulea-Pitea (ULP), University of Massachusetts–Lowell (UL), University of Miami(UOM), University of North Carolina at Asheville (UNCA), University of Southern California (USC), Upper Midwest (UMW),Uruguay (ROU), Utah (UT), Vancouver (BC), Vancouver Student (BCS), Venezuela (VEN), Vienna (VI), West Michigan (WM),William Paterson University (WPU), Worcester Polytechnic Institute (WPI), Wroclaw University of Technology (WUT),Yugoslavian (YU).

INFORMATION

MEMBERSHIP

Paul C. Adlaf6208 Allan St. #5, Halifax Nova Scotia, B3L1G6, Canada

Paulo Andrade EsquefLeivosentie 17A, FI 00780, Helsinki, Finland(FIN)

Stel Anthony40 The Green Ave., Porthcawl, Wales, CF363AX UK (BR)

Olav M. ArntzenSeas Fabikker AS, P.O. Box 600, NO 1522,Moss, Norway (NOR)

Craig S. BelcherP.O. Box 335, Wyncote, PA 19095 (PHIL)

Fernando B. BolanosJose Jae 7 No. 4, 3 Flr., ES 46117, BeteraValencia, Spain (SPA)

Sean D. Browne3409 Jetty Ln., Columbia, PA 17512 (PHIL)

Gustavo A. Celis16424 N.W. 15th St., Pembroke Pines, FL33028

Deepan ChatterjiGukmohr 35 West Ave., Santacruz (W),Mumbai 400 054, India (IND)

Wongab JH ChoiRPG Korea Diffuser Systems, Missy 860b/d Ste. 2122, Seoul 135 240, Korea (RK)

Edward Coleman9612 Summit Circle, La Mesa, CA 91941(LA)

Jose A. Dias TavaresRua da Estrada #16, PT 3865-011, Canelas-Etr Estarreja, Portugal (PT)

Vincenzo Ferraravia Rosata 73, IT 0012, GuidoniaMontecelio (RM), Italy (IT)

Michael D. Gatlin87-89 Wakeman Ave. #G3, Newark, NJ07104 (NY)

Andre S. Gauthier166 Second Ave. #12L, New York, NY10003 (NY)

Jay N. Goldman2328 Thomas Ln., Wilmington, DE 19810(DC)

Guillermo Grez

Av. Cenenario 994, Depto. B-203, SanMiguel, Santiago, Chile (RCH)

Paolo Guidorzivia Malavasi 26, IT 40033, Casalecchio diReno (BO), Italy (IT)

Satish GuptaSahara India T.V. Network, S.V. Road,Goregaon (W), Mumbai 400104, India(IND)

Shrikrishna H. M.F-1 649 47th St. 9th Sector, K.K. Nagar,Chennai 600 078, India (IND)

Elizabeth M. Havenor3500 Bryant Ave. S., Minneapolis, MN55408-4119 (UMW)

Daniel L. Janko2513 S. Ginger St., Cornelius, OR 97113(POR)

Benjamin F. PizzutoCrawford Comminications, 3845Pleasantdale Rd., Atlanta, GA 30340 (AT)

Ghislain PomerleauSociete Telediffusion du Quebec, 1000 AlleFulldum, Montreal, Quebec H2K 3L7Quebec, Canada (TOR)

MEMBERS

These listings represent new membership according to grade.

MEMBERSHIP

INFORMATION

Theodore W. Prohinsie26 Haring Ave., Hicksville, NY 11803 (NY)

Ferran G. PuigPlaca D'Europa No. 3, Zon Zona, GironaCatalunya, ES 17005, Spain (SPA)

William PutnamUniversal Audio, 2125 Delaware, Santa Cruz,CA 95060 (SF)

Zakaria Bin RasliBlock 359 Woodlands Ave. 5, #12-368,Singapore 30359, Singapore (SGP)

Dayakar C. Reddy3335 Kifer Rd., Santa Clara, CA 95051 (SF)

Carlos H. ReyesWilson, Ihrig & Associates, Inc., 5776Broadway, Oakland, CA 94618-1531 (SF)

Norbert T. RichardsonPolk Audio, 5601 Metro Dr., Baltimore, MD21215 (DC)

Robert M. Rinsky152 Stanwick Ct., Holmdel, NJ 07733 (NY)

Ricardo M. RodriguesRua Euclides da Cunha 176, Bloco “B” Ap602 Sao Cristovao RJ 20940-060, Brazil(BZ)

Gress L. Roth11837 S.E. Raymond St. #F, Portland, OR97266 (POR)

Stuart F. RubinDigital 5, 101 Grovers Mill Rd., Ste. 200,Lawrenceville, NJ 08648 (PHIL)

Thomas N. RyanGPD Group, 520 S. Main St., Akron, OH 44311

Franklin G. SalazarReel FX Creative Studios, 2211 N. Lamar,Ste. 100, Dallas, TX 75202

Eric D. Schlosser1604 Stanford Ave., Metarie, LA 70003(NO)

Douglas J. ShearerSanctuary Mastering, Sanctuary House, 45-53 Sinclair Rd., London W14 0NS, UK (BR)

Dindae Sheena117 Bukit Batok W. Ave. 6, #15-240,Singapore 650117, Singapore (SGP)

Barbara Shinn-CunninghamBoston Univ., Dept. Cogs & Neural Systems& Biomed Eng., 677 Beacon St., Boston, MA02215 (BOS)

Arthur W. Smith Jr.145 Southard Dr., Manahawkin, NJ 08050(NY)

Ivo Soares de JesusBaden Brasil SA, Rua Cardosa Marinho 28,Santo Cristo RJ 20220370, Brazil (BZ)

Alois K. SontacchiMuchargasse 39/3, Graz AT 8010, Austria(AU)

Gerald T. Stewart

15 Bothwell St., Balaclava, VC 3183, New South Wales, Australia (SYD)

Mohamed A TasribBlk 428, #02-2636 Ave. 3, Ang Mo Kio,Singapore 560428, Singapore (SGP)

Bobby W. TaylorAll Pro Sound, 806 Beverly Pkwy.,Pensacola, FL 32505

Nicolas Teichner10 rue Etienne Dolet, Paris, FR 75020,France (FR)

Shin-chi UeokaSona Corp., Yayoicho 2-19-9 Nakano-ku,Tokyo 164-0013, Japan (JA)

Hans J. VesterbergStorgatan 48, Lulea SE 97231, Sweden (SWE)

Victor VisserGeneral Stedmanstraat 180, Eindhoven NL5623 HZ, Netherlands (NE)

Martin A. Wand175 Sunset Dr., Plano, TX 75094

Robert V. Wawoe2953 Maple Grove Pl., Oveido, FL 32765

William B. WeeksHCJB Work Radio, Casilla 17-17-691, Quito,Ecuador

Robert B. Weymouth4 Birch Ln., Pelhan, NH 03076 (BOS)

David L. Williams17 School Ave., Bedford, MA 01730 (BOS)

Neill A. WoodgerArup Acoustics, 155 Ave. of the Americas,New York, NY 10013 (NY)

Fabio Augusto SF ZacariaFz Audio, Av: Alberto Jafet-191, Diadema,SP 09951 110, Brazil (BZ)

Manfred ZazziSonnenbergstrasse 13, Oberengstringen, CH8102, Switzerland (SWI)

John S. Zubroski820 Tenplos St., Bettendorf, IA 52722(UMW)

Johua M. Zulu5115 Caroline, Houston, TX 77004

Uki AyyappanNo. 6 Bharath Colany, Alwarthirunagar (Pn),Valasaravakkam, Chennai 600 087, India(IND)

Tom J. Beno3645 S. Bascom Ave. #1, Campbell, CA95008 (SF)

Michael P. Chanay29166 County Rd. M4, Dobres, CO 81323(CO)

Peter CobbinAbbey Road Studios, 3 Abbey Road, St.Johns Wood NW8 9AY, UK (BR)

ASSOCIATES


MANAGER OF THE RECORDINGSTUDIOS AT THE

UNIVERSITY OF IOWA TO BEGIN AUGUST 1, 2002.Salary is $31,830-commensurate.

Administrative responsibilities; develop andmonitor budget. Supervise the recording,editing, and digital mastering of musicalevents and recording sessions in addition tofaculty CD projects. Coordinate the support ofequipment for faculty and the School ofMusic. Teach "Recording Techniques" classand other possible recording/technologyclasses. Please contact the search committeefor a complete job description. Requiredqualifications are Masters degree in relatedfield or an equivalent combination ofeducation and experience. Ability to read andedit music from score. Demonstratedleadership ability and experience in workingeffectively with a professionally activeperformance faculty. Desirable: 1-3 yearsexperience directly related to the duties andresponsibilities is desirable, 3-5 year highlydesirable. Sensitivity for, and knowledge of thestandard repertoire and contemporary music.Undergraduate teaching experience. Review ofapplications will begin March 27, 2002 andcontinue until position is fi l led. Sendapplication letter, résumé of qualifications andexperience, and the names, addresses andphone numbers of three references to: Dan Moore, Chair, Recording Studios Search

Committee, 1006 Voxman Music Building,University of Iowa, Iowa City, IA 52242.The University of Iowa is an affirmative

action/equal opportunity employer. Womenand minorities are encouraged to apply.

AdvertiserInternetDirectoryETANI Electronics ..............................313www.etani.co.jp

Liberty Instruments, Inc.....................309www.libinst.com

Purebits ...............................................313www.purebits.com

*That Corporation ...............................277www.thatcorp.com

University of Iowa...............................312www.uiowa.edu/~music

*AES Sustaining Member.

MEMBERSHIP

INFORMATION

Tony CollingeEMT Building, 1A Kings Rd., Heath,Birmingham B14 6TV, UK (BR)

Russell Dawson-Butterworth17 Blandford Gardens, Peterborough PE14RW, UK (BR)

Barry G. Dean1902 S. Locust, Pittsburg, KS 66762 (KC)

Lenny Franchi32 Marley Rd., London NW10 8BB, UK(BR)

Byron J. Funk1501 S. Pine, Pittsburg, KS 66762 (KC)

Nicky GrahamMaximum Music, 9 Heathmans Rd., ParsonsGreen, London SW6 4TJ, UK (BR)

Drago Grandic1515 Litchfield Rd., Oakville, Ontario L6H5P4, Canada

Ian GreavesThe Sanctuary, 98D Goldhurst Terrace,South Hampstead, London NW6 3HS, UK(BR)

Phil HardingWillow Barn, Wrenshall Farm, Walsham-Willows, Bury St. Eds, Suffolk IP31 3AS,UK (BR)

Christian Haunt3B Belsham St., Hackney, London E9 6NG,UK (BR)

Ray Hedges36 Sandalwood Ave., Chertsey, SurreyKT16 9PB, UK (BR)

Ozgur Izmirli23 River Ridge Rd., New London, CT 06320(BOS)

Thomas W. JohnsonEffigy Studios LLC, 803 Vester Ave.,Ferndale, MI 48220 (DET)

Arjun Kalavidaru171/3 2nd Main, Chamarajpet, Bangalore560018, India (IND)

Aravind KiggalAravind Studios, 617 4th Cross 5th Main,Honunantrages Karnataka, Bangalore, India(IND)

Angela J. Limb11 Brookview Ct., Ho-Ho-Kus, NJ 04723(NY)

John LundstonAlchemea College of Audio Engineering,Windsor Centre, Windsor 57, N1 8QG UK(BR)

Ian C. McDonaldIvy Cottage, Middle Winterslow, Salisbury,Wilts. SP5QS, UK (BR)

Jon I. NeumanVeneklasen Associates Inc., 1711 SixteenthSt., Santa Monica, CA 90404 (LA)

Magesh Palaniappan

No. 9 Akshaya Apt., #4 I Floor, PinjalaSuaramanian St., T. Naran Channai-17, India(IND)

Stephen ParrHear No Evil Studios, 6 Lillie Yard, SW61UB London, UK (BR)

Neil Pickles77 Thornton Ave., London SW2 4BD, UK(BR)

Tony PlattInch House, Lowell St., Pury End, UK (BR)

Krishnan RajendranNew No. 109 Old No. 5812 V.V. Koil St.,Choolai, Chennai Tamilnadu 600112, India(IND)

M. RaviNew No. 24 Chari Street, T. Nagar Chennai,Tamilnadu, India (IND)

Srinivas Reddy30 B/2 Safalya, 19in ‘A’ Road, Khar (W)Mumbai, India (IND)

Thomas Ruff811 Tinton Ave., Tinton Falls, NJ 07724(NY)

Harish G. SamtaniStereovision, 846-A Mount Rd., Chennai 600002, India (IND)

Torsten SielaffStargarder Str. 54, DE 60388, Frankfurt,Germany

Thomas B. Simpson111 Norma Ln. (RR#1), Carp, Ontario K0A1L0, Canada

Dan Smith301 Floral Dr., Tampa, FL 33613

Albert M. SoansAnnexe 2 Kares Campus, Balmatta,Mangaione 575001, India (IND)

Kevin K. SpannSeverns Reid and Associates, 821 S. Neil St.,Champaign, IL 61820 (CH)

M. SubburahNew No. 4 13 M.G. Sangarabarani St.,Saligraman, Chennai 600 093, India (IND)

Mike Thorne59 Dingletown Rd., Greenwich, CT 06830(BOS)

J. VenkatesanPrime Source Technologies Pvt. Ltd., New24 Ramakrishna St., T-Nagar, Chennai600017, India (IND)

Bjarne Vonger-LorenzenGrantoftevej 4B, DK 3500, Vaerlose,Denmark (DA)

Jayney B. WallickMackie Designs Inc., 16220 Wood Red Rd.NE, Woodinville, WA 98072 (PNW)

Blue Weaver28 Craven Hills Mews, London W2 3DY,UK (BR)


Loudspeaker Test System : S-251• High-speed tests for production lines• Rub & buzz tests & statistical processing

Audio Analysis System : S-260• For testing speakers & microphones• Up to 100 kHz band for DVD specifications

ETANI ELECTRONICS Co., Ltd.1-10-15 Ohmori-Honcho, Ohta-Ku,Tokyo,143-0011 JapanTel: +81-3-5763-1391 Fax: +81-3-5763-1394http://www.etani.co.jp

Audio Analysis and Sound Design

Victor Weng-Onn Voo22 Road 17/21C, Petaling Jaya, 46400Malaysia, Malaysia (MY)

Keola Akana5246 Agnes Ave. #105, Valley Village, CA91607 (CTC)

Ronny AnderssonAnkowskatavagen 83A, SE 94134, Pitea,Sweden (ULP)

Joseph D. Ash127 Allison Rd., Eagleville, TN 37132(MTSU)

Bradley W. BaisleyBox C594, 1301 E. Main St., Murfreesboro,TN 37132 (MTSU)

Cory R. Ballentine730 19th St. #A, Boulder, CO 80302

Michal Bamford16584 Nosoni Way, Apple Valley, CA 92307(SDSU)

Felix Bautista552 W. 185th St. #3, New York NY 10033(IAR)

Peter T. Bense613 Mallard Lakes Dr., Lexington, SC 29072

Andrew C. Bloomberg2520 Halls Hill Pike, Murfreesboro, TN37130 (MTSU)

Chrissa S. Bowman4330 Key Biscayne Ln. #3012, Winter Park,FL 32792 (FS)

Cortez L. Brazley15238 9th Ave., Phoenix, IL 60426 (CC)

Miguel S. BrownHampton University, 407 Moton Hall,Hampton, VA 23668 (HPTU)

Brian P. Buchanan1245 Sherwood Dr., Christansburg, VA24073 (HPTU)

Ryan M. Campbell3408 Zuni St., Denver, CO 80211 (DEN/S)

Franks J. Capobianco9821 Mira Del Rio Dr., Sacramento, CA95827 (ARC)

Sarah E. Scepansky65 Iowa Ave., Absecon, NJ 08201 (WPU)

Kaspar SchaubyNorre Alle 75, 708, Kobenhavn 0 DK 2100, Denmark (DAS)

Mathew L. Schoenberger71 Wilder St., MS 7038, Lowell, MA 01854(UL)

Ryan D. Schwartz1311 Greenland Dr., #F-17, Murfreesboro,TN 37130 (MTSU)

Aaron W. Sefton1311 Greenland, #D9, Murfreesboro, TN

37130 (MTSU)

Keith R. Sengbusch3616 Summerwind Dr., Winter Park, FL32792 (FS)

Michael J. Shaw Jr.11 Sedgwick Dr., Engelwood, CO 80110(DEN/S)

Jason A. Seibert114 S. Parkside Rd., Normal, IL 61761 (CC)

James E. Simms4700 College Oak Dr., Sacramento, CA95841 (ARC)

Brett A. Simon217 E. Davenport, Iowa City, IA 52245(BSU)

Geordy C. Sincavage4352 Autumn Breeze Way, Winter Park, FL32792 (FS)

Joseph A. Skarulis250 East Ave., Valley Stream, NY 11580(IAR)

Bryan W. Smith1946 E. 120th, 2nd Flr., Cleveland, OH44106 (OU)

Artie W. Smith III145 Southard Dr., Manahawkin, NJ 08050(WPU)

Evan B. Sonderegger41-I South Ave., Harrisonburg, VA 22801(HPTU)

Victoria Spencer2046 1/2 Delrosa Ave., Los Angeles, CA90041 (USC)

Jason W. Spencer2206 Summer Wind Dr., Winter Park, FL32792 (FS)

Vivek R SrivastavaAlexandria Research Institute, Virginia Tech,206 N. Washington St. Ste. 400, Alexandria,VA 22314

Andrew C. Starr1350 Columnie St., #7, Denver, CO 80206(DEN/S)

Joseph F. Steinwand43 Linmor Ave., Newton, NJ 07860 (WPU)

Darren S. Stroupe15 Murdock Ave., #2, Asheville, NC 28801(UNCA)

Jonathan E. Talley1301 E. Main, Box 1857, Murfreesboro, TN37132-0001 (MTSU)

Glenn P. Tietjen3063 Wiese Way, Sacramento, CA 95833(ARC)

Shannon R. TraynorP.O. Box 1964, New York, NY, 10009(IAR)

Alan M. Trevena12 Bulbecks Walk, South Woodham Ferrers,Chelmsford, Essex CM3 5ZN, UK (BR)

Carl G. Troia Jr.11781 Goodwood Blvd., Baton Rouge, LA70815 (NO)

Eric T. Velazquez2602 Calverac Pl., Ontario, CA 91761 (CTC)

Michael E. VidettoLebanon Valley College, North College 100,101 N. College Ave., Annville, PA 17003

Chris R. Villanueva7875 Fawn Trail Way, Antelope, CA 95843(ARC)

Andrew L. WallnerVander Werp 266, Calvin College, GrandRapids, MI 49546

Toby L. Warren16732 Muni Rd., Apple Valley, CA 92307(SDSU)

Danielle L. Wasmer5634 S. Lansing Way, Englewood, CO 80111(DEN/S)

Luke C. WassermanSmith House #101, 11311 Juniper Rd.,Cleveland, OH 44106 (OU)

Eric D. Weaver150 S. Rossevelt Rd., #1159, Mesa, AZ85202-1078 (CRAS)

Jared M. Wechsler344 N. Goldenrod Rd., #134, Winter Park,FL 32792 (FS)

Michael W. Weinstein303 E. Lytle St., Murfreesboro, TN 37130(MTSU)

Rebecca E. Welford Sale111 Cumberland Ave., #1, Asheville, NC28801 (UNCA)

Leslie A. Will920 Greenland Dr., #301, Murfreesboro, TN37130 (MTSU)

Ben D. Willt1601 Hillcrest Dr., #Y30, Manhattan, KS66502 (KC)

Corey A. Winer3440 Goldenrod Rd., #915, Winter Park, FL32792 (FS)

Patrick J. WolfeChurchill College, Cambridge CB3 0DS, UK(BR)

Elizabeth H. Wood19A Northumberland Park, Tottenham,London N17 0TA, UK (BR)

Ricardo A. Zayas3233 Haven Ave., Ocean City, NJ 08226(WPU)

Jonathan J. Zenz35 S. Congress St., Rm. 604, Athens, OH45701 (UC)

STUDENTS

MEMBERSHIP

INFORMATION


ANTHOLOGY SERIES

Collected papers from theAES’s international confer-ences are reprinted here from the authors' original manu-scripts. Books are bound indurable paper covers and areshrinkwrapped.

Proceedings of the AES 3rd Interna-tional Conference: PRESENT ANDFUTURE OF DIGITAL AUDIO, Tokyo,Japan, 1985 June 20-21. Presented in Tokyo at an internationalgathering of experts in digital audio, thesepapers include digital audio overview, dig-ital audio technology in studio broadcast-ing, DAD technology, and DAT recordertechnology. 228 pages

Proceedings of the AES 4th Interna-tional Conference: STEREO AUDIOTECHNOLOGY FOR TELEVISIONAND VIDEO, Rosemont, Illinois, 1986May 15-18.

These manuscripts are devoted to thechallenges and problems of stereo audio. The papers cover transmission,duplication, consumer product technol-ogy, and economic considerations.

336 pages

Proceedings of the AES 5th Interna-tional Conference: MUSIC AND DIGITAL TECHNOLOGY, Los Angeles,California, 1987 May 1-3.These papers were presented by engi-neers and experts in music and comput-ers. The subjects include the past, pre-sent, and future of digital music-making;digital synthesizer design; digital worksta-tions for music; composition; computernetworks for music; music software forpersonal computers. 248 pages

Proceedings of the AES 6th International Conference: SOUNDREINFORCEMENT, Nashville, Ten-nessee, 1988 May 5-8.

These papers were written by engi-neers and the savants of sound rein-forcement. They cover the history ofsound reinforcement, new frontiers inapplications, computers, new concepts,electronic architecture, and sound rein-forcement in the future. 600 pages

Proceedings of the AES 7th Interna-tional Conference: AUDIO IN DIGITALTIMES, Toronto, Ontario, Canada,1989 May 14-17.Experts in digital audio explore digital au-dio from the history, basics, hardware,software to the ins and outs. 384 pages

Proceedings of the AES 8th Interna-tional Conference: THE SOUND OFAUDIO,Washington,D.C., 1990 May3-6.These papers are devoted to theprogress of sound, including perception,measurement, recording and reproduc-tion. The book is fully illustrated.

384 pages

The AES's renowned seriesof collected papers ofarchival quality are repro-duced exactly as they ap-peared in the Journal andother authoritative sources.These books measure 81⁄4inches (209.6 mm) by 111⁄2inches (285.8 mm), are

bound in durable paper covers, andare shrinkwrapped for safe shipment.

DISK RECORDING VOL.1: GROOVEGEOMETRY AND THE RECORDINGPROCESS edited by Stephen F.Temmer. These papers describe themajor contributions to the art of diskrecording in the areas of groove geome-try, cutterheads and lathes, styli andlacquers, pressings, and high-densitydisk technology. 550 pages

DISK RECORDING VOL. 2: DISK PLAY-BACK AND TESTING edited by StephenF. Temmer. Written by experts, thesepapers discuss the subjects of diskplayback, disk pickups, tone arms andturntables, and quality control.

550 pages

LOUDSPEAKERS VOL.1 edited byRaymond E. Cooke. These papers(from 1953 to 1977) were written by theworld's greatest transducer experts andinventors on the design, construction,

and operation of loudspeakers.448 pages

LOUDSPEAKERS VOL.2 edited by Ray-mond E. Cooke. Papers from 1978 to1983 cover loudspeaker technology, extending the work initiated in Vol. 1.

464 pages

LOUDSPEAKERS VOL. 3: Systemsand Crossover Networks edited byMark R. Gander. These papers with com-ments and corrections were publishedfrom 1984 through 1991 in the area ofloudspeaker technology. With a compan-ion volume on transducers, measurementand evaluation, the publication extendsthe work of the first two volumes. An ex-tensive list of related reading is included.

456 pages

LOUDSPEAKERS VOL. 4: Transduc-ers, Measurement and Evaluation edited by Mark R. Gander. Papers withcomments and corrections explore thissubcategory from 1984 through 1991. Abibliography lists essential titles in thefield. 496 pages

MICROPHONES edited by Louis A.Abbagnaro. These papers cover cali-bration and testing, general purposemicrophones, directional microphones,miniature types, and associated elec-tronic circuits. 392 pages

SOUND REINFORCEMENT edited byDavid L. Klepper. These papers deal withthe significant aspects of the develop-ment of sound-reinforcement technologyand its practical application to sound sys-tem design and installation. 339 pages

SOUND REINFORCEMENT VOL. 2 edit-ed by David L. Klepper. These paperswith comments and corrections were orig-inally published between 1967 and 1996.In addition to extending the work of thefirst anthology on this vital topic, Volume 2adds earlier papers now considered semi-nal in the original development of thetechnology. 496 pages

STEREOPHONIC TECHNIQUES edit-ed by John M. Eargle. These articlesand documents discuss the history, de-velopment, and applications of stereo-phonic techniques for studio technology,broadcasting, and consumer use.

390 pages

TIME DELAY SPECTROMETRY editedby John R. Prohs. Articles of Richard C.Heyser’s works on measurement, analy-sis, and perception are reprinted from thepages of the JAES and other publica-tions, including Audio magazine andIREE Australia. A memorial to the author’s work, it contains fundamentalmaterial for future developments in audio.

280 pages

Continued

papers, and conference papers published by the AES between 1953and 2000. The approximately 9000papers and articles are stored inPDF format, preserving the original

documents to the highest fidelitypossible, while also permitting full-text and field searching. The librarycan be viewed on Windows, Mac,and UNIX platforms.

This 17-disk elec-tronic librarycontains most

of the Journal articles, convention

ELECTRONIC LIBRARY

AES SPECIALPUBLICATIONS

PROCEEDINGS

Proceedings of the AES 9th International Conference: TELE-VISION SOUND TODAY AND TOMORROW, Detroit , Michigan,1991 February 1-2.These fully illustrated papers explore thelatest in audio and video technologies.

256 pages

Proceedings of the 10th InternationalAES Conference: IMAGES OF AUDIO,London, UK, 1991 September 7–9.Papers cover recording and postproduc-tion, digital audio bit-rate reduction, digi-tal audio signal processing and audio forhigh definition television plus a 100-pagetutorial on digital audio. 282 pages

Proceedings of the 11th InternationalAES Conference: AUDIO TEST &MEASUREMENT, Portland, Oregon,1992 May 29–31.These papers describe both the engi-neering and production aspects of test-ing including state-of-the-art techniques.Authors examine electronic, digital, andacoustical measurements, bridging thegap between subjective and objectivemeasurement to advance the science ofaudio measurement. 359 pages

Proceedings of the AES 12th Inter-national Conference: PERCEPTIONOF REPRODUCED SOUND, Copen-hagen, Denmark, 1993 June 28–30.Papers by experts in the science of human perception and the application ofpsychoacoustics to the audio industry explore the performance of low bit-ratecodecs, multichannel sound systems,and the relationships between sound andpicture. 253 pages

Proceedings of the AES 13th Interna-tional Conference: COMPUTER-CON-TROLLED SOUND SYSTEMS, Dallas,Texas, 1994 December 1–4.A complete collection of the papers pre-sented at this conference covers all aspects of computer-controlled soundsystems including product design, imple-mentation and real-world applications.

372 pages

Proceedings of the AES 15th Internation-al Conference: AUDIO, ACOUSTICS &SMALL SPACES, Copenhagen, Den-mark, 1998 October 31–November 2.Reproduction of sound in small spaces,such as cabins of automobiles, trucks, andairplanes; listening and control rooms; anddomestic rooms is addressed in detail inthe papers included. 219 pages

Proceedings of the AES 16th Inter-nat ional Conference: SPATIALSOUND REPRODUCTION, Rovaniemi,Finland, 1999 April 10–12.Var ious aspects of spat ial sound reproduction (perception, signal pro-

cessing, loudspeaker and headphonereproduction, and applications) arecovered in this volume. 560 pages

Also available on CD-ROM

Proceedings of the AES 17th Interna-tional Conference: High-Quality Audio Coding, Florence, Italy, 1999September 2-5.The introduction of new, high-capacity media, such as DVD and the Super Audio CD, along with the latest develop-ments in digital signal processing, ICdesign, and digital distribution of audiohave led to the widespread utilization ofhigh-quality sound. These new tech-nologies are discussed. 352 pages


Proceedings of the AES 18th Interna-tional Conference: Audio for Informa-tion Appliances, Burlingame, Califor-nia, 2001 March 16-18.This conference looked at the new breedof devices, called information appliances,created by the convergence of consumerelectronics, computing, and communica-tions that are changing the way audio iscreated, distributed, and rendered.

Available on CD-ROM only

Proceedings of the AES 19th Inter-national Conference: SurroundSound—Techniques, Technology,and Perception, Schloss Elmau, Germany, 2001 June 21-24.The emphasis of the conference was onsurround sound for mainstream recordingand broadcasting applications, according tothe so-called "5.1" or 3/2-stereo standardspecified in ITU-R BS.775. 464 pages


Proceedings of the AES 20th Interna-tional Conference: Archiving, Restora-tion, and New Methods of Recording,Budapest, Hungary, 2001 October 5-7.This conference assessed the latest developments in the fields of carrierdegradation, preservation measures, digi-tization strategies, restoration, and newperspectives in recording technology.

211 pagesAlso available on CD-ROM

Proceedings of the AES DSP UK Confer-ence: DIGITAL SIGNAL PROCESSING,London, UK, 1992 September 14–15.Papers cover issues crucial to the appli-cation of DSP in both domestic and pro-fessional audio. These include proces-sor choice, filter design and topology,code development, psychoacoustic con-siderations, and a variety of applications.

239 pages

Proceedings of the AES DAI UK Conference: DIGITAL AUDIO INTER-CHANGE, (DAI) London, UK, 1993May 18–19.

Since audio is part of a multimedia envi-ronment, there are more questions relat-ed to the effective exchange of digital audio signals between equipment. Thesepapers explore them. 135 pages

Proceedings of the AES UK Confer-ence: Managing the Bit Budget, (MBB)London, UK, 1994 May 16–17.The boundaries of digital audio haveextended in different directions in termsof bit rate and sound quality. These papers by experts address the manycomplex aspects of digital/analog con-version, signal processing, dynamicrange, low bit-rate coding, and perfor-mance assessment. 189 pages

Proceedings of the AES DAB UK Con-ference: THE FUTURE OF RADIO,London, UK, 1995 May 2–3.These papers provide cutting-edge information on digital audio broadcast-ing and a review of competing digitalradio services. 143 pages

Proceedings of the AES ANM UK Con-ference: AUDIO FOR NEW MEDIA,London, UK, 1996 March 25–26.The papers in this valuable book are avital reference for those involved in thetechnologies. 117 pages

Proceedings of the AES UK Confer-ence: THE MEASURE OF AUDIO(MOA), London, UK, 1997 April 28–29.Audio test and measurement is beingrevolutionized by advancing technology.Learn about the various aspects of thisimportant topic from papers written byprofessionals in the field. 167 pages

Proceedings of the AES UK Conference:MICROPHONES AND LOUDSPEAK-ERS: THE INS AND OUTS OF AUDIO,LONDON, UK, 1998 March 16–17. These papers update the transducerspecialist and nonspecialist with thelatest in microphone and loudspeak-er development. They explore the influence on equipment and workingpractices in the audio industry.

135 pages

Proceedings of the AES UK Confer-ence: Audio—The Second Century,London, UK, 1999 June 7-8.The convergence of the computer andaudio industries has introduced benefitsand challenges. These are covered inpapers written by leading experts fromaround the world. 176 pages

Proceedings of the AES UK Confer-ence: Moving Audio, Pro-Audio Net-working and Transfer, 2000 May 8-9.These papers describe how the capacityand speed of new computer systems andnetworks bring flexibility, convenience, andutility to professional audio. 134 pages

(Amsterdam, 1998 May 16-19) areavailable on one CD-ROM. IndividualCD-ROMs are available for the 105th to110th conventions. Contact Andy Velozat Headquarters [email protected].

Internet: For preprint lists, pricesand a search engine see the AESWeb site.

and ordering see the AES Web site orcontact Andy Veloz at [email protected].

CD-ROMs: Preprints presented at the103rd Convention in New York (1997September) and the 104th Convention

Printed Form: Many of the papers pre-sented at AES conventions are preprint-ed and available in printed form individu-ally or as sets. For preprint lists, prices

CONVENTION PREPRINTS

Complete the order form on the following page and mail it to one of the addresses below:(Please note appropriate methods of payment for each location).

Headquarters: Orders must be prepaid by Visa, Mastercard,or American Express (AMEX) credit card or in U.S. dollarsdrawn on a U.S. bank. Make checks payable to Audio Engi-neering Society. Send completed order form to:

Audio Engineering Society60 East 42nd Street, Room 2520New York, NY 10165-2520, USA

Tel: +1 212-661-8528Fax: +1 212-682-0477e-mail: [email protected]

Continental Europe: Send completed order form, WITH-OUT payment but indicating method of payment (credit card,Eurocheque or cheque) in local currency to:

Michael Williams, Ile du Moulin62 bis Quai de l'Artois, F 94170 Le Perreux sur Marne, France.

Tel: +33 1 4881 4632 • Fax: +33 1 4706 0648e-mail: [email protected]

You will receive a pro forma invoice in the required currencyindicating details of payment.

AUDITORY ILLUSIONSAND AUDIO, Vol. 31,No. 9.Edited by Diana Deutsch.The 1983 September issueof the Journal, devoted to paradoxes in human

audio perception, explores audi-tory illusions from varied viewpoints (with

two demonstration Soundsheets)

DIGITIZATION OF AUDIO: A Compre-hensive Examination of Theory,Implementa t ion , and CurrentPractice, Vol. 26, No. 10.

JOURNAL ISSUES

ALSO AVAILABLE

The 1978 October issue of the Jour-nal features the internationally refer-enced tutor ia l paper by Barr y A.Blesser on analog-to-digital conver-sion. Implementation questions arealso examined.

SHIELDS AND GROUNDS: Safety,Power Mains, Studio, Cable andEquipment, (special excerpt).The June 1995 issue of the Journal wasa definitive and comprehensive collec-tion of information on this important top-ic. The seven papers by Neil Muncyand other experts in the field have been

Perceptual Audio Coders:What to Listen For. This isthe first educational/tutorialCD-ROM presented by theAES Technical Council on

a particular topic, combining back-ground information with specific audioexamples. To facilitate the use of highquality home playback equipment forthe reproduction of audio excerpts, thedisk can also be played back on allstandard audio CD players. Perceptual

audio coding combines elements fromdigital signal processing, coding theory,and psychoacoustics.The Audio Engineering Society Pre-sents Graham Blyth in Concert: ACD of seven selected pieces fromGraham Blyth’s recitals performed onsome of the great pipe organs.Membership pin: A gold-colored lapelpin with AES logo in blue and white. Membership certificate: A personalized

membership certificate suitable for fram-ing measures 81⁄2 inches by 11 inches.Please print your name exactly as youwould like it to appear on the certificate.

VIDEO CASSETTES:“An Afternoon with Jack Mullin” is a 1⁄2-inch VHS and PAL format cassette tapecapturing the growth of entertainmenttechnology. “A Chronology of American TapeRecording ” (VHS format)

reprinted into a convenient guide for designers and practitioners. 82 pages

Commemorative Issue... The AES:50 Years of Contributions to AudioEngineering, Vol. 46, No. 1/2. As-sembled by John J. Bubbers, guesteditor, 1998 January/February.This special issue covers the founding,development and internationalizationof the society. It includes an impres-sive group of review papers on the essential technologies in the audiofield. It is an indispensable addition toany audio library. 134 pages

CUMULATIVE INDEX 2:Journal of the Audio En-gineering Society, 1981-1987, Volumes 29-35.This edition lists all articlespublished in the Journal ofthe Audio Engineering Society from 1981–1987. Itis an easy-to-use reference

to some of the most pertinent writingspublished on the subject of audio engi-neering. It lists subject categories withpaper titles, author and coauthor, volume,issue numbers, first page, and Journalcategory classification. 88 pages

CUMULATIVE INDEX 3: Journal ofthe Audio Engineering Society ,1988-1995, Volumes 36-43. This edition is a continuation of Index 2.

104 pages

CUMULATIVE INDEX OF CONVEN-TION PREPRINTS, VOL. 1: Audio En-gineering Society, 9th Convention,1957 October–64th Convention,1979 November.This edition is a complete referenceguide to convention preprints fromthe 9th through the 64th conventions.It is divided into three sections with

author, sub jec t , and numer ica l listings. 288 pages

CUMULATIVE INDEX OF CONVEN-TION PREPRINTS, VOL. 2: AudioEngineering Society, 65th Convention,1980 February–93rd Convention, 1992October.This edition is a continuation of preprintslisted in Vol. 1 above. 368 pages

Books measure 8 1/4 inches (209.6 mm)by 11 inches (279.4 mm) and are paperbound.

CUMULATIVE INDICES

digital signals is amply covered in thisimportant publication. Pertinent topicsand authors—all experts in theirfields—were carefully selected by the edi-tors. The 16 reviewed and editedmanuscripts are presented here for thefirst time. It is an essential reference for

understanding the current and futuretechnology of audio codecs. 208 pages

MAGNETIC RECORDING: THE UPSAND DOWNS OF A PIONEER—THEMEMOIRS OF SEMI JOSEPH BEGUN,edited by Mark Clark. 168 pages

COLLECTED PAPERSON DIGITAL AUDIO BIT-RATE REDUCTION editedby Neil Gilchrist andChrister Grewin.The emerging technologyof reducing the bit rate of

SPECIAL PUBLICATIONS

continued

ORDERING INFORMATION

ORDER FORM FOR ALL PUBLICATIONS EXCEPT STANDARDS: Check box next to the item you are ordering andwrite the quantity and total amount in the space provided. Mail the form to one of the addresses shown. Postage isprepaid. Fill in all information and allow 4-6 weeks for delivery.

QUANTITY US DOLLARS ($) TOTAL AMT.

ANTHOLOGIES ____Disk Recording, Vol. 1 $ 30.00 $ 40.00 _____ ____Disk Recording, Vol. 2 30.00 40.00 _____ ____Loudspeakers, Vol. 1 30.00 40.00 _____ ____Loudspeakers, Vol. 2 30.00 40.00 _____ ____Microphones 30.00 40.00 _____ ____Sound Reinforcement 30.00 40.00 _____ ____StereophonicTechniques 30.00 40.00 _____ ____Time Delay Spectrometry 30.00 40.00 _____

ORDERS OF 2 OR MORE (ANY COMBINATION OF THE ABOVE), PER VOLUME 27.00 37.00 _____

____Loudspeakers, Vol. 3 38.00 50.00 _____ ____Loudspeakers, Vol. 4 38.00 50.00 _____ ____Sound Reinforcement, Vol. 2 38.00 50.00 _____

ORDERS OF 2 OR MORE, PER VOLUME 35.00 47.00 _____

____3rd International Conference $ 28.00 $ 40.00 _____ ____4th International Conference 28.00 40.00 _____ ____5th International Conference 28.00 40.00 _____ ____6th International Conference 28.00 40.00 _____ ____7th International Conference 28.00 40.00 _____ ____8th International Conference 28.00 40.00 _____ ____9th International Conference 28.00 40.00 _____ ____10th International Conference 28.00 40.00 _____ ____11th International Conference 28.00 40.00 _____ ____12th International Conference 28.00 40.00 _____ ____13th International Conference 28.00 40.00 _____ ____15th International Conference 28.00 40.00 _____ ____DSP UK Conference, 1992 28.00 40.00 _____ ____DAI UK Conference, 1993 28.00 40.00 _____ ____MBB UK Conference, 1994 28.00 40.00 _____ ____DAB UK Conference, 1995 28.00 40.00 _____ ____ANM UK Conference, 1996 28.00 40.00 _____ ____MOA UK Conference, 1997 28.00 40.00 _____ ____MAL UK Conference, 1998 28.00 40.00 _____ ____ASC UK Conference, 1999 28.00 40.00 _____ ____Moving Audio UK Conference, 2000 28.00 40.00 _____

ORDERS OF 2 OR MORE (ANY COMBINATION OF THE ABOVE), PER VOLUME 26.00 36.00 _____ ____16th International Conference 40.00 60.00 _____ ____16th CD-ROM 40.00 60.00 _____ ____17th International Conference 40.00 60.00 _____ ____17th CD-ROM 40.00 60.00 _____ ____18th CD-ROM only 40.00 60.00 _____ ____19th International Conference 40.00 60.00 _____ ____19th CD-ROM 40.00 60.00 _____ ____20th International Conference: Archiving, Restoration, 2001 40.00 60.00 ____ ____20th CD-ROM 40.00 60.00 _____

ORDERS OF 2 OR MORE (ANY COMBINATION OF THE ABOVE), PER VOLUME 37.00 56.00 _____

CONFERENCE PROCEEDINGS AND COLLECTED PAPERS

MEMBER NONMEMBER

ELECTRONIC LIBRARY

____17-disk electronic library $ 395.00 $ 495.00 _____ — — Institution Price, single user 795.00 895.00 _____ — — Institution Price, networked users 1195.00 1295.00 _____

____Search Disk Only 50.00 75.00 _____

Continued

ORDERING INFORMATION Continued

United Kingdom: Orders must be prepaid in pounds sterling. Contact the U.K. office for prices. Send completed order form to:

AES British SectionP.O. Box 645

Slough, SL1 8BJ, UKTel: Burnham +44 (0) 1628 663725

Fax: +44 (0) 1628 667002 • e-mail: [email protected]

AES STANDARDS AND INFORMATION DOCUMENTS

NAME ____________________________________________________________________ AES MEMBERSHIP NUMBER _____________________________

ADDRESS ________________________________________________________________________________________________________________________

CITY, STATE & COUNTRY ______________________________________________________________ ZIP OR POSTAL CODE ________________________

PHONE _________________________________ FAX _____________________________ E-MAIL _______________________________________________

SIGNATURE ______________________________________________________________________________________________________________________

VISA, Master Card, and AMEX_____________________________ EXP. DATE ____________________ AMOUNT ENCLOSED _________________________

(First) (Family)

Standards may be obtained by clicking on “Standards in print” at www.aes.org/standards/.

Oct. 01

QUANTITY US DOLLARS ($) TOTAL AMT.MEMBER NONMEMBER

SPECIAL PUBLICATIONS

For Preprint lists, prices and ordering (in printed form or on CD-ROM)contact Andy Veloz @ [email protected] or see the AES Web site.

____Papers on Digital Audio Bit-Rate Reduction $ 34.00 $ 68.00 _____ ____Magnetic Recording: The Memoirs of Semi Joseph Begun 15.00 20.00 _____

CUMULATIVE INDICES ____Cumulative Index 2 (Journal) $ 15.00 $ 21.00 _____ ____Cumulative Index 3 (Journal) 15.00 21.00 _____ ____Cumulative Index, Vol. 1 (Preprints) 25.00 40.00 _____ ____Cumulative Index, Vol. 2 (Preprints) 25.00 40.00 _____

____Perceptual Audio Coders CD-ROM $ 15.00 20.00 _____ ____Graham Blyth in Concert CD $ 14.00 _____ ____Membership Certificate $ 30.00 _____ ____AES Lapel pin 15.00 _____

An Afternoon with Jack Mullin _____ ____ NTSC VHS Tape 29.95 39.95 _____ ____ PAL-VHS format 39.95 49.95 _____ ____A Chronology of American Tape Recording (VHS only) 35.00 45.00 _____ ____Back Issues (Please specify volume and number)

JOURNAL ISSUES

ALSO AVAILABLE

Prices subject to change without notice.Total ———

____Auditory Illusions and Audio $ 10.00 $ 15.00 _____ ____Digitization of Audio 10.00 15.00 _____ ____Shields and Grounds (special excerpt) 10.00 15.00 _____ ____Commemorative Issue... AES: 50 Years... 10.00 15.00 _____

Free Downloading of StandardsSingle copies of AES Standards in PDF form may be downloaded free from the AESSC Web page. Becausethe AESSC reserves the right to make changes in these documents, the latest printing must be download-ed before each use. These are copyrighted documents that must not be printed or circulated except in partwhere useful in standards body deliberations or with written permission of the AES Standards Committee.Any use of AES Standards information obtained from this Web site in a republishing or selling activity isexpressly prohibited.

STUDENTMEMBER

10.00

In Memoriam

Emory Cook, AES memberand audio innovator, diedon February 19. He was

89 years old. Cook was knownfor the left-right binaural disc andlaunching a recording label devoted to pushing the bound-aries of high fidelity in the 1950sand 60s.

Born in Albany, New York in1913, Cook was the only child ofLavinia and Harry Cook. He attended Phillips Exeter Academyand spent a year at M.I.T. beforeenlisting in the Army Air Corpsin 1932. In 1934, when he wasdischarged, he enrolled at CornellUniversity and graduated with anengineering degree in 1938.

After graduation, Cook workedfor the New York Power andLight Company, and then forCBS in the general engineeringand construction department. Subse-quently, Western Electric hired him towork in the Field Engineering Force.During World War II, while workingfor Western Electric, he designed andsupervised the construction of a fire-control radar “Trainer” and installedradar on destroyers. For this achieve-ment, he received a Commendationwhich stated, “The Trainer [was]enthusiastically received by the Ser-vice and is, insofar as destroyer Fire-Control Radar is concerned, one ofthe most valuable training aids everdeveloped.” Cook was extremelyproud of this achievement.

A rugged New Englander who wascontinually formulating new ideas,plans and projects, Cook played acritical role in the development ofvinyl disk technology and the intro-duction of high fidelity recordings.His company, Cook Laboratories,which he started in the basement ofhis home in Floral Park, New York,

in 1945, served as an umbrella orga-nization for many of his inventions.

A newly designed feedback diskcutting head, which improved thequality of the transfer from tape tovinyl, was the first invention. Cookmarketed this cutter to be sold direct-ly to recording companies when mostmanufacturers were renting their cut-ters to record makers on a per projectbasis. In response to Cook’s cutter,MGM Records commissioned him todesign and equip three recording stu-dios. This endeavor allowed him tofamiliarize himself with the myriaddevices and aspects of recording. Hethen began to think about other modi-fications that could be made toachieve what he considered to be thatever-elusive aim of high fidelity.

As a result of his experiments onmicrophones, amplifiers and electron-ic devices such as oscillators, Cooksoon discovered that he was able torecord frequency spectrums ranging

up to 20 kHz. He made a fewsample recordings of piano andorgan recitals to show what couldbe accomplished with the properjuxtaposition of tools, then renteda small suite at the 1949 AudioFair in New York to demonstratethese novelty records. Hundredsof audiophiles flocked to his suiteto listen to these records; and thereaction from attendees was sooverwhelming that Cook immedi-ately knew where his future lay.

In 1950, he founded the compa-ny Sounds of Our Times, whichproduced and released high fideli-ty records. These recordingsdemonstrated increased frequencyrange and innovative microphoneplacement techniques. The initialrecordings were of storms, trains,airplane noises, ships whistles,babies crying, bull frogs, and a

variety of environmental sounds. Although initially designed for high-fibuffs, the records sold out as fast asthey could be pressed and proved tobe a hit among the general public.

There are many anecdotes told ofCook and his pursuit of sound. Hecould spend hours listening to andrecording the sound of rain until itsounded “wet,” or lightening until itbrought up the hairs on one’s neck.One of his most famous recordings,“Rail Dynamics,” featured the click-ety-clack of the New York CentralLine between Peekskill and NewYork City. To get the sounds, Cookperched atop trestles, hopped acrossswitches and third rails and hung pre-cariously out of the window of speed-ing trains. The results, however, wereso realistic that a neighbor, who hadjust moved in to the house next doorto Cook’s Pound Ridge barn, becamedisturbed after hearing the screechingof wheels and whistles and com-


Emory Cook1913–2002


plained the next day to his real estateagent for not informing him about therailroad that ran close to his property.The record was a sensation at the1951 Audio Fair and sealed Cook’sreputation as a high fidelity guru.

Cook was always looking for waysto improve recorded sound. In 1952,he developed the first binauralrecord. This prototype differed fromthe current monaural type in that ithad two sets of grooves to be playedwith a binaural arm with two needles.Each needle connected to a loud-speaker at either end of a room, tocreate one of the first simulatedstereo recordings.

Eventually, Cook’s interests turnedto the vinyl, itself. In 1955, he intro-duced a process called microfusion,which resulted in a powdered form ofvinyl, which had fewer imperfectionsthan the harder vinyl then in use. Cookset up a small factory out of Stamford,Connecticut, the new location of hishome-based company, in order toprocess microfusion records for all ofhis releases. The Sunday Times citedCook’s microfusion records as one ofthe outstanding developments insound recording during 1955.

Cook was a founding member andfellow of the Audio Engineering Society. He published many papers,and in 1985 received the Silver Medalfor four decades of achievement inthe recording field. Many articleshave been written about Cook duringhis lifetime, including a two-part pro-file written by Daniel Lang for theNew Yorker. In the profile, Langquotes an engineer for a large manu-facturer, who described the ceaseless-ly striving Cook as somewhat of acatalyst. “His [Cook’s] business issmall enough so that he can act as akind of trial balloon for the rest ofus… Things might get pretty stodgywithout Emory around,” he said. Described as a modest man with areverence for sound, he was loyal tofriends, family and his staff.

Cooke is survived by his wifeMartha, two stepchildren, and threegrandchildren.

In Memoriam

FOR ORDERING INFORMATIONSend e-mail to [email protected] or

contact any AES office•

USATel: +1 212 661 8528Fax: +1 212 682 0477

•Europe

Tel: +32 2 345 7971Fax: +33 1 4706 0648

•United Kingdom

Tel: +44 1628 663725Fax: +44 1628 667002

THE PROCEEDINGS OF THE AES 18TH

INTERNATIONAL CONFERENCE

AUDIO FORINFORMATION APPLIANCESChallenges, Solutions, and Opportunities

$40.00 Members $60.00 Nonmembers

This conference looked at the new breed of devices,called information appliances, created by the convergence of consumer electronics, computing, andcommunications that are changing the way audio is created, distributed, and rendered.

Available on

CD-ROM


EASTERN REGION,USA/CANADA

Vice President:Jim Anderson12 Garfield PlaceBrooklyn, NY 11215Tel. +1 718 369 7633Fax +1 718 669 7631E-mail [email protected]

UNITED STATES OFAMERICA

CONNECTICUT

University of HartfordSection (Student)Howard A. CanistraroFaculty AdvisorAES Student SectionUniversity of HartfordWard College of Technology200 Bloomfield Ave.West Hartford, CT 06117-1599Tel. +1 860 768 5358Fax +1 860 768 5074 E-mail [email protected]

FLORIDA

Full Sail Real WorldEducation Section (Student)Noble HibbsAES Student SectionFull Sail Real World Education3300 University Blvd., Ste. 160Winter Park, FL 32792Home Tel. +1 407 657 7332Fax +1 407 671 1112E-mail [email protected]

University of Miami Section(Student)Ken Pohlmann, Faculty AdvisorAES Student SectionUniversity of MiamiSchool of MusicPO Box 248165Coral Gables, FL 33124-7610Tel. +1 305 284 6252Fax +1 305 284 4448E-mail [email protected]

GEORGIA

Atlanta SectionRobert Mason

2712 Leslie Dr.Atlanta, GA 30345Office Tel./Fax +1 770 938 7320E-mail [email protected]

MASSACHUSETTS

Berklee College of MusicSection (Student)Eric Reuter, Faculty AdvisorBerklee College of MusicAudio Engineering Societyc/o Student Activities1140 Boylston St., Box 82Boston, MA 02215Tel. +1 978 443 7871Fax +1 978 443 7873E-mail [email protected]

Boston SectionEric Reuter1416 South Rd.Rye, NH 03870Tel. +1 603 964 6383E-mail [email protected]

University ofMassachusetts–Lowell Section(Student)John Shirley, Faculty AdvisorAES Student ChapterUniversity of Massachusetts–LowellDept. of Music35 Wilder St., Ste. 3Lowell, MA 01854-3083Tel. +1 978 934 3886Fax +1 978 934 3034E-mail [email protected]

Worcester PolytechnicInstitute Section (Student) William MichalsonFaculty AdvisorAES Student SectionWorcester Polytechnic Institute100 Institute Rd.Worcester, MA 01609Tel. +1 508 831 5766E-mail [email protected]

NEW JERSEY

William Paterson UniversitySection (Student)David Kerzner, Faculty AdvisorAES Student SectionWilliam Paterson University300 Pompton Rd.Wayne, NJ 07470-2103Tel. +1 973 720 3198

Fax +1 973 720 2217E-mail [email protected]

NEW YORK

Fredonia Section (Student)Bernd Gottinger, Faculty AdvisorAES Student SectionSUNY–Fredonia1146 Mason HallFredonia, NY 14063Tel. +1 716 673 4634Fax +1 716 673 3154E-mail [email protected]

Institute of Audio ResearchSection (Student)Noel Smith, Faculty AdvisorAES Student SectionInstitute of Audio Research 64 University Pl.New York, NY 10003Tel. +1 212 677 7580Fax +1 212 677 6549E-mail [email protected]

New York SectionRuss HammG Prime Limited1790 Broadway, Ste. 402New York, NY 10019Tel. +1 212 765 3415Fax +1 212 581 8938E-mail [email protected]

NORTH CAROLINA

University of North Carolinaat Asheville Section (Student)Wayne J. KirbyFaculty AdvisorAES Student SectionUniversity of North Carolina at

AshevilleDept. of MusicOne University HeightsAsheville, NC 28804Tel. +1 828 251 6487Fax +1 828 253 4573E-mail [email protected]

PENNSYLVANIA

Carnegie Mellon UniversitySection (Student)Thomas SullivanFaculty AdvisorAES Student SectionCarnegie Mellon UniversityUniversity Center Box 122Pittsburg, PA 15213

Tel. +1 412 268 3351E-mail [email protected]

Duquesne University Section(Student)Francisco RodriguezFaculty AdvisorAES Student SectionDuquesne UniversitySchool of Music600 Forbes Ave.Pittsburgh, PA 15282Fax +1 412 396 5479E-mail [email protected]

Pennsylvania State UniversitySection (Student)Brian TuttleAES Penn State Student ChapterGraduate Program in Acoustics217 Applied Science Bldg.University Park, PA 16802Home Tel. +1 814 863 8282Fax +1 814 865 3119E-mail [email protected]

Philadelphia SectionPrice Stevenson1591 Broad Run Rd.Downingtown, PA 19335Tel. +1 610 383 1083Fax +1 610 383 4007E-mail [email protected]

VIRGINIA

Hampton University Section(Student)Bob Ransom, Faculty AdvisorAES Student SectionHampton UniversityDept. of MusicHampton, VA 23668Office Tel. +1 757 727 5658,

+1 757 727 5404Home Tel. +1 757 826 0092Fax +1 757 727 5084E-mail [email protected]

WASHINGTON, DC

American University Section(Student)Benjamin TomassettiFaculty AdvisorAES Student SectionAmerican UniversityPhysics Dept.4400 Massachusetts Ave., N.W.Washington, DC 20016

DIRECTORY

SECTIONS CONTACTS

The following is the latest information we have available for our sections contacts. If youwish to change the listing for your section, please mail, fax or e-mail the new informationto: Mary Ellen Ilich, AES Publications Office, Audio Engineering Society, Inc., 60 East42nd Street, Suite 2520, New York, NY 10165-2520, USA. Telephone +1 212 661 8528.Fax +1 212 661 7829. E-mail [email protected].

Updated information that is received by the first of the month will be published in thenext month’s Journal. Please help us to keep this information accurate and timely.

Tel. +1 202 885 2746Fax +1 202 885 2723E-mail [email protected]

District of Columbia SectionJohn W. ReiserDC AES Section SecretaryP.O. Box 169Mt. Vernon, VA 22121-0169Tel. +1 703 780 4824Fax +1 703 780 4214E-mail [email protected]

CANADA

McGill University Section(Student)John Klepko, Faculty AdvisorAES Student SectionMcGill UniversitySound Recording StudiosStrathcona Music Bldg.555 Sherbrooke St. W.Montreal, Quebec H3A 1E3CanadaTel. +1 450 465 0955E-mail [email protected]

Toronto SectionLee White26 Flaremore CrescentToronto, Ontario M2K 1V1CanadaTel. +1 416 222 2447Fax +1 416 222 8546E-mail [email protected]

CENTRAL REGION,USA/CANADA

Vice President:Jim KaiserMaster Mix1921 Division St.Nashville, TN 37203Tel. +1 615 321 5970Fax +1 615 321 0764E-mail [email protected]


INDIANA

Ball State University Section(Student)Michael Pounds, Faculty AdvisorAES Student SectionBall State UniversityMET Studios2520 W. BethelMuncie, IN 47306Tel. +1 765 285 5537Fax +1 765 285 8768E-mail [email protected]

Central Indiana SectionJames LattaSound Around6349 Warren Ln.Brownsburg, IN 46112Office Tel. +1 317 852 8379

Fax +1 317 858 8105E-mail [email protected]

ILLINOIS

Chicago SectionRobert ZurekMotorola2001 N. Division St.Harvard, IL 60033Tel. +1 815 884 1361Fax +1 815 884 2519E-mail [email protected]

Columbia College Section(Student)Dominique J. ChéenneFaculty AdvisorAES Student Section676 N. LaSalle, Ste. 300Chicago, IL 60610Tel. +1 312 344 7802Fax +1 312 482 9083

KANSAS

Kansas City SectionJim MitchellCustom Distribution Limited12301 Riggs Rd.Overland Park, KS 66209Tel. +1 913 661 0131Fax +1 913 663 5662

LOUISIANA

New Orleans SectionJoseph Doherty6015 Annunication St.New Orleans, LA 70118Tel. +1 504 891 4424Fax +1 504 891 6075

MICHIGAN

Detroit SectionEric BuschDLC Design47677 Avante Dr.Wixom, MI 48393Tel. +1 248 305 5534Fax +1 248 305 5536E-mail [email protected]

Michigan TechnologicalUniversity Section (Student)Andre LaRoucheAES Student SectionMichigan Technological

UniversityElectrical Engineering Dept.1400 Townsend Dr.Houghton, MI 49931Home Tel. +1 906 847 9324E-mail [email protected]

West Michigan SectionCarl HordykCalvin College3201 Burton S.E.Grand Rapids, MI 49546Tel. +1 616 957 6279Fax +1 616 957 6469E-mail [email protected]

MINNESOTA

Music Tech College Section(Student)

Michael McKernFaculty AdvisorAES Student SectionMusic Tech College19 Exchange Street EastSaint Paul, MN 55101Tel. +1 651 291 0177Fax +1 651 291 [email protected]

Ridgewater College,Hutchinson Campus Section(Student)Dave Igl, Faculty AdvisorAES Student SectionRidgewater College, Hutchinson

Campus2 Century Ave. S.E.Hutchinson, MN 55350E-mail [email protected]

Upper Midwest SectionGreg ReiersonRare Form Mastering4624 34th Avenue SouthMinneapolis, MN 55406Tel. +1 612 327 [email protected]

MISSOURI

St. Louis SectionJohn Nolan, Jr.693 Green Forest Dr.Fenton, MO 63026Tel./Fax +1 636 343 4765

NEBRASKA

Northeast Community CollegeSection (Student)Anthony D. BeardsleeFaculty AdvisorAES Student SectionNortheast Community CollegeP.O. Box 469Norfolk, NE 68702Tel. +1 402 644 0581Fax +1 402 644 0650E-mail [email protected]

OHIO

Ohio University Section(Student)Erin M. DawesAES Student SectionOhio UniversityRTVC Bldg.9 S. College St.Athens, OH 45701-2979Home Tel. +1 740 597 6608E-mail [email protected]

University of CincinnatiSection (Student)Thomas A. HainesFaculty AdvisorAES Student SectionUniversity of CincinnatiCollege-Conservatory of MusicM.L. 0003Cincinnati, OH 45221Tel. +1 513 556 9497Fax +1 513 556 0202

TENNESSEE

Belmont University Section(Student)Wesley Bulla, Faculty AdvisorAES Student SectionBelmont UniversityNashville, TN 37212

Middle Tennessee StateUniversity Section (Student)Doug Mitchell, Faculty AdvisorAES Student SectionMiddle Tennessee State University301 E. Main St., Box 21Murfreesboro, TN 37132Tel. +1 615 898 2553E-mail [email protected]

Nashville Section Tom EdwardsMTV Networks2806 Opryland Dr.Nashville, TN 37214Office Tel. +1 615 457 8009Fax +1 615 457 8855E-mail [email protected]

SAE Nashville Section (Student)Mark Martin, Faculty AdvisorAES Student Section7 Music Circle N.Nashville, TN 37203Tel. +1 615 244 5848Fax +1 615 244 3192E-mail [email protected]

TEXAS

Southwest Texas StateUniversity Section (Student)Mark C. EricksonFaculty AdvisorAES Student Section Southwest Texas State

University224 N. Guadalupe St.San Marcos, TX 78666Tel. +1 512 245 8451Fax +1 512 396 1169E-mail [email protected]

WESTERN REGION,USA/CANADA

Vice President:Bob MosesIsland Digital Media Group,

LLC26510 Vashon Highway S.W.Vashon, WA 98070Tel. +1 206 463 6667Fax +1 810 454 5349E-mail [email protected]


ARIZONA

Conservatory of TheRecording Arts and SciencesSection (Student)

SECTIONS CONTACTSDIRECTORY


Glen O’HaraFaculty AdvisorAES Student Section Conservatory of The Recording

Arts and Sciences2300 E. Broadway Rd.Tempe, AZ 85282Tel. +1 480 858 9400, 800 562

6383 (toll-free)Fax +1 480 829 [email protected]

CALIFORNIA

American River CollegeSection (Student)Eric Chun, Faculty AdvisorAES Student SectionAmerican River College Chapter4700 College Oak Dr.Sacramento, CA 95841Tel. +1 530 888 9440E-mail [email protected]

Cal Poly San Luis ObispoState University Section(Student)Jerome R. BreitenbachFaculty AdvisorAES Student SectionCalifornia Polytechnic State

UniversityDept. of Electrical EngineeringSan Luis Obispo, CA 93407Tel. +1 805 756 5710Fax +1 805 756 1458E-mail [email protected]

California State University–Chico Section (Student)Keith Seppanen, Faculty AdvisorAES Student SectionCalifornia State University–Chico400 W. 1st St.Chico, CA 95929-0805Tel. +1 530 898 5500E-mail [email protected]

Citrus College Section(Student)Gary Mraz, Faculty AdvisorAES Student SectionCitrus CollegeRecording Arts1000 W. Foothill Blvd.Glendora, CA 91741-1899Fax +1 626 852 8063

Cogswells PolytechnicalCollege Section (Student)Tim Duncan, Faculty SponsorAES Student SectionCogswell Polytechnical CollegeMusic Engineering Technology1175 Bordeaux Dr.Sunnyvale, CA 94089Tel. +1 408 541 0100, ext. 130Fax +1 408 747 0764E-mail [email protected]

Ex’pression Center for NewMedia Section (Student)Scott Theakston, Faculty AdvisorAES Student SectionEx’pression Center for New

Media6601 Shellmount St.

Emeryville, CA 94608Tel. +1 510 654 2934Fax +1 510 658 3414E-mail [email protected]

Long Beach City CollegeSection (Student)Nancy Allen, Faculty AdvisorAES Student SectionLong Beach City College4901 E. Carson St.Long Beach, CA 90808Tel. +1 562 938 4312Fax +1 562 938 4118E-mail [email protected]

Los Angeles SectionAndrew Turner400 South Main St., #403Los Angeles, CA 90013Tel. +1 213 625 1790E-mail [email protected]

San Diego SectionJ. Russell Lemon2031 Ladera Ct.Carlsbad, CA 92009-8521Home Tel. +1 760 753 2949E-mail [email protected]

San Diego State UniversitySection (Student)John Kennedy, Faculty AdvisorAES Student SectionSan Diego State UniversityElectrical & Computer

Engineering Dept.5500 Campanile Dr.San Diego, CA 92182-1309Tel. +1 619 594 1053Fax +1 619 594 2654E-mail [email protected]

San Francisco SectionBrian E. Cheney3429 Morningside Dr.El Sobrante, CA 94803Tel. +1 510 222 4276Fax +1 510 232 3837E-mail [email protected]

San Francisco StateUniversity Section (Student)John Barsotti, Faculty AdvisorAES Student SectionSan Francisco State UniversityBroadcast and Electronic

Communication Arts Dept.1600 Halloway Ave.San Francisco, CA 94132Tel. +1 415 338 1507E-mail [email protected]

Stanford University Section(Student)Jay Kadis, Faculty AdvisorStanford AES Student SectionStanford UniversityCCRMA/Dept. of MusicStanford, CA 94305-8180Tel. +1 650 723 4971Fax +1 650 723 8468E-mail [email protected]

University of SouthernCalifornia Section (Student)Richard McIlvery

Faculty AdvisorAES Student SectionUniversity of Southern California840 W. 34th St.Los Angeles, CA 90089-0851Tel. +1 213 740 3224Fax +1 213 740 3217E-mail [email protected]

COLORADO

Colorado SectionRobert F. MahoneyRobert F. Mahoney & Associates310 Balsam Ave.Boulder, CO 80304Tel. +1 303 443 2213Fax +1 303 443 6989E-mail [email protected]

Denver Section (Student)Roy Pritts, Faculty AdvisorAES Student SectionUniversity of Colorado at

DenverDept. of Professional StudiesCampus Box 162P.O. Box 173364Denver, CO 80217-3364Tel. +1 303 556 2795Fax +1 303 556 2335E-mail [email protected]

OREGON

Portland SectionTony Dal MolinAudio Precision, Inc.5750 S.W. Arctic Dr.Portland, OR 97005Tel. +1 503 627 0832Fax +1 503 641 8906E-mail [email protected]

UTAH

Brigham Young UniversitySection (Student)Jim Anglesey, Faculty AdvisorBYU-AES Student SectionSchool of MusicBrigham Young UniversityProvo, UT 84602Tel. +1 801 378 1299Fax +1 801 378 5973 (Music

Office)E-mail [email protected]

Utah SectionDeward Timothyc/o Poll Sound4026 S. MainSalt Lake City, UT 84107Tel. +1 801 261 2500Fax +1 801 262 7379

WASHINGTON

Pacific Northwest SectionGary LouieUniversity of Washington

School of Music4522 Meridian Ave. N., #201Seattle, WA 98103Office Tel. +1 206 543 1218Fax +1 206 685 9499E-mail [email protected]

The Art Institute of SeattleSection (Student)David G. ChristensenFaculty AdvisorAES Student SectionThe Art Institute of Seattle2323 Elliott Ave.Seattle, WA 98121-1622 Tel. +1 206 239 [email protected]

CANADA

Alberta SectionFrank LockwoodAES Alberta SectionSuite 404815 - 50 Avenue S.W.Calgary, Alberta T2S 1H8CanadaHome Tel. +1 403 703 5277Fax +1 403 762 6665E-mail [email protected]

Vancouver SectionPeter L. JanisC-Tec #114, 1585 BroadwayPort Coquitlam, B.C. V3C 2M7CanadaTel. +1 604 942 1001Fax +1 604 942 1010E-mail [email protected]

Vancouver Student SectionGregg Gorrie, Faculty AdvisorAES Greater Vancouver

Student SectionCentre for Digital Imaging and

Sound3264 Beta Ave.Burnaby, B.C. V5G 4K4, CanadaTel. +1 604 298 [email protected]

NORTHERN REGION,EUROPE

Vice President:Søren BechBang & Olufsen a/sCoreTechPeter Bangs Vej 15DK-7600 Struer, DenmarkTel. +45 96 84 49 62Fax +45 97 85 59 [email protected]

BELGIUM

Belgian SectionHermann A. O. WilmsAES Europe Region OfficeZevenbunderslaan 142, #9BE-1190 Vorst-Brussels, BelgiumTel. +32 2 345 7971Fax +32 2 345 3419

DENMARK

Danish Section




Knud Bank ChristensenSkovvej 2DK-8550 Ryomgård, DenmarkTel. +45 87 42 71 46Fax +45 87 42 70 10E-mail [email protected]

Danish Student SectionTorben Poulsen, Faculty AdvisorAES Student SectionTechnical University of DenmarkØrsted-DTU, Acoustic

TechnologyDTU - Building 352DK-2800 Kgs. Lyngby, DenmarkTel. +45 45 25 39 40Fax +45 45 88 05 77E-mail [email protected]

FINLAND

Finnish SectionKalle KoivuniemiNokia Research CenterP.O. Box 100FI-33721 Tampere, FinlandTel. +358 7180 35452Fax +358 7180 35897E-mail [email protected]

NETHERLANDS

Netherlands SectionRinus BooneVoorweg 105ANL-2715 NG ZoetermeerNetherlandsTel. +31 15 278 14 71, +31 62

127 36 51Fax +31 79 352 10 08E-mail [email protected]

Netherlands Student SectionDirk FischerAES Student SectionGroenewegje 143aDen Haag, NetherlandsHome Tel. +31 70 [email protected]

NORWAY

Norwegian SectionJan Erik JensenNøklesvingen 74NO-0689 Oslo, NorwayOffice Tel. +47 22 24 07 52Home Tel. +47 22 26 36 13 Fax +47 22 24 28 06E-mail [email protected]

RUSSIA

All-Russian State Institute ofCinematography Section(Student)Leonid Sheetov, Faculty SponsorAES Student SectionAll-Russian State Institute of

Cinematography (VGIK)W. Pieck St. 3RU-129226 Moscow, RussiaTel. +7 095 181 3868Fax +7 095 187 7174E-mail [email protected]

Moscow SectionMichael Lannie

Research Institute forTelevision and Radio

Acoustic Laboratory12-79 Chernomorsky bulvarRU-113452 Moscow, RussiaTel. +7 095 2502161, +7 095

1929011Fax +7 095 9430006E-mail [email protected]

St. Petersburg SectionIrina A. AldoshinaSt. Petersburg University of

TelecommunicationsGangutskaya St. 16, #31RU-191187 St. PetersburgRussiaTel. +7 812 272 4405Fax +7 812 316 1559E-mail [email protected]

St. Petersburg Student SectionNatalia V. TyurinaFaculty AdvisorProsvescheniya pr., 41, 185RU-194291 St. Petersburg, RussiaTel. +7 812 595 1730Fax +7 812 316 [email protected]

SWEDEN

Swedish SectionMikael OlssonAudio Data LabKatarinavägen 22SE-116 45 Stockholm, SwedenTel. +46 8 30 29 98Fax +46 8 641 67 91E-mail [email protected]

University of Luleå-PiteåSection (Student)Lars Hallberg, Faculty SponsorAES Student SectionUniversity of Luleå-PiteåSchool of MusicBox 744S-94134 Piteå, SwedenTel. +46 911 726 27Fax +46 911 727 10E-mail [email protected]

UNITED KINGDOM

British SectionHeather LaneAudio Engineering SocietyP.O. Box 645Slough GB-SL1 8BJUnited KingdomTel. +44 1628 663725Fax +44 1628 667002E-mail [email protected]

CENTRAL REGION,EUROPE

Vice President:Markus ErneScopein ResearchSonnmattweg 6CH-5000 Aarau, Switzerland

Tel. +41 62 825 09 19Fax +41 62 825 09 [email protected]

AUSTRIA

Austrian SectionFranz LechleitnerLainergasse 7-19/2/1AT-1238 Vienna, AustriaOffice Tel. +43 1 4277 29602Fax +43 1 4277 9296E-mail [email protected]

Graz Section (Student)Robert Höldrich, Faculty SponsorInstitut für Elektronische Musik

und AkustikInffeldgasse 10AT-8010 Graz, AustriaTel. +43 316 389 3172Fax +43 316 389 3171E-mail [email protected]

Vienna Section (Student)Jürg Jecklin, Faculty SponsorVienna Student SectionUniversität für Musik und

Darstellende Kunst WienInstitut für Elektroakustik und

Experimentelle MusikRienösslgasse 12AT-1040 Vienna, AustriaTel. +43 1 587 34 78Fax +43 1 587 34 78 20E-mail [email protected]

CZECH REPUBLIC

Czech SectionJiri OcenasekDejvicka 36CZ-160 00 Prague 6Czech Republic Home Tel. +420 2 24324556E-mail [email protected]

Czech Republic StudentSectionLibor Husník, Faculty AdvisorAES Student SectionCzech Technical University atPragueTechnická 2, CZ-116 27 Prague 6Czech RepublicTel. +420 2 2435 2115E-mail [email protected]

GERMANY

Berlin Section (Student)Bernhard Güttler Zionskirchstrasse 14DE-10119 Berlin, GermanyTel. +49 30 4404 72 19Fax +49 30 4405 39 03E-mail [email protected]

Central German SectionErnst-Joachim VölkerInstitut für Akustik und

BauphysikKiesweg 22-24DE-61440 Oberursel, GermanyTel. +49 6171 75031

Fax +49 6171 85483E-mail [email protected]

Darmstadt Section (Student)G. M. Sessler, Faculty SponsorAES Student SectionTechnical University of

DarmstadtInstitut für ÜbertragungstechnikMerkstr. 25DE-64283 Darmstadt, GermanyTel. +49 6151 [email protected]

Detmold Section (Student)Werner Czesla, Faculty AdvisorAES Student SectionHochschule für Musik DetmoldNeustadt 22DE-32756 Detmold, GermanyOffice Tel. +49 5231 9755Home Tel. +49 5232 990638Fax +49 5231 975972E-mail [email protected]

Düsseldolf Section (Student)Ludwig KuglerAES Student SectionBilker Allee 126DE-40217 Düsseldorf, GermanyTel. +49 211 3 36 80 [email protected]

Ilmenau Section (Student)Karlheinz BrandenburgFaculty SponsorAES Student SectionInstitut für MedientechnikPF 10 05 65DE-98684 Ilmenau, GermanyTel. +49 3677 69 2676Fax +49 3677 69 1255E-mail [email protected]

North German SectionReinhard O. SahrEickhopskamp 3DE-30938 Burgwedel, GermanyTel. +49 5139 4978Fax +49 5139 5977E-mail [email protected]

South German SectionGerhard E. PicklappLandshuter Allee 162DE-80637 Munich, GermanyTel. +49 89 15 16 17Fax +49 89 157 10 31E-mail [email protected]

HUNGARY

Hungarian SectionFerenc György TakácsSzellö u. 2. VII. 18.HU-1035 Budapest, HungaryHome Tel. +36 1 368 47 70Office Tel. +36 1 463 20 47Fax +36 1 463 32 66E-mail [email protected]

LITHUANIA

Lithuanian SectionVytautas J. StauskisVilnius Gediminas Technical


UniversitySauletekio al. 11LT-2040 Vilnius, LithuaniaTel. +370 2 700 492Fax +370 2 700 498E-mail [email protected]

POLAND

Polish SectionJan A. AdamczykUniversity of Mining and

MetallurgyDept. of Mechanics and

Vibroacousticsal. Mickiewicza 30PL-30 059 Cracow, PolandTel. +48 12 617 30 55Fax +48 12 633 23 14E-mail [email protected]

Technical University of GdanskSection (Student)Krzysztof KakolAES Student Section Technical University of GdanskSound Engineering Dept.ul. Narutowicza 11/12PL-809 52 Gdansk, PolandHome Tel. +48 501 058 279Fax +48 58 3471114E-mail [email protected]

Wroclaw University ofTechnology Section (Student)Andrzej B. DobruckiFaculty SponsorAES Student SectionInstitute of Telecommunications

and AcousticsWroclaw University of

TechnologyWybrzeze Wyspianskiego 27PL-503 70 Wroclaw, PolandTel. +48 71 320 30 68Fax +48 71 320 31 89E-mail [email protected]

REPUBLIC OF BELARUS

Belarus SectionValery ShalatoninBelarusian State University of

Informatics and Radioelectronics

vul. Petrusya Brouki 6BY-220027 MinskRepublic of BelarusTel. +375 17 239 80 95Fax +375 17 231 09 14E-mail [email protected]

SLOVAK REPUBLIC

Slovakian Republic SectionRichard VarkondaCentron Slovakia Ltd.Podhaj 107SK-841 03 BratislavaSlovak RepublicTel. +421 7 6478 0767Fax. +421 7 6478 0042E-mail [email protected]

SWITZERLAND

Swiss SectionAttila Karamustafaoglu

AES Swiss SectionSonnmattweg 6CH-5000 AarauSwitzerlandE-mail [email protected]

UKRAINE

Ukrainian SectionValentin AbakumovNational Technical University

of UkraineKiev Politechnical InstitutePolitechnical St. 16Kiev UA-56, UkraineTel./Fax +38 044 2746093

SOUTHERN REGION,EUROPE

Vice President:Daniel ZalayConservatoire de ParisDept. SonFR-75019 Paris, FranceOffice Tel. +33 1 40 40 46 14Fax +33 1 40 40 47 [email protected]

BOSNIA-HERZEGOVINA

Bosnia-Herzegovina SectionJozo TalajicBulevar Mese Selimovica 12BA-71000 SarajevoBosnia–HerzegovinaTel. +387 33 455 160Fax +387 33 455 163E-mail [email protected]

BULGARIA

Bulgarian SectionKonstantin D. KounovBulgarian National RadioTechnical Dept.4 Dragan Tzankov Blvd. BG-1040 Sofia, BulgariaTel. +359 2 65 93 37, +359 2

98 52 46 01Fax +359 2 963 1003E-mail [email protected]

CROATIA

Croatian SectionSilvije StamacHrvatski RadioPrisavlje 3HR-10000 Zagreb, CroatiaTel. +385 1 634 28 81Fax +385 1 611 58 29E-mail [email protected]

Croatian Student SectionHrvoje DomitrovicFaculty AdvisorAES Student SectionFaculty of Electrical

Engineering and ComputingDept. of Electroaocustics (X. Fl.)Unska 3HR-10000 Zagreb, Croatia

Tel. +385 1 6129 640Fax +385 1 6129 852E-mail [email protected]

FRANCE

Conservatoire de ParisSection (Student)Alessandra Galleron36, Ave. ParmentierFR-75011 Paris, FranceTel. +33 1 43 38 15 94

French SectionMichael WilliamsIle du Moulin62 bis Quai de l’Artois FR-94170 Le Perreux sur

Marne, FranceTel. +33 1 48 81 46 32Fax +33 1 47 06 06 48E-mail [email protected]

Louis Lumière Section(Student)Alexandra Carr-BrownAES Student SectionEcole Nationale Supérieure

Louis Lumière7, allée du Promontoire, BP 22FR-93161 Noisy Le Grand

Cedex, FranceTel. +33 6 18 57 84 41E-mail [email protected]

Strasbourg Section (Student)Alban Moraud section etudiante AES France

Strasbourg8, quai du Chanoine WintererFR-67000 Strasbourg, FranceHome Tel. +31 6 13 08 38 30E-mail [email protected]

GREECE

Greek SectionSoterios SalamourisRoister Sapfous St. 145 GR-17675 Kallithea, GreeceTel. +30 1 9599088, +30 1

9522283Fax +30 1 9582730E-mail [email protected]

ISRAEL

Israel SectionBen Bernfeld Jr.H. M. Acustica Ltd.1/11 Ha’alumim St.IL-46308 Herzlia, IsraelTel. +972 9 9574448Fax +972 9 9574254E-mail [email protected]

ITALY

Italian SectionCarlo Perrettac/o AES Italian SectionPiazza Cantore 10IT-20134 Milan, ItalyTel. +39 338 9108768

Fax +39 02 58440640E-mail [email protected]

Italian Student SectionFranco Grossi, Faculty AdvisorAES Student SectionViale San Daniele 29 IT-33100 Udine, ItalyTel. +39 [email protected]

PORTUGAL

Portugal SectionRui Miguel Avelans CoelhoR. Paulo Renato 1, 2APT-2745-147 Linda-a-VelhaPortugalTel. +351 214145827E-mail [email protected]

ROMANIA

Romanian SectionMarcia TaiachinRadio Romania60-62 Grl. Berthelot St.RO-79756 Bucharest, RomaniaTel. +40 1 303 12 07Fax +40 1 222 69 19

SLOVENIA

Slovenian SectionTone SeliskarRTV SlovenijaKolodvorska 2SI-1550 Ljubljana, SloveniaTel. +386 61 175 2708Fax +386 61 175 2710E-mail [email protected]

SPAIN

Spanish SectionJuan Recio MorillasSpanish SectionC/Florencia 14 3oDES-28850 Torrejon de Ardoz

(Madrid), SpainTel. +34 91 540 14 03E-mail [email protected]

TURKEY

Turkish SectionSorgun AkkorSTDSelamicesme, Gulden sok. 2/2Kadikoy TR-81060, IstanbulTurkeyTel. +90 216 4671814Fax +90 216 4671815E-mail [email protected]

YUGOSLAVIA

Yugoslavian Section Tomislav StanojevicSava centreM. Popovica 9YU-11070 Belgrade, YugoslaviaTel. +381 11 311 1368Fax +381 11 605 [email protected]




LATIN AMERICAN REGION

Vice President:Mercedes OnoratoTalcahuano 141Buenos Aires, ArgentinaTel./Fax +5411 4 375 [email protected]

ARGENTINA

Argentina SectionMercedes Onorato Talcahuano 141Buenos Aires, ArgentinaTel./Fax +5411 4 375 0116E-mail [email protected]

BRAZIL

Brazil SectionRosalfonso BortoniAve. João de Camargo, 510BR-37540-000 Santa Rita do

Sapucaí, MG, BrazilTel. +55 35 3471 9230Fax +55 35 3471 9328E-mail [email protected]

CHILE

Chile SectionAlejandro Soto de ValleUniversidad Tecnológica

Vicente Pérez RosalesBrown Norte 290Nunoa, Santiago de ChileTel. +56 2 274 5432Fax +56 2 223 8825

COLOMBIA

Colombia SectionTony Penarredonda CaraballoCarrera 51 #13-223Medellin, ColombiaTel. +57 4 265 7000Fax +57 4 265 2772E-mail [email protected]

University of Javeriana, BogotaSection (Student)Ronald Beals, Faculty AdvisorAES Student SectionUniversidad JaverianaDept. de MusicaCarrera 7 #40-62, Piso 2Bogota, ColombiaTel./Fax +57 1 338 4548E-mail [email protected]

MEXICO

Mexican SectionJavier Posada Div. Del Norte #1008Col. Del ValleMexico, D.F. MX-03100MexicoTel. +52 5 669 48 79Fax +52 5 543 60 37E-mail [email protected]

URUGUAY

Uruguay SectionRafael AbalSondor S.A.Calle Rio Branco 1530C.P. UY-11100 MontevideoUruguayTel. +598 2 91 26 70, +598 2 92

53 88Fax +598 2 92 52 72E-mail [email protected]

VENEZUELA

Taller de Arte Sonoro,Caracas Section (Student)Carmen Bell-Smythe de LealFaculty AdvisorAES Student SectionTaller de Arte SonoroAve. Rio de Janeiro Qta. Tres PinosChuao, VE-1061 CaracasVenezuelaTel. +58 14 9292552Tel./Fax +58 2 9937296E-mail [email protected]

Venezuela SectionElmar LealAve. Rio de JaneiroQta. Tres PinosChuao, VE-1061 CaracasVenezuelaTel. +58 14 9292552Tel./Fax +58 2 9937296E-mail [email protected]

INTERNATIONAL REGION

Vice President:Neville Thiele10 Wycombe St.Epping, NSW AU-2121,AustraliaTel. +61 2 9876 2407Fax +61 2 9876 2749E-mail [email protected]

AUSTRALIA

Adelaide SectionDavid MurphyKrix Loudspeakers14 Chapman Rd.Hackham AU-5163South AustraliaTel. +618 8 8384 3433Fax +618 8 8384 3419E-mail [email protected]

Brisbane SectionDavid RingroseAES Brisbane SectionP.O. Box 642Roma St. Post OfficeBrisbane, Qld. AU-4003, AustraliaOffice Tel. +61 7 3364 6510E-mail [email protected]

Melbourne SectionGraham J. Haynes

P.O. Box 5266Wantirna South, VictoriaAU-3152, AustraliaTel. +61 3 9887 3765Fax +61 3 9887 [email protected]

Sydney SectionHoward JonesAES Sydney SectionP.O. Box 766Crows Nest, NSW AU-2065AustraliaTel. +61 2 9417 3200Fax +61 2 9417 3714

HONG KONG

Hong Kong SectionHenry Ma Chi FaiHKAPA, School of Film and

Television1 Gloucester Rd. Wanchai, Hong KongTel. +852 2584 8824Fax +852 2588 1303E-mail [email protected]

INDIA

India SectionAvinash OakWestern Outdoor Media Tech.16, Mumbai Samachar MargMumbai IN-400 023, IndiaTel. +91 22 2046181Fax +91 22 2043038E-mail [email protected]

JAPAN

Japan SectionKatsuya (Vic) Goh2-15-4 Tenjin-cho, Fujisawa-shiKanagawa-ken 252-0814, JapanHome Tel. +81 466 81 0681 Fax +81 466 81 0698E-mail [email protected]

KOREA

Korea SectionSeong-Hoon KangTaejeon Health Science CollegeDept. of Broadcasting

Technology77-3 Gayang-dong Dong-guTaejeon, Korea Tel. +82 42 630 5990Fax +82 42 628 1423E-mail [email protected]

MALAYSIA

Malaysia SectionC. K. Ng King Musical Industries

Sdn BhdLot 5, Jalan 13/2MY-46200 Kuala LumpurMalaysiaTel. +603 7956 1668Fax +603 7955 4926E-mail [email protected]

PHILIPPINES

Philippines SectionDario (Dar) J. Quintos125 Regalia Park TowerP. Tuazon Blvd., CubaoQuezon City, PhilippinesTel./Fax +63 2 4211790, +63 2

4211784E-mail [email protected]

SINGAPORE

Singapore SectionP. V. Anthonyc/o MIND & MEDIA1G Paya Lebar Rd. SG-408999 SingaporeRepublic of SingaporeTel. +65 0 547 1067Fax +65 0 743 0096E-mail [email protected]

Chair:Scott CannonStanford University Section (AES)P.O. Box 15259Stanford, CA 94309Tel. +1 650 346 4556Fax +1 650 723 8468E-mail [email protected]

Vice Chair:Dell HarrisHampton University Section(AES)125A Mariners CoveHampton, VA 23669Tel +1 757 723 4374E-mail [email protected]

Chair:Blaise ChabanisConservatoire de Paris

(CNSMDP) Student Section (AES)

14, rue de la FaisanderieFR-77200 Torcy, FranceTel. +336 62 15 29 97E-mail [email protected]

Vice Chair:Werner de BruijnThe Netherlands Student

Section (AES)Korvezeestraat 541NL-2628 CZ DelftThe NetherlandsHome Tel. +31 15 2622995Office Tel. +31 15 2782021E-mail [email protected]

EUROPE/INTERNATIONALREGIONS

NORTH/SOUTH AMERICA REGIONS

STUDENT DELEGATEASSEMBLY


AES CONVENTIONS AND CON


The latest details on the following events are posted on the AES Website: http://www.aes.org

2002

Los Angeles

Convention chair:Floyd TooleHarman International8500 Balboa Blvd.Northridge, CA 91329, USATelephone: +1 818 895 5761Fax: +1 818 893 7139Email: [email protected]

Papers cochair: John StrawnS Systems, Inc.15 Willow AvenueLarkspur, CA 94939, USATelephone: +1 415 927 8856Email: [email protected]

Papers cochair:Eric BenjaminDolby Laboratories, Inc.100 Potrero AvenueSan Francisco, CA 94103-0200 USA

113th ConventionLos Angeles, California, USADate: 2002 October 5–8Location: Los AngelesConvention Center,Los Angeles, California, USA

Espoo

Conference cochairs:Jyri Huopaniemi and Nick ZacharovNokia Research CenterSpeech and Audio SystemsLaboratoryEmail: [email protected]

Papers chair: Vesa VälimäkiHelsinki University of TechnologyLab. of Acoustics and Audio SignalProcessingP. O. Box 3000, FIN-02015 HUTEspoo, FinlandFax: +358 9 460 224Email: [email protected]

22nd International ConferenceEspoo, Finland“Virtual, Synthetic, andEntertainment Audio”Date: 2002 June 15–17Location: Helsinki Universityof TechnologyEspoo, Finland

2002St. Petersburg,

Russia

Telephone: +7 812 2724405Fax: +7 812 3161559Email: [email protected]

Papers chair: Natalia V. TyurinaBaltic State Technical University1st Krasnoarmeyskaya Str. 1RU-198005 St. Petersburg, RussiaTelephone/Fax: +7 812 5951730Email: [email protected]

Conference chair:Nickolay I. IvanovBaltic State Technical UniversityTelephone: +7 812 1101573Fax: +7 812 3161559Email: [email protected]

Scientific Committee chair:Irina A. AldoshinaSt. Petersburg University ofTelecommunications

21st International ConferenceSt. Petersburg, Russia“Architectural Acoustics andSound Reinforcement”Date: 2002 June 1–3Location: Hotel MoscowSt. Petersburg, Russia

Munich2002

NEW DATESConvention Chair:Martin WöhrBayerischer RundfunkStudioproduktion HFDE-80300 Munich, GermanyTelephone: +49 89 59002434Fax: +49 89 59003393Email: [email protected]

Papers chair:Ben BernfeldKrozinger Str. 22DE-79219 Staufen, GermanyTelephone: +49 (0)7633 982570Fax: +49 (0)40 3603046919Email: [email protected]

112th ConventionMunich, GermanyDate: 2002 May 10–13Location: MOC Center,Munich, Germany

FERENCESPresentationManuscripts submitted should betypewritten on one side of ISO size A4(210 x 297 mm) or 216-mm x 280-mm(8.5-inch x 11-inch) paper with 40-mm(1.5-inch) margins. All copies includingabstract, text, references, figure captions,and tables should be double-spaced.Pages should be numbered consecutively.Authors should submit an original plustwo copies of text and illustrations.ReviewManuscripts are reviewed anonymouslyby members of the review board. After thereviewers’ analysis and recommendationto the editors, the author is advised ofeither acceptance or rejection. On thebasis of the reviewers’ comments, theeditor may request that the author makecertain revisions which will allow thepaper to be accepted for publication.ContentTechnical articles should be informativeand well organized. They should citeoriginal work or review previous work,giving proper credit. Results of actualexperiments or research should beincluded. The Journal cannot acceptunsubstantiated or commercial statements.OrganizationAn informative and self-containedabstract of about 60 words must beprovided. The manuscript should developthe main point, beginning with anintroduction and ending with a summaryor conclusion. Illustrations must haveinformative captions and must be referredto in the text.

References should be cited numerically inbrackets in order of appearance in thetext. Footnotes should be avoided, whenpossible, by making parentheticalremarks in the text.

Mathematical symbols, abbreviations,acronyms, etc., which may not be familiarto readers must be spelled out or definedthe first time they are cited in the text.

Subheads are appropriate and should beinserted where necessary. Paragraphdivision numbers should be of the form 0(only for introduction), 1, 1.1, 1.1.1, 2, 2.1,2.1.1, etc.

References should be typed on amanuscript page at the end of the text inorder of appearance. References toperiodicals should include the authors’names, title of article, periodical title,volume, page numbers, year and monthof publication. Book references shouldcontain the names of the authors, title ofbook, edition (if other than first), nameand location of publisher, publication year,and page numbers. References to AESconvention preprints should be replacedwith Journal publication citations if thepreprint has been published.IllustrationsFigure captions should be typed on aseparate sheet following the references.Captions should be concise. All figures

should be labeled with author’s name andfigure number.Photographs should be black and white prints without a halftone screen,preferably 200 mm x 250 mm (8 inch by10 inch).Line drawings (graphs or sketches) can beoriginal drawings on white paper, or high-quality photographic reproductions.The size of illustrations when printed in theJournal is usually 82 mm (3.25 inches)wide, although 170 mm (6.75 inches) widecan be used if required. Letters on originalillustrations (before reduction) must be largeenough so that the smallest letters are atleast 1.5 mm (1/16 inch) high when theillustrations are reduced to one of the abovewidths. If possible, letters on all originalillustrations should be the same size.Units and SymbolsMetric units according to the System ofInternational Units (SI) should be used.For more details, see G. F. Montgomery,“Metric Review,” JAES, Vol. 32, No. 11,pp. 890–893 (1984 Nov.) and J. G.McKnight, “Quantities, Units, LetterSymbols, and Abbreviations,” JAES, Vol.24, No. 1, pp. 40, 42, 44 (1976 Jan./Feb.).Following are some frequently used SIunits and their symbols, some non-SI unitsthat may be used with SI units (), andsome non-SI units that are deprecated ( ).

Unit Name Unit Symbolampere Abit or bits spell outbytes spell outdecibel dBdegree (plane angle) () °farad Fgauss ( ) Gsgram ghenry Hhertz Hzhour () hinch ( ) injoule Jkelvin Kkilohertz kHzkilohm kΩliter () l, Lmegahertz MHzmeter mmicrofarad µFmicrometer µmmicrosecond µsmilliampere mAmillihenry mHmillimeter mmmillivolt mVminute (time) () minminute (plane angle) () ’nanosecond nsoersted ( ) Oeohm Ωpascal Papicofarad pFsecond (time) ssecond (plane angle) () ”siemens Stesla Tvolt Vwatt Wweber Wb

INFORMATION FOR AUTHORS

Telephone: +1 415 558 0236Email: [email protected]

Exhibit information:Chris PlunkettTelephone: +1 212 661 8528Fax: +1 212 682 0477Email: [email protected]

Call for papers: Vol. 50, No. 1/2,p. 100 (2002 January/February)

Call for workshops participants: Vol. 50, No. 3, p. 206 (2002 March)

Call for papers: Vol. 49, No. 9,p. 852 (2001 September)

Conference preview: This issue,pp. 290–301 (2002 April)

Call for papers: Vol. 49, No. 9,p. 851 (2001 September)

Conference preview: This issue,pp. 274–289 (2002 April)

Exhibit information:Thierry BergmansTelephone: +32 2 345 7971Fax: +32 2 345 3419Email: [email protected]

Call for papers: Vol. 49, No. 7/8,p. 720 (2001 July/August)

Convention preview: Vol. 50, No. 3,pp. 178–200 (2002 March)

sustainingmemberorganizations AESAES

VO

LU

ME

50,NO

.4JO

UR

NA

L O

F T

HE

AU

DIO

EN

GIN

EE

RIN

G S

OC

IET

Y2002 A

PR

IL

JOURNAL OF THE AUDIO ENGINEERING SOCIETYAUDIO / ACOUSTICS / APPLICATIONSVolume 50 Number 4 2002 April

In this issue…

Synthesis As Music Restoration

TelecommunicationsReproduction Quality

Impulse Response Methods

Virtual Auditory-SpaceHeadphone Issues

Features…

21st ConferenceSt. Petersburg, Russia—Preview

22nd ConferenceEspoo, Finland—Preview

Update: Sections Directory

The Audio Engineering Society recognizes with gratitude the financialsupport given by its sustaining members, which enables the work ofthe Society to be extended. Addresses and brief descriptions of thebusiness activities of the sustaining members appear in the Octoberissue of the Journal.

The Society invites applications for sustaining membership. Informa-tion may be obtained from the Chair, Sustaining Memberships Com-mittee, Audio Engineering Society, 60 East 42nd St., Room 2520,New York, New York 10165-2520, USA, tel: 212-661-8528. Fax: 212-682-0477.

ACO Pacific, Inc.Air Studios Ltd.AKG Acoustics GmbHAKM Semiconductor, Inc.Amber Technology LimitedAMS Neve plcATC Loudspeaker Technology Ltd.Audio LimitedAudiomatica S.r.l.Audio Media/IMAS Publishing Ltd.Audio Precision, Inc.AudioScience, Inc.Audio-Technica U.S., Inc.AudioTrack CorporationAudio-Visual Land Pte Ltd.Autograph Sound Ltd.B & W Loudspeakers LimitedBMP RecordingBritish Broadcasting CorporationBSS Audio Cadac Electronics PLCCalrec AudioCanford Audio plcCEDAR Audio Ltd.Celestion International LimitedCentre for Signal ProcessingCerwin-Vega, IncorporatedCommunity Professional Loudspeakers, Inc.Cox Audio EngineeringCrystal Audio Products/Cirrus Logic Inc.D.A.S. Audio, S.A.D.A.T. Ltd.dCS Ltd.Deltron Emcon LimitedDigigramDigital Audio Disc CorporationDolby Laboratories, Inc.DRA LaboratoriesDTS, Inc.DYNACORD, EVI Audio GmbHEastern Acoustic Works, Inc.

Event Electronics, LLCFerrotec (USA) CorporationFocusrite Audio Engineering Ltd.Fostex America, a division of Foster Electric

U.S.A., Inc.FreeSystems Private LimitedFTG Sandar TeleCast ASGentner Communications Corp.Harman BeckerHHB Communications Ltd.Innova SONInnovative Electronic Designs (IED), Inc.International Federation of the Phonographic

IndustryJBL ProfessionalJensen Transformers Inc.Kawamura Electrical LaboratoryKEF Audio (UK) LimitedKenwood U.S.A. CorporationKlark Teknik Group (UK) PlcKlipsch L.L.C.Lectret Precision Pte. Ltd.Leitch Technology CorporationLindos ElectronicsMagnetic Reference Laboratory (MRL) Inc.Martin Audio Ltd.Meridian Audio LimitedMetropolis Studios and MasteringMiddle Atlantic Products Inc.Mosses & MitchellM2 Gauss Corp.National Semiconductor CorporationGeorg Neumann GmbH Neutrik AGNVisionNXT (New Transducers Ltd.)1 LimitedOntario Institute of Audio Recording TechnologyOutline sncPRIMEDIA Business Magazines & Media Inc.Prism Sound

Pro-Bel LimitedPro-Sound NewsRadio Free AsiaRane CorporationRecording ConnectionRocket NetworkRoyal National Institute for the BlindRycote Microphone Windshields Ltd.SADiESanctuary Studios Ltd.Sekaku Electron Ind. Co., Ltd.Sennheiser Electronic CorporationShure Inc.Snell & Wilcox Ltd.Solid State Logic, Ltd.Sony Broadcast & Professional EuropeSound Devices LLCSound On Sound Ltd.Soundcraft Electronics Ltd.Soundtracs plcSowter Audio TransformersSRS Labs, Inc.Stage AccompanySterling Sound, Inc.Studer North America Inc.Studer Professional Audio AGTannoy LimitedTASCAMTHAT CorporationTOA Electronics, Inc.TommexTouchtunes Music Corp.United Entertainment Media, Inc.Uniton AGUniversity of Essex, Dept. of Electronic

Systems EngineeringUniversity of SalfordUniversity of Surrey, Dept. of Sound RecordingWenger CorporationJ. M. Woodgate and AssociatesYamaha Research and Development