Human Factors Course Lec14 Sound Exposure

8/3/2019 Human Factors Course Lec14 Sound Exposure

1/43

MeasuringLoudness


2/43

Loudnessisacomplex

subjectiveexperience

relatedtoboththe

intensityandthefrequencyofthesound.

Muchresearchhas

beenperformedover

theyearstodevelop

loudnessindices,two

earlyattemptsbeing

thephon andthesone.

Loudness


3/43

Thephon wasdevelopedbyexperiments

whichusedpuretonesoundsignalsof

fixedfrequencyandamplitude.

Ineachtesttheparticipantpresenteda1000Hzpuretonesoundasareference,

thenthesoundfrequencywaschanged

andtheparticipantwasaskedtoadjust

theamplitudeofthenewsignaluntilit

wasofequalloudness.

Byperformingthetestmanytimeswith

differentfrequenciesanddifferentpeople

itwaspossibletogenerateasetofequal-

loudnesscurves.

Loudness


4/43

Fromtheequalloudnesscurvesitcanbe

seenthathumanperceptionofloudness

variesasafunctionoffrequency.Humansareparticularlysensitivetofrequenciesin

therangefrom1000to6000Hz.

Loudness


5/43

Thephon wasdesignatedtheunitofloudness

andwassetequaltothedecibellevelofthe

1000Hzreferencetone. Forexample,alltones

judgedtobeofequalloudnesstothe60dBreferencetonearedesignatedashavinga

loudnessof60phons.

Loudness


6/43

Thefrequencyweightingnetworksusedin

soundlevelmetersarebasedonthephon

curvesdevelopedbyFletcherandMunson.

TheAandBfrequencyweightingsarethe

40and70phoncontours,butwithsomeminormodificationstosimplytherequired

electricalfilternetwork.

Loudness


7/43

Phoncurvesprovideinformationaboutthe

equivalenceofsounds,butnotabouttheabsolutelevelofperceivedloudness.We

cannotsay,forexample,howmanytimes

loudera40phonsoundiswithrespecttoa

20phonsound.

FletcherandMunsonthereforeperformed

furthertestswitharatingscalewhichwas

laternamedthesone.Onesoneisdefined

astheloudnessofa1000Hztoneof40dB

(40phons).

Asoundwhichisjudgedtobetwiceasloud

asthe1000Hzstandardreferencetonehasa

loudnessvalueof2sones,asoundjudgedthreetimesasloudis3sones,etc..

Loudness


8/43

Thegraphpresentstherelationshipbetweenthelevel

inphonsandtheperceivedloudnessin sones forpure

tonesounds.Theperceivedloudnessgrowsrapidly

withincreasingsoundpressure,particularlyatlower

levels.

Loudness


9/43

Thegraphabovegivesanapproximate

indicationofthesonevaluesofsometypicalsoundsfromeverydaylife.

Loudness


10/43

Theloudnessofbroadbandsounds

canbeestimatedbymeansofthe

StevensLoudnessMethod(ISO532A).

Inthismethodthesoundenergyis

firstdividedintooctaveor1/3octave

bands.Aloudnessvalueforeachbandisthendeterminedbymeansof

aloudnessnomogram.

Thetotalloudnessisthendetermined

fromtheindividualbandvaluesbymeansofasummationformula.The

formulatakesacousticmaskinginto

accountbyweightingtheloudnessof

thebandwiththegreatestvalueaboutthreetimesasmuchastheother

bands.

StevensLoudness


11/43

StevensLoudness

TheloudnessnomogramusedintheStevensmethod

providesaloudnessindexwhichistherelativeloudnessofagivenoctaveorthirdoctaverandomnoisesoundto

thatofareferenceoctaveorthirdoctaverandomnoise

bandcentredon1000Hz.


12/43

StevensLoudness

Thesummationformulaforobtainingthetotalsone loudnessvaluefromthesone valuesof

theindividualbandsis


13/43

Zwicker Loudness

Anothermethodforestimatingtotal

loudnessistheZwickermethod.

LiketheStevensmethod,theZwicker

methodisbasedontheuseofoctave

or1/3octavebandanalysisofthesoundsignal.


14/43

Zwicker LoudnessTheZwicker loudnessprocedureismorecomplex

thantheStevenloudnessprocedurebecause

maskingeffectsareevaluatedateachstage.

Maskingoccurswhenasoundisnothearddueto

thepresenceofanintensesoundatanearby

frequency.Forexample,a90dBtoneat1200Hz

willcompletelymaska50dBtoneat4000Hz.


15/43

Onetermthatissometimesusedtodescribe

theeffectsofunwantedsoundisannoyance.

Annoyanceisasubjectivequantityassociated

withtheinappropriatenessorunwantedness

ofthesound.

Itisimportanttonotethattheloudnessvalue

ofagivensoundisonlyweaklycorrelatedwith

itsannoyance.

Annoyance


16/43

Annoyancecanbecausedbysoundsthatare

toointensewithrespecttotheenvironmentin

whichtheyoccur.

Evenlowintensitysoundscan,however,

causeannoyancewhentheyareunexpected

orunusual. Forexample,eveniflowin

amplitude,apuretonesoundfromafanorair

conditionercanproducegreatannoyanceifit

isnotexpectedornotwantedinaspecific

room.

Annoyance


17/43

AuditoryEnvironment


18/43

Whendesigningforpeopletheacoustic

environmentshould:

AuditoryEnvironment

avoidhearingloss

minimisesoundsrelatedto

annoyanceandstress

minimisethedisruptionof

speechcommunications

transmitdesiredsoundsreliably

andpleasantlytothelisteners


19/43

TheA-weightedSoundPressureLevelLAisdefinedas

WherepA(t)istheinstantaneoussound

pressuremeasuredusingthestandardAscalefrequencyweightingshownbelow.

SoundLevel

dBp

tpLogL

referecne

AA

2

10

)(10

=


20/43

TheAverageSoundLevelLav,T isdefinedas

whereTisthetimeoverwhichthemeasurementis

averaged.TheAverageA-weightedSoundLevelLA,Tisdefinedas

SoundLevel

dBp

dttpT

LogLreference

T

Tav

=

2

0

2

10,

)(1

10

dBp

dttpT

LogLreference

T

A

TA

=

2

0

2

10,

)(1

10


21/43

TheDay-NightSoundLevelLdn isdefinedas

wherethefirsttermcoversthedaytimehoursfrom

7:00to22:00andthesecondtermcoversthenight

timehoursfrom22:00to7:00.Thenightlevelsare

takentobe10dBmorethantheyactuallymeasure.

SoundLevel

dBp

dttp

p

dttp

LogLreference

A

reference

A

dn

+=

2

00:7

00:22

2

2

00:22

00:7

2

10

)(10)(

24

110


22/43

TheA-weightedSoundExposureEAT isdefinedas

Whilethe A-weightedNoiseExposureLevelLEA,T is

definedas

WhereE0 isareferencevaluenormallytakentobe

(20Pa)2swhichis(4x10-10Pa)2s

SoundLevel

[ ]sPadttpET

AAT = 2

0

2)(

dBEELogL TATEA

=

0

,

10,10


23/43

EffectsofNoiseonHumanPerformance

Alittlenoiseintheworkenvironmentis

beneficialasitservestoincreasearousal,whichcanproduceimprovedtask

performance.

Beyondacertainlevelofintensitythetask

performancebeginstodegrade.

Sudden,unexpected,noisecanproducea

startleresponsewhichinterrupts

concentrationandtaskperformance.

Continuousnoisenormallyreduces

performanceoncomplextaskssuchas

visualtracking.Thereductionis

proportionaltothenoiselevel.

Psychologicaleffectsofnoisemayinclude

anxiety,helplessness,narrowedattentionandotheradverseeffectsthatdegradetask

performance.


24/43

EffectsofNoise

onHumanPerformance


25/43

CriteriaforSpeechCommunication

Acousticdesignoftenseekstoachieve

environmentalnoiselevelsthatarelowenoughtopermitacceptablespeech

communicationbetweenindividuals.

Thedesignmethodsusedforspeechcanalsobeappliedtomusicsincethe

frequenciesinvolvedaresimilar.For

example,thetonalrangeoftheviolinis

from200to8000Hz.


26/43

ArticulationIndex

Speechintelligibilityrequiresthatthe

listenerreceivesoundthatisabove

thethresholdofhearingandbelowthe

regionofhearingoverload.

Speechintelligibilityalsorequiresthat

thespeechsignalbestrongerthanthe

backgroundnoisepresentinthe

environment.


27/43

ArticulationIndex

Thehearingthreshold,theoverloadregionandthe

typicalspeechregionforamaleraisedvoiceat1

metredistancearepresentedbelowasafunction

offrequency.


28/43

ArticulationIndex

Anarticulationindexof100percentcorrespondstoa

situationwherethespectrumlevelsofspeechatthe

listenersearlieabovethethresholdofhearingand

belowtheoverloadline.Thespeechspectrumlevels

mustalsobeabovethespectrumofthebackground

noise.

Ifontheotherhandthenoisespectrumcoverspartof

theshadedspeechregion,orifpartofthespeech

regionfallsbelowthethresholdcurve,thearticulation

indexislessthan100percent.


29/43

ArticulationIndexForanarticulationindexof0.6ormoreconversationwill

besatisfactory,whileforanAIof0.3orlessthespeech

communicationwillbeunsatisfactory.

Thereare,however,otherfactorsinfluencingword

intelligibilitythereforetheAIcannotbeconsideredthe

absolutemeasureoftheacousticenvironment.

AnexampleoftherelationshipbetweenAIandspeech

intelligibilityisprovidedbythegraphbelowdeveloped

byNASA.


30/43

SpeechInterferenceLevel

TheSpeechInterferenceLevel(SIL)isasimple

meansofquantifyingtheeffectofbackground

noisemeasuredatthelocationofthelistener.

TheAmericanNationalStandarddefinitionof

theSILis

SpeechInterferenceLevel(SIL)in

decibelsisthearithmeticaverage

ofthesoundpressurelevelsofthe

interferingnoisere20Painthe

fouroctavebandscentredonthe

frequencies500,1000,2000and

4000Hz.


31/43


ANASAdocumentprovidesatablefor

interpretingSpeechInterferenceLevels.


32/43


ANASAdocumentprovidesagraphforinterpretingSpeechInterferenceLevels.


33/43

SpeechCommunication

CriterionCurvesCriterioncurvesweredevelopedusingthespeechinterference

level(SIL).Thecurvesassumethatthetalkerandlistenerare

locatedinafreefield,andgivethepermissibleSILatthe

listenerspositionasafunctionofthetalker-listenerseparation

randthetalkersvoicelevel.

ThecurvesweredevelopedforanAIof0.5whichcorresponds

toamonosyllabicwordintelligibilityofatleast85%,andwere

developedseparatelyformenandwomensincewomens

voicesareabout4dBweaker.


34/43

Non-Verbal

AuditoryWarnings


35/43

Non-VerbalAuditoryWarnings

Non-VerbalAuditoryWarningreferstothe

transmissionofinformationbymeansof

bells,buzzersandothersimplesounds.

Examplesincludethebuzzersusedon

medicalmonitorsandthevariouswarning

soundsusedtosignalemergenciesin

buildingsandinvehicles.


36/43


Althoughvisualdisplaysareabletotransmit

moreinformation,therearemanysituationsin

whichanauditorywarningispreferable. For

examplewhen

Theoperatorismobileandhencenotable

toseeavisualwarninggivenbymeansof

symbolsorphrases.

Theoperatorisinahighmentalworkloadenvironmentinwhichvisualsignalsare

likelytobemissed.

Aneconomicalmeansofconveyingsimple

informationtoalargenumberofpeopleissought.


37/43


Therearemanysituationsinwhichverbal

auditorywarningsarenoteffective. For

example:

Inahospitaltheuseofverbalwarnings

whichcanbeoverheardbythepatient

maybeinsensitive.

Inmanyemergencysituationsthereis

notenoughtimetoconveytheproblem

throughspeechsignals.

Inmanyworkenvironmentsthereare

highlevelsofbackgroundnoisewhich

makespeechsignalsdifficulttointerpret

correctly.


38/43

Non-Verbal

AuditoryWarnings

Therearetwomainrequirementsforan

effectivenon-verbalauditorywarning

system:

Itneedstobeheard,butshouldnot

betooloud.

Itneedstobepsychologically

appropriate.


39/43

Non-Verbal

AuditoryWarningsSettingthesoundlevelofauditorywarningsrequires

knowledgeofthebackgroundnoiselevelsforthe

environment. Oncethebackgroundlevelsareknown

thewarningscanbesettolevelssuchas15-25dB

abovethebackgroundwhicharehighenoughtonotbemasked.

Asanexamplethegraphbelowpresentsdataforthe

flightdeckofaBoeing727aircraft.Thelowerlinesare

measuredbackgroundnoisespectrainvariousflight

conditions.Therangeofappropriatelevelsforauditory

warningsisgivenbytheshadedregionatthetop.


40/43

Non-Verbal

AuditoryWarnings

15-25dBabovethebackgroundlevelisatypicaldesign

guidelineforauditorywarning.Severalproblemsoccurof

warningsaresetattoohigh:

Thereisthepotentialforhearinglossifthewarnings

areactivatedfrequently.

Loudwarningsleadpeopletoswitchthemoffandnot

turnthembackonagain.

Inmanyenvironmentsaloudwarningmaydisturb

people.Thisisespeciallytrueinahospitalsetting

wherepatientsmayalreadyhavealowtolerancefor

noiseduetoillness.

Loudwarningstendtobeturnedoffbeforethesituationtheyaresignallingisattendedto.Inmany

situationsthetimerequiredtoturnoffthealarmmay

becritical.

Loudwarningsoundsproducestartlereactionswhich

hinderconcentrationatatimewhenthismaybeimportant.

Loudwarningsoundshinderverbalcommunication.


41/43


Somesoundsaremoreappropriatethanotherfor

indicatingaparticularsituation. Ifasoundcanbe

chosenwhichhelpsthelistenertounderstandthe

natureofthewarning,learningtimesareminimised

andconfusionavoided.Thetablebelowlistssome

potentialsourcesofconfusioninauditorywarnings.


42/43

Non-Verbal

AuditoryWarningsSometraditionaldesignguidelinesare:

Usefrequenciesbetween200and5000Hz,especially

from3000to5000Hzsincetheearismostsensitivein

thismiddlerange.

Usefrequenciesbelow1000Hzwhensignalshaveto

travellongdistances(morethan200metres)because

highfrequenciesarerapidlydissipatedinair.

Usefrequenciesbelow500Hzwhensignalshavetobendaroundobstaclesorpassthroughpartitions.

Usemodulatedsignals(1to8beepspersecond)since

theyaredifferentenoughfromnaturalsoundstodemand

attention.

Usesignalswithfrequenciesdifferentfromthoseofanydominantbackgroundnoisetominimisemasking.

Ifdifferentwarningsignalsrepresentingdifferent

conditionsareused,eachsoundshouldbeidentifiable

fromtheothers.

Wherepossible,useaseparatecommunicationsystem

forthewarnings.Examplesincludehorns,bellsand

loudspeakers.


43/43


Severalstudieshavemappedtheperceived

urgencyofwarningsoundstotheirfrequency

contentandtemporalpattern. Inonestudythe

StevensPowerLawwasusedtodetermine

urgencyexponentsforthespeed,numberofrepetitionsandfrequencyofthewarningsignal.

Thehigherthevalueoftheexponent,themore

Speed 1.35

NumberofRepetitions 0.50

Frequency 0.38

Human Factors Course Lec14 Sound Exposure

Documents