LightingDesign

By reading this book, youwill develop the skills to perceive a space and its contents inlight,andbeabletodevisealayoutofluminairesthatwillprovidethatlitappearance.

WrittenbyrenownedlightingexpertChristopher(Kit)Cuttle,thebook:

explainsthedifferencebetweenvisionandperception,whichisthedistinctionbetweenprovidinglightingtomakethingsvisible,andprovidingittoinfluencetheappearanceofeverythingthatisvisible;demonstrateshowlightingpatternsgeneratedbythree-dimensionalobjectsinteractingwithdirectionallightingarestronglyinfluentialuponhowthevisualperceptionprocessenablesustorecogniseobjectattributes,suchaslightness,colourfulness,textureandgloss;revealshowadesignerwhounderstandstheroleoftheselightingpatternsintheperceptualprocessmayemploythemeithertoreveal,ortosubdue,ortoenhancetheappearanceofselectedobjectattributesbycreatingappropriatespatialdistributionsoflight;carefullyexplainscalculationaltechniquesandprovideseasy-to-usespreadsheets,sothatlayoutsoflampsandluminairesarederivedthatcanbereliedupontoachievetherequiredilluminationdistributions.

Practical lighting design involves devising three-dimensional light fields that createluminoushierarchiesrelatedtothevisualsignificanceofeachelementwithinascene.Byproviding you with everything you need to develop a design concept – from theunderstandingofhowlightinginfluenceshumanperceptionsofsurroundings,throughtoengineering efficient and effective lighting solutions –KitCuttle instils in his readers anew-foundconfidenceinlightingdesign.

Christopher ‘Kit’ Cuttle,MA, FCIBSE, FIESANZ, FIESNA, FSLL, is visiting lecturer inAdvanced Lighting Design at the Queensland University of Technology, Brisbane,Australia, and is author of two books on lighting (Lighting by Design, 2nd edition,ArchitecturalPress,2008;andLight forArt’sSake,ButterworthHeinemann,2007).His previous positions include Head of Graduate Education in Lighting at the LightingResearchCenter,RensselaerPolytechnic Institute,Troy,NewYork;SeniorLecturerat theSchools of Architecture at the University of Auckland and the Victoria University ofWellington, both in New Zealand; Section Leader in the Daylight Advisory Service,

PilkingtonGlass;andLightingDesignerwithDerekPhillipsAssociates,bothintheUK.HisrecentawardsincludetheSocietyofLightandLighting’sLeonGaster2013AwardforhisLR&T paper ‘A New Direction for General Lighting Practice’, and the LifetimeAchievement Award presented at the 2013 Professional Lighting Design Conference inCopenhagen.

Publisher’sNote:

Todownloadthespreadsheetsthatareusedtofacilitatethecalculationsinthisbook,gotothe e-resources link shown on the back cover of the book and click oneResource/Downloads.

LightingDesign

Aperception-basedapproach

ChristopherCuttle

Firstpublished2015

byRoutledge

2ParkSquare,MiltonPark,Abingdon,OxonOX144RN

andbyRoutledge

711ThirdAvenue,NewYork,NY10017

RoutledgeisanimprintoftheTaylor&FrancisGroup,aninformabusiness

©2015ChristopherCuttle

TherightofChristopherCuttletobeidentifiedasauthorofthisworkhasbeenassertedbyhiminaccordancewith

sections77and78oftheCopyright,DesignsandPatentsAct1988.

Allrightsreserved.Nopartofthisbookmaybereprintedorreproducedorutilisedinanyformorbyanyelectronic,

mechanical,orothermeans,nowknownorhereafterinvented,includingphotocopyingandrecording,orinany

informationstorageorretrievalsystem,withoutpermissioninwritingfromthepublishers.

Trademarknotice:Productorcorporatenamesmaybetrademarksorregisteredtrademarks,andareusedonlyfor

identificationandexplanationwithoutintenttoinfringe.

BritishLibraryCataloguinginPublicationData

AcataloguerecordforthisbookisavailablefromtheBritishLibrary

LibraryofCongressCataloginginPublicationData

Cuttle,Christopher.

Lightingdesign:aperception-basedapproach/ChristopherCuttle.

pagescm

Includesbibliographicalreferencesandindex.

ISBN978-0-415-73196-6(hardback:alk.paper)—ISBN978-0-415-73197-3

(pbk.:alk.paper)—ISBN978-1-315-75688-2(ebook)1.Lighting,

Architecturalanddecorative—Design.2.Visualperception.I.Title.

NK2115.5.L5C882015

747’.92—dc23

2014009980

ISBN:978-0-415-73196-6(hbk)

ISBN:978-0-415-73197-3(pbk)

ISBN:978-1-315-75688-2(ebk)

TypesetinBembo

bySaxonGraphicsLtd,Derby

Contents

Listoffigures

Listoftables

Acknowledgements

Introduction

1Theroleofvisualperception

2Ambientillumination

3Illuminationhierarchies

4Spectralilluminationdistributions

5Spatialilluminationdistributions

6Deliveringthelumens

7Designingforperception-basedlightingconcepts

Appendix

Index

Figures

1.1TheCheckerShadowIllusion.SquaresAandBareidentical

1.2AwhitesheethasbeendrawnovertheCheckerShadowIllusion,withcut-outsforsquaresAandB,andnowtheyappeartobeidentical

1.3Previouslythecylindricalobjectappearedtobeuniformlygreen

1.4Theobjectattributesofthisbuildingareclearlyrecognisable(ChartresCathedral,France)

1.5ChartresCathedral,Francebutavastlydifferentappearance

2.1Tostartthethoughtexperiment,imaginearoomforwhichthesumofceiling,walls,andfloorareais100m2

2.2Totheroomisaddedaluminaire

2.3Allroomsurfacesaregivenaneutralgreyfinishsothatρrs=0.5

2.4Roomsurfacereflectanceisincreasedsothatρrs=0.8

2.5Roomsurfacereflectanceisreducedtozero,soρrs=0

2.6Thefinalstageofthethoughtexperiment

2.7ReflectanceplottedagainstMunsellValue

2.8Usinganinternallyblackenedtubemountedontoalightmetertoobtainameasurementofsurfacereflectance

2.9Thevalueofthereflectance/absorptanceratioisproportionaltomeanroomsurfaceexitance,MRSE

3.1Demonstrationset-upforgainingassessmentsofnoticeable,distinct,strongandemphaticilluminationdifferences

3.2Flowchartforachievingmeanroomsurfaceexitance,MRSE,andtask/ambientillumination,TAIR,designvalues

4.1RelativesensitivityfunctionsforV(λ),andthethreeconetypes;long-,medium-andshort-wavelength;L(λ),M(λ)andS(λ)

4.2TheVB3(λ)spectralsensitivityofbrightnessfunctionfordaytimelightlevels

4.3TheV(λ)andV′(λ)relativeluminousefficiencyfunctionsrelatetophotopicandscotopicadaptationrespectively

4.4Rea’sproposedVC(λ)functionfortherelativecircadianresponse

4.5Theblack-bodylocus(solidline)plottedontheCIE1931(x,y)chromaticitychart

4.6ThereciprocalmegaKelvinscale(MK−1)comparedwiththeKelvin(K)scale

4.7Contoursofperceivedleveloftint

4.8Kruithof’schartrelatingcorrelatedcolourtemperature(TC)andilluminance(E)tocolourappearance

4.9OutputfromCIE133W.execomputerprogramtocalculateCRIs,foraWarmWhitehalophosphatefluorescentlamp

4.10Colour-mismatchvectordataforahalophosphateCoolWhitecolour33fluorescentlamp

4.11GamutareasforsomefamiliarlightsourcesplottedontheCIE1976UCS(uniformchromaticityscale)diagram

4.12TheGretagMacbethColorCheckercolourrenditionchartbeingexaminedunderdaylight

5.1Thetripleobjectlightingpatternsdevice

5.2Forthethreelightingconditionsdescribedinthetext

5.3ThestrikingfirstviewoftheinterioroftheQELAboutique,Doha

5.4QELA–Thedisplaylightinginthecentralareahasstrongdownward‘flow’,with‘sharpness’creatingcrispshadowandhighlightpatterns

5.5QELA–Inthisdisplayarea,whichisadjacenttothecentralarea,thelowermeanroomsurfaceexitance(MRSE)levelhastheeffectofstrengtheningtheshadingpatterns

5.6QELA–Inthisdisplayarea,themannequinappearsisolatedbythestrongshadingpatterngeneratedbytheselectivelighting

5.7QELA–Ontheupperfloor,the‘fire’ontherightmatchesthewarmwhiteilluminationusedthroughouttheboutique

5.8ThepointPislocatedattheintersectionofthex,yandzorthogonalaxes

5.9Thethree-dimensionalilluminationdistributionaboutpointP

5.10TheilluminationsolidisnowthesumofcomponentsolidsduetosourcesS1andS2

5.11Theilluminationsolidatapointinaspacewherelightarrivesfromeverydirection

5.12Themagnitudeanddirectionof(EA–EB)maxdefinestheilluminationvector,whichisdepictedasanarrowactingtowardsthepoint

5.13Thisisthesymmetricsolid

5.14In(a),asmallsourceSprojectsluminousfluxofFlmontoadiscofradiusr,producingasurfaceilluminanceE=F/(π.r2).In(b),thediscisreplacedbyasphereofradiusr,givingasurfaceilluminanceE=F/(4π.r2)

5.15(a):VerticalsectionthroughPshowingilluminationvectoraltitudeangleα,and(b):HorizontalsectionthroughPshowingazimuthangleφofthehorizontalvectorcomponent

5.16ThepointPisonasurface,andisilluminatedbyadisc-shapedsourcethatisnormaltothesurfaceandofangularsubtenceα

5.17Thiscomparisonsurfacehastwomountedsamplesthatresponddifferentlytothediscsource

5.18Asthesubtenceofalargediscsourceisreduced,thesourceluminancerequiredtomaintainanilluminancevalueof100luxincreasesrapidlyassubtencefallsbelow30degrees

5.19Forsmallsources,theincreaseinluminancerequiredtomaintain100luxincreasesdramaticallyforsubtenceangleslessthan3degrees

5.20HighlightcontrastpotentialHLCforthreevaluesoftargetreflectance

5.21Lightsourcesofsmallersubtenceangleproducelesspenumbra,increasingthe’sharpness’ofthelighting

6.1Measuringsurfacereflectance,usinganinternallyblackenedcardboardtubefittedoveranilluminancemeter

6.2Applicationofthepoint-to-pointformula

6.3DeterminingtheilluminanceatpointPonaverticalplane

6.4ThepointPisilluminatedbytwoalternativesources

6.5ThecorrectionfactorC(D/r)tobeappliedtopointsourceilluminationformulae

6.6TheCubicIlluminationconcept

6.7ThelocationofsourceSrelativetoathree-dimensionalobjectisdefinedintermsofX,Y,andZdimensions

6.8Assessmentoflikelyprospectsforvariousrolesforfenestrationinbuildings

6.9AsimplewayofmakinganapproximatemeasurementofMRSEusingaconventionallightmeter

6.10Asix-photocellcubicilluminationmeter

6.11Themeasurementcubeistiltedsothatalongaxisiscoincidentwiththezaxis,and

threefacetsfaceupwardsandthreedownwards

6.12Averticalsectionthroughthetiltedcubeontheuaxis,whichliesinthesameverticalplaneastheyaxis,againstwhichitistiltedthroughtheanglea

6.13Aphotocellheadmountedonaright-anglebracket,ontoaphotographictripod

6.14Thephotocelltiltedto+35degreesrelativetothehorizontalplane

7.1Alightingdesignflowchart

7.2TAIRvaluesforthehorizontalworkingplane,whenitisthetarget

7.3Theinfluenceofroomsurfacereflectionproperties.

Tables

2.1Perceivedbrightnessordimnessofambientillumination

2.2Perceiveddifferencesofexitanceorilluminance

4.1The14CIETCS(Testcoloursamples).TCS1–8comprisetheoriginalsetofmoderatelysaturatedcoloursrepresentingthewholehuecircle,andthesearetheonlysamplesusedfordeterminingCRI.Theothersixhavebeenaddedforadditionalinformation,andcomprisefoursaturatedcolours,TCS9–12,andtwosurfacesofparticularinterest.Regrettably,detailsofcolourshiftsfortheseTCSareseldommadeavailable

5.1Vector/scalarratioandtheperceived‘flow’oflight

7.1Valuesoftarget/ambientilluminanceratio,TAIR,againstroomindexwherethehorizontalworkingplane,HWP,isthetargetsurfaceandalldirectfluxisincidentontheHWP.Lightsurfacereflectancesareassumed

Acknowledgements

Thecontentsof thisbookhavegrown from theAdvancedLightingDesigncourse that Ihavetaughteveryyearsince2005attheQueenslandUniversityofTechnologyinBrisbane,Australia, forwhichI thanktheprogrammecoordinator,Professor IanCowling,andalsothesuccessionoflivelyandenquiringCPD(continuingprofessionaldevelopment)studentswhohavecausedmetokeepthecurriculuminastateofcontinualrevision.

Whilemanypeoplehavecontributedtothedevelopmentoftheideascontainedinthisbook,whethertheyrealiseditatthetimeornot,threeformercolleagueswithwhomIhavemaintainedemailcontacthaverespondedtospecificissuesthatIencounteredinpreparingthe text. They are, in no particular order, Joe Lynes and ProfessorsMarkRea and PeterBoyce.Mythankstoeachofthem.

Thosewhohavegivenpermissionformetoreproducefiguresareacknowledgedinthecaptions, but Iwant tomakeparticularmentionofEdwardAdelson,ProfessorofVisionScience at the Massachusetts Institute of Technology, who not only permitted me toreproduce his Checker Shadow Illusion (Figure 1.1), but also two of my own modifiedversionsofhisbrilliantfigure.

Introduction

Theaimofthisbookistoenablepeoplewhoarefamiliarwiththefundamentalsoflightingtechnology to extend their activities into the field of lighting design.While the text isaddressed primarily to students, it is relevant to professionals working in the fields ofbuildingservices,interiordesignandarchitecture.

Thepremiseof this book is that thekey to lightingdesign is the skill tovisualise thedistribution of light within the volume of a space in terms of how it affects people’sperceptionsofthespaceandtheobjects(includingthepeople)withinit.Theaimisnottoproduce lighting that will be noticed, but rather, to provide an envisioned balance ofbrightnessthatsetstheappearanceofindividualobjectsintoanoveralldesignconcept.

This isdifferentfromcurrentnotionsof ‘goodlightingpractice’,whichaimtoprovideforvisibility,whereby‘visualtasks’maybeperformedefficientlyandwithoutpromotingfatigueordiscomfort. It isalsoquitedifferent fromsome lightingdesignpractice,wherespectacular effects are achieved by treating the architecture as a backdrop onto whichpatternsofcolouredlight,orevenbrilliantimages,areprojected.

Several perception-based lighting concepts are introduced to enable distributions ofillumination tobedescribed in termsofhow theymay influence theappearanceofa litspace.Thesedescriptionsinvolveperceivedattributesofillumination,suchasilluminationthatbringsout‘colourfulness’,orhasaperceived‘flow’,orperhaps‘sharpness’.Itisshownthatthethree-dimensionaldistributionsofilluminationthatunderliethisunderstandingoflighting can be analysed in quantitative terms, enabling their characteristics to bemeasured and predicted. The principles governing these distributions are explained, andspreadsheets are used to automatically perform the calculations that relate perceivedattributestophotometricquantities.

Theobjectiveis toenablea lightingdesignertodiscuss lightingwithclientsandotherprofessionals in terms of how illuminationmay influence the appearance of spaces andobjects.Whenagreementisreached,thedesigneristhenabletoapplyproceduresthatleadtolayoutsofluminairesandstrategiesfortheircontrol,andtodothiswithconfidencethattheenvisionedappearancewillbeachieved.

1TheRoleofVisualPerception

Chaptersummary

The Checker Shadow Illusion demonstrates a clear distinction between the processes ofvision and perception, where vision is concerned with discrimination of detail andperceptioninvolvesrecognitionofsurfaceandobjectattributes.Theroleoflightinginthisrecognition process involves the formation of lighting patterns created by interactionsbetween objects and the surrounding light field. Confident recognition comprises clearperception of both object attributes and the light field. Three types of object lightingpatterns are identified, being the shading, highlight, and shadow patterns, and it is bycreating light fields that produce controlled balances of these three-dimensional lightingpatterns that designers gain opportunities to influence how room surface and objectattributesarelikelytobeperceived.

Theevidenceofyoureyes

Figure1.1 shows theChecker Shadow Illusion, andat first sight, thequestionhas to be,whereistheillusion?Everythinglooksquitenormal.TheanswerliesinsquaresAandB:theyareidentical.Thatistosay,theyarethesameshadeofgreyandtheyhavethesamelightness,ortobemoretechnical,theyhavethesamereflectance(andtherebyluminance)andthesamechromaticity.

Doyoufindthiscredible?Theycertainlydonotlookthesame.NowlookatFigure1.2,whichshowsawhitesheetdrawnoverthefigurewithcut-outsforthetwosquares.Seeninthiswaytheydolookthesame,andifyoutakeapieceofcardandpunchaholeinit,youcanslideitoverthepreviousfigureandconvinceyourselfthatthetwosquaresareinfactidenticalandasshowninFigure1.2.

Thisraisesaquestion:howisitthat,whentheimagesofthesetwoidenticalsquaresaresimultaneouslyfocussedontotheretina,inonecase(Figure1.2)theyappearidenticalandintheother(Figure1.1)theyappeardistinctlydifferent?

Figure1.1TheCheckerShadowIllusion.SquaresAandBareidentical.Theyarepresentedhereasrelatedcolours,thatis

tosay,theyappearrelatedtotheirsurroundings.Thelightingpatternsthatappearsuperimposedoverthesurrounding

surfacescauseaviewertoperceivea‘flow’oflightwithinthevolumeofthisspace,andwhichleadstothematching

luminancesofAandBbeingperceivedquitedifferently.

(Source:en.wikipedia.org/wiki/Checker_shadow_illusion.html,downloadedJanuary2013)

Figure1.2AwhitesheethasbeendrawnovertheCheckerShadowIllusion,withcut-outsforsquaresAandB,andnow

theyappeartobeidentical.Inthiscasetheyarepresentedasunrelatedcolours.

Relatedandunrelatedcolours

The essential difference is that in Figure 1.1 the two squares are presented as relatedcolours, that is to say, colours are perceived to belong to surfaces or objects seen inrelationtoothercolours,andinFigure1.2,theyareshownasunrelatedcolours,meaningtheyareseeninisolationfromothercolours(Fairchild,2005).Asunrelatedcolours(greyisacolour), theyareperceived tocomprisenothingmore thanrectangularcolouredshapesonaplainwhitebackground,butwhentheyaresetintothecontextofFigure1.1,theyareperceived as solid elements in a three-dimensional scene that have recognisable objectattributes. It is this change in the way they are perceived that causes them to appeardifferently.

So what are the components of the surrounding scene that make this illusion soeffective?Askyourself,whyisthecylindricalobjectthere?Doesitcontributesomething?Infact,itisavitalcomponentoftheillusion.So,whatcolourisit?Obviously,green.Isituniformlygreen?Well,yes…butlookmorecarefullyattheimageoftheobjectandyouwill see that both its greenness and its lightness vary hugely. The image is far fromuniform,sohowdidyousupposetheobjecttobeuniformlygreen?Theansweristhatyouperceivedadistinctive lightingpatternsuperimposedover theuniformlygreenobject. InFigure 1.3, the area enclosed by the object outline is shown as uniformly green and itappearsasnothingmorethanaformlessblob.

Thesolid,three-dimensionalobjectperceivedinFigure1.1isobservedtobeinteractingwithadirectional ‘flow’of light,which causes ashadingpattern to be generated, andthis appears superimposed over the green object surface. Note also that the cylinder’ssurface is not perfectly matt, and there is just a hint of a highlight pattern due to aspecularcomponentofreflectionthatisapparentattheroundedrimofthecylinder’stopedge.Theselightingpatternsinformyouabouttheobject’sattributes(Cuttle,2008).

Figure1.3Previouslythecylindricalobjectappearedtobeuniformlygreen.Nowitisuniformlygreen,butitdoesnot

looklikeacylinder.Thatisbecauseitisnowlackingthelightingpatternduetointeractionwiththe‘flow’oflight.

Nowlookatthecheckerboardsurface.Againwehaveapatternduetothelighting,butinthiscaseitisashadowpattern,whichhasadifferentappearancefromtheshadingandhighlight patterns, but nonetheless is quite consistentwith our perception of the overall‘flow’oflightwithinthevolumeofthespace.Itwillbeobvioustoyouthatiftwosurfaceshavethesamelightness(whichalsomeanstheyhavethesamereflectance)andoneoccurswithin theshadowpatternandoneoutside it, theywillhavedifferent luminancevalues.Thecreatorofthisbrilliantillusion,EdwardH.Adelson,ProfessorofVisionScienceattheMassachusettsInstituteofTechnology,hascarefullysetitupsothatsquaresAandBhavethe same luminance value,whichmeans of course, that their images on your retina areidentical.However,thefunctionofthevisualprocessistoprovideinformationtothevisualcortexofthebrain,andhereyourperceptualprocessistellingyouthat,althoughthesetwosquaresmatchforluminance,theycannothavethesamelightness.Theoneintheshadowmustbelighter,thatistosay,itmusthavehigherreflectance,thantheoneinfulllight.Youhold this innate understanding of lighting in your brain, and you cannot apply yourconsciousmindtooverruleit.

Inthisway,itcanbeseenthattheimagefocussedontotheretinaissimplyanopticalprojection of the visual scene that corresponds directly with the luminance andchromaticityvaluesoftheelementswithintheexternalscene.Sinceitsinception,thestudyoflightinghasconcentratedonthevisualprocessandhowilluminationmaybeappliedtoprovideforvisibility,laterdefinedintermsofvisualperformance,buttheroleofvisionistoservetheprocessofperception,andthisoccursnotattheretina,butinthevisualcortex

ofthebrain.Whatweperceiveisnotapatternofbrightnessandcolour,butagestalt,thisbeingapsychologicaltermthatdescribestheholisticentitythatenablesustorecogniseallthe forms and objects that make up our surroundings (Purves and Beau Lotto, 2003).Consciously,weareawareofthree-dimensionalspacesdefinedbysurfacesandcontainingobjects,butinordertomakethismuchsenseoftheflowofinformationarrivingthroughtheopticnerve,wehavetobesubconsciouslyawareofalightfieldthatfillsthevolumeofthe space.This ishowwemake senseof squaresAandB. Seen in thisway, it becomesobvious why attempts to analyse scenes in terms of luminance and chromaticity wereboundtoleadtofrustration.

Theroleofambientillumination

Formostofthetime,weliveinaworldofrelatedcolours.Wearesurroundedbysurfacesand objects which, providing the entire scene is adequately illuminated, our perceptualfaculties reliably recognise andmake us aware of, sometimes so thatwe can copewitheverydaylife,andsometimestoelevateoursensestohigherlevelsofappreciation,aswhenwe encounter artworks or beauties of nature. Recognition involves identifying objectattributes associatedwith all of the things thatmake up our surrounding environments,and our innate skill in doing this is truly impressive. Scientists working on artificialintelligencehavetriedtoprogramsupercomputerstoperforminthisway,butsofartheirbesteffortsfallfarshortofwhathumanperceptionachieveseverymomentthroughoutourwakinghours.

Provided that ambient illumination is sufficient, we are able to enter unfamiliarenvironments, orientate ourselves, and go about our business without hesitating toquestion the reliabilityof theperceptionswe formof the surroundingenvironment. It isclearthatsubstantialprocessinghastooccur,veryrapidly,betweentheretinalimageandformation of the perception of the environment. There is no good reason why ourperceptions of elements of the scene should show in-step correspondence with theirphotometriccharacteristics.Visualperceptionmaybethoughtofastheprocessofmakingsenseoftheflowofsensoryinputthroughtheopticnervetothebrain,wherethepurposeis to recognise surfaces and objects, rather than to record their images. Colours areperceived as related to object attributes, and effects of illumination are perceived aslightingpatternssuperimposedoverthem.AswerecognisedthecylinderinFigure1.1 tobe uniformly green with a superimposed shading pattern, so we also recognised theidenticalsquarestodifferinlightnessbecauseofthesuperimposedshadowpattern.

There will, however, be situations where we are confronted with elements seen inisolation from each other, and this is particularly likely to occur in conditions of lowambient illumination.Whenwefindourselvesconfrontedbydarksurroundings,relianceuponrelatedcoloursandidentificationofobjectattributesmaygivewaytoperceptionofunrelatedcolours,andwhenthisoccurs,ourperceptionsdonotdistinguishlightnessandilluminanceseparately,andluminancepatternsdominate.Thatistosay,theappearancesofindividual objects within the scene relate to their brightness and chromaticity values,ratherthanuponrecognitionoftheirintrinsicattributes.

Figures1.4and1.5showtwoviewsofthesamebuilding.InFigure1.4,weseeaviewofthismagnificentcathedralinitssetting,andwereadilyformasenseofitssubstantialmassandthematerialsfromwhichitisconstructed.Also,evenifwearenotconsciousofit,weperceivetheentirelightfieldthatgeneratesthisappearance.InFigure1.5,ourperceptionof this building is quite different. We have no notion of a natural light field, and thebuildingseemstofloat,unattachedtotheground.Itisrevealedbyaglowinglightpattern

thatdoesnotdistinguishbetweenmaterials,andactuallymakesthebuildingappearself-luminous.Thebuilding’sappearanceisdominatedbybrightness,andobjectattributesarenotdiscernible.Thesetwoviewsshowclearlythedifferencebetweenrelatedcolours,inthedaylight view, and unrelated colours in the night-time view. They also give us dueappreciationoftherolethatlightingmayplayinbringingaboutfundamentaldifferencesinourperceptions.

Undernormaldaytimelighting, two-wayinteractionsoccurthatenableourperceptualprocesses tomake sense of the varied patterns of light and colour that are continuouslybeing focussed onto our retinas. Working in one direction, there is the process ofrecognisingobjectattributesthatarerevealedbythe lightingpatterns,whileat thesametime,andworkingintheoppositedirection,itistheappearanceoftheselightingpatternsthat provides for the viewer’s understanding of the light field that occupies the entirespace.

Figure1.4Theobjectattributesofthisbuildingareclearlyrecognisable,andtheambientilluminationprovidesamplyfor

allelementstoappearasrelatedcolours.(ChartresCathedral,France.)

Figure1.5Thesamebuilding,butavastlydifferentappearance.Lowambientilluminationprovidesadarkbackdrop

againstwhichthecathedralglowswithbrightness.Objectattributesareunrecognisableinthisexampleofunrelated

colours.

Perceptionasabasisforlightingdesign

Fromadesignpointofview,lightingpracticemaybeseentofallintotwobasiccategories.Ononehand,forilluminationconditionsrangingfromoutdoordaylighttoindoorlightingwheretheambientlevelissufficienttoavoidanyappearanceofgloom,weliveinaworldof related colours in which we distinguish readily between aspects of appearance thatrelate to the visible attributes of surfaces and objects, and aspects which relate to thelightingpatternsthatappearsuperimposeduponthem.

Ontheotherhand,inconditionsoflowambientillumination,wherewehaveasenseofdarknessorevengloom,whetherindoorsor,mostnotably,outdoorsatnight,wetypicallyexperience unrelated colours and this may lead to the appearances of objects andsurroundings dominated by brightness patterns that may offer no distinction betweenobjectlightnessandsurfaceilluminance.

Theimplicationsofthisdichotomyforlightingdesignareprofound.Outdoornight-timelighting practice, such as floodlighting and highway illumination, is based on creatingbrightnesspatterns thatmaybear littleornorelationship to surfaceorobjectproperties.Alternatively,forsituationswhereambientilluminationisatleastsufficienttomaintainanappearance of adequacy (apart from outdoor daylight, this may be taken to include allindoorspaceswheretheilluminationcomplieswithcurrentstandardsforgenerallightingpractice)wetakeinentirevisualscenesincludingobjectattributes,andinvolvinginstantrecognitionoffamiliarobjectsandscrutinyofunfamiliarorotherwiseinterestingobjects.The identification of object attributes may become a matter of keen interest, as whenadmiring an art object or seeking to detect a flaw in a manufactured product, and wedepend upon the lighting patterns to enable us to discriminate and to respond todifferencesofobjectattributes.

Betweenthesetwosetsofconditionsisarangeinwhichsomeuncertaintyprevails.Wehave, for example, all experienced ‘tricks of the light’ that can occur at twilight, andgenerally, recommendations for good lighting practice aim to avoid such conditions.Perhaps surprisingly, it iswithin this range that lightingdesignersachieve someof theirmostspectaculardisplayeffects.By isolatingspecificobjects fromtheirbackgroundsandilluminating them from concealed light sources, lighting can be applied to alter theappearance of selected object attributes, such as making selected objects appear moretextured, or colourful, or glossy. All of this thinking will be developed in followingchapters.

Beforeweclosethischapter,askyourself,whydowecallFigure1.1anillusion?Ifthepage is evenly illuminated, squaresA andBwill have the same luminance and so theystimulate their corresponding areas of our retinas to the same level. The fact that theseequalstimulidonotcorrespondtoequalsensationsofbrightnessiscitedasanillusion.Thepointneedstobemadethatvisionservestheprocessofperception,andperceptionisnot

concerned with assessing or responding to luminance. Its role is to continually seek torecognise object attributes from the flow of data arriving from the eyes.When we areconfrontedwithFigure1.1inaconditionofadequateillumination,ourperceptionprocessperforms its task toperfection.A iscorrectly recognisedasadarkcheckerboardsquare,andBasalightsquare.RatherthanlabellingFigure1.1asanillusion,perhapsweshouldrefertoitasaninsightintotheworkingsofthevisualperceptionprocess.

However, therealpurposeforexaminingthis imagehasbeentoshowhowperceptiondependsuponandisinfluencedbythelightingpatternsthatobjectsandsurfacesgeneratethroughinteractionswiththeirsurroundinglightfields.Theselightingpatternsmayhavetheeffectsofrevealing,subduing,orenhancingselectedobjectattributes,anditisthroughcontroloflightfielddistributionsthatlightingdesignersinfluencepeople’sperceptionsofobject attributes. Skill in exercising this control, particularly for indoor lighting, is theessenceoflightingdesignandthecentralthemeofthisbook.

References

Cuttle,C.(2008).LightingbyDesign,SecondEdition.Oxford:ArchitecturalPress.

Fairchild,M.D.(2005).ColorAppearanceModels,SecondEdition.Chichester:Wiley.

Purves,D.andR.BeauLotto(2003).Whyweseewhatwedo:Anempiricaltheoryofvision.Sunderland,MA:SinauerAssociates.

2AmbientIllumination

Chaptersummary

Theperceptionofambientilluminationconcernswhetheraspaceappearstobebrightlylit,dimlylit,orsomethinginbetween.Atfirstthismightseemarathersuperficialobservationuntilweconsideralloftheassociationsthatwehavewith‘brightlight’and‘dimlight’,atwhich point ambient illumination becomes a key lighting design concept. It provides abasis for planning lighting based on the perceived difference of illumination betweenadjacentareas,orspacesseeninsequenceaswhenpassingthroughabuilding.Athoughtexperiment is introducedwhich leads to theconclusionthatmeanroomsurfaceexitance(MRSE)providesausefulindicatorofambientillumination,whereMRSEisameasureofinter-reflectedlightfromsurroundingroomsurfaces,excludingdirectlightfromwindowsorluminaires.TheAmbientIlluminationspreadsheetfacilitatesapplicationofthisconcept.

Theamountoflight

An important decision in lighting design is, ‘What appearance of overall brightness (ordimness) is this space to have?’General lighting practice gives emphasis to the issue ofhowmuchlightmustbeprovidedtoenablepeopletoperformthevisualtasksassociatedwithwhateveractivityoccurswithinthespaceand,ofcourse,thismustalwaysbekeptinmind. In a banking hall, for example,we need to ensure that the counters are lit to anilluminance that is sufficient to enable the tellers to perform theirwork throughout theworking day without suffering strain. While that aspect of illumination must not beoverlooked, there is an overarching design decision to be made, which is whether theoverall appearanceof the space is tobe abright, lively and stimulating environment, orwhetheramoredimoverallappearanceiswanted.Theaimofadimappearancemaybetopresentasubdued,andperhapssombre,appearance,oralternatively,tocreateasettinginwhichilluminationcanbedirectedontoselectedtargetstopresenttheminhighcontrastrelative to their surroundings. Of course, the surroundings cannot be made too dim asilluminationmustalwaysbesufficientforsafemovement,butthereissubstantialscopeforadesignertochoosewhether,inaparticularsituation,theoverallimpressionistobeofabrightspace,orofadimspace,orofsomethinginbetween.Clearly,theimpressionsthatvisitorswouldformofthespacewillbesubstantiallyaffectedbythedesigner’sdecision.

This raises a question. Ifwe are not lighting a visual task plane for visibility, but areinstead illuminating a space for a certain appearance of overall brightness, how do wespecify the level of illumination that will achieve this objective? All around the world,lightingstandards,codes,andrecommendedpracticedocumentsspecifyilluminationlevelsfor various indoor activities in terms of illuminance (lux) and a uniformity factor. Ifsomeone states that ‘This is a 400 lux installation’, that means that illuminance valuesmeasuredonthehorizontalworkingplane,usuallyspecifiedasbeing700mmabovefloorlevelandextendingfromwalltowallwithinthespace,shouldaverageatleast400lux,andfurthermore,atnopointshouldilluminancedroptolessthan80percentofthataveragevalue.

Thereasonsforthisarehistorical.Itwasinthelatenineteenthcenturythatthepracticeof measuring illumination emerged, and for indoor lighting, the prime purpose was toenable working people to remain productive for the full duration of the working day,despite daylight fluctuations.While the recommended illuminance levels have increasedmorethantenfoldsincethosedays,themeasurementproceduresareessentiallyunchangedeven though light meters have undergone substantial development. The two specifiedmeasures, an average illuminance and the uniformity factor, are the means by whichlighting quantity is specified, andmore than that, they govern how people think aboutilluminationquantity.Perhapstheworstfeatureofthesespecificationsisthattheyhavetheeffectofinhibitingexplorationofdifferentwaysinwhichthelightmightbedistributedin

aspace,andhowlightingmaybeappliedtocreatealitappearancethatrelatestoaspaceand the objects it contains. For lighting designers, these aspects of appearance are all-important,andinfact,itmaybesaidthattheyformtheverybasisofwhatlightingdesignisallabout.Tobeobligedtoensurethatalllightingis‘codecompliant’isnothingshortofadenialtopursuethemostfundamentallightingdesignobjectives.

Athoughtexperiment

Weare going to conduct a thought experiment as a first step to exploring how lightingdoes more than simply make things visible, and in fact, we are going to explore howlightingaffectstheappearanceofeverythingwesee.Tostart,youneedtogetyourselfintoan experimentalmindset. The first requirement is to forget everything you know.Then,imagine an indoor spacewhere the sum total of ceiling,wall and floor areas add up to100m2,asshowninFigure2.1.

Then,intothisspaceisaddedaluminairethatemitsatotalaluminousflux,F,of5000lumens(Figure2.2).

How brightly litwill the space appear? Thismight seem to be a difficult question toanswer,whichisas itshouldbebecauseavitalpieceof informationis lacking.Until theroom surface reflectance values are specified, you have noway of knowing howmuchlightthereisinthisspace.

Figure2.1Tostartthethoughtexperiment,imaginearoomforwhichthesumofceiling,walls,andfloorareais100m2.

Figure2.2TotheroomisaddedaluminairewithatotalfluxoutputF=5000lumens.

Tokeeplifesimple,wewillspecifythatallroomsurfaceshaveareflectancevalue,ρrs,of0.5,thatistosay,50percentofincidentlumensareabsorbedand50percentarereflected(Figure2.3).Nowwecanworkouthowmanylumensthereareinthespace.

Howmuchlightdowehave?

addition total

Initialflux(F) 5000 5000Firstreflection 2500 7500Secondreflection 1250 8750Thirdreflectionandsoon… 625 9375

Figure2.3Allroomsurfacesaregivenaneutralgreyfinishsothatρrs=0.5.

All of the initial 5000 lumens from the luminaire are incident on room surfaces thatreflect 50 per cent back into the space, so the first reflection adds 2500 lm, bringing thetotalluminousfluxinthespaceupto7500lm.Thesereflectedlumensareagainincidenton room surfaces, and the second reflection adds another 1250 lumens to the total. Theprocess repeats, so that you could go on adding reflected components of the initial fluxuntil theybecome insignificantlysmall.Alternatively, theeffectofan infinitenumberofreflectionsisgivenbydividingtheinitialfluxby(1–ρ),sothat:

An interestingpointemergeshere.Wehavesurroundedthe luminairewithsurfaces thatreflect 50 per cent of the light back into the space, and this has doubled the number oflumens.Keepthispoint inmind.Nowwedividethetotal fluxbythetotalroomsurfaceareatogettheaverageroomsurfaceilluminance:

Atlastwehaveameasurewecanunderstand.Thiswouldbeenoughlightforustoseeour

wayaroundthespace,butnotenoughtomaketheroomappearbrightlylit.Let’ssupposethat we want a reasonably bright appearance. Well, we could fit a bigger lamp in theluminaire, butbeforewe take that easyoption, let’s thinkabitmoreabout the effectofroom surface reflectance.Wehave seen that it canhave a quite surprising effect on theoverallamountoflightinthespace.

Whatwouldbetheeffectofincreasingρrsto0.8,asshowninFigure2.4?Combiningtheexpressionsweusedbefore,itfollowsthatthemeanroomsurfaceilluminance:

Thisdeserves somecarefulattention.We increasedρrs from0.5 to0.8,which isa60percentincrease,andthetotalfluxincreasedtwo-and-a-halftimes!Howcanthisbeso?Thinkaboutitthisway.Itisconventionaltorefertosurfacereflectancevalues,buttrythinkinginsteadofsurfaceabsorptancevalues,whereα=(1–ρ).Whatwehavedonehasbeentoreduceαrsfrom0.5to0.2,andthatiswherethe2.5factorcomesfrom.

Figure2.4Roomsurfacereflectanceisincreasedsothatρrs=0.8.

Asthisisathoughtexperiment,thinkaboutwhatwouldhappenifwecouldreduceαrs

to zero.Well, the lumenswould just keep bouncing around inside the room.When youswitchedontheluminaire,thetotalfluxwouldkeeponincreasing.Ifyoudidnotswitchoffintime,theroomprobablywouldexplode!Ifyoudidswitchoffintime,thelightlevelwouldremainconstant.Youcouldcomebackamonthlateranditwouldbeundiminished,untilyouopenthedoorandinaflashallthelumenspouroutandtheroomwouldbein

darkness.Thoughtexperimentsreallycanbefun.Nowthinkaboutgoingintheoppositedirection.

What would be the effect of reducing ρrs to zero? How brightly lit would the roomappear? The question is of course meaningless. The only thing visible would be theluminaire,asshowninFigure2.5.Ifyouweresufficientlyadventurous,youcouldfeelyourwayaroundtheroomandyoucouldusea lightmeter toconfirmthevalueof themeanroomsurfaceilluminance:

The meter would respond to those 50 lux, but your eye would not. Here is anotherimportantpoint.Thedirectfluxfromtheluminairehasnoeffectontheappearanceoftheroom. It is not until the flux has undergone at least one reflection that it makes anycontribution towards our impression of howbrightly, or dimly, lit the roomappears. Tohaveausefulmeasureofhowtheambientilluminationaffectstheappearanceofaroom,weneedtoignoredirectlightandtakeaccountonlyofreflectedlight.

Figure2.5Roomsurfacereflectanceisreducedtozero,soρrs=0.

Let’sthinknowaboutageneralexpressionforambientilluminationasitmayaffectourimpressionofthebrightnessofanenclosedspace.Theluminaireistobeignored,andsoinFigure2.6,itisshownblack.Admittedly,ablackluminaireemitting5000lmisrathermoredemanding of the imagination, but bear with the idea. To take account of only light

reflected fromroomsurfaces,weneedanexpression formeanroomsurfaceexitance,MRSE,whereexitanceexpressestheaveragedensityofluminousfluxexiting,oremergingfrom,asurfaceinlumenspersquaremetre,lm/m2.

Figure2.6Thefinalstageofthethoughtexperiment.AblackluminaireemitsFlminaroomofareaAanduniform

surfacereflectanceρ,andmeanroomsurfaceexitance,MRSE,ispredictablefromFormulae2.1and2.2.

TheupperlineofFormula2.2isthefirstreflectedfluxFRF,whichistheinitialfluxafterit has undergone its first reflection. This is the energy that initiates the inter-reflectionprocess that makes the spaces we live in luminous. More descriptively, it is sometimesreferredtoasthe‘firstbounce’lumens.

The bottom line is the room absorption, Aα. One square metre of perfectly blacksurfacewouldcomprise1.0m2ofroomabsorption;alternatively,itmaycomprise2.0m2ofamaterialforwhichα=0.5,oragain,4.0m2ifα=0.25.Itisafactthatwhenyouwalkintoaroom,theambientilluminationreducesbecauseyouhaveincreasedtheroomabsorption.Youcouldminimisethateffectbywearingwhiteclothing,butthatisunlikelytocatchonamonglightingdesigners.Myownobservationisthatiflightingdesignerscanbesaidtohaveauniform,itisblack.Itseemsweaspiretobeperfectlightabsorbers!

TheMRSEconcept

Ofcourse,realroomsdonothaveuniformreflectancevalues,butthiscanbecopedwithwithoutunduecomplication.

OnthetoplineofFormula2.1,FρistheFirstReflectedFlux,FRF,whichisthesumof‘firstbounce’ lumens fromallof the roomsurfaces, suchasceiling,walls,partitionsandanyothersubstantialobjectsintheroom.Itisobtainedbysummingtheproductsof:

directilluminanceofeachsurfaceEs(d)

surfaceareaAs

surfacereflectanceρs

So,inaroomhavingnsurfaceelements:

On the bottom line of Formula 2.1, A(1 – ρ) is the RoomAbsorption, indicated by thesymbolAα,anditisameasureoftheroom’scapacitytoabsorblight.Asitisconventionaltodescribesurfacesintermsofreflectanceratherthanabsorptance;

Thegeneralexpressionformeanroomsurfaceexitance,Formula2.2,maybesummarisedas:

Themeanroomsurfaceexitanceequalsthefirstbouncelumensdividedbytheroomabsorption.

MRSEhasthreevaluableuses:

1TheMRSEvalueprovidesanindicationoftheperceivedbrightnessordimnessofambientillumination.Table2.1givesanapproximateguideforthetwodecadesof ambient illumination that cover the range of indoor general lighting practice.Thesevaluesarebasedonvariousstudiesconductedbytheauthorandreportedbyother researchers, and it should be noted that ambient illumination relates to aperceived effect, while MRSE is a measurable illumination quantity, likeilluminance,butnottobeconfusedwithworkingplaneilluminance.

Table2.1Perceivedbrightnessordimnessofambientillumination

Meanroomsurfaceexitance(MRSE,lm/m2)

Perceivedbrightnessordimnessofambientillumination

10 Lowestlevelforreasonablecolourdiscrimination

30 Dimappearance100 Lowestlevelfor‘acceptablybright’appearance300 Brightappearance1000 Distinctlybrightappearance

2TheMRSEratioforadjacentspacesprovidesanindexoftheperceiveddifferenceofillumination.Table2.2givesanapproximateguideforthisperceiveddifferenceas one moves from space to space within a building, or to the appearance ofdifferently

Table2.2Perceiveddifferencesofexitanceorilluminance

Exitanceorilluminanceratio Perceiveddifference

1.5:1 Noticeable3:1 Distinct10:1 Strong40:1 Emphatic

illuminatedsurfacesorobjectswithinaspace.Thereismoreaboutthisperceiveddifferenceeffectinthefollowingchapter.

3Itmayprovideanacceptablemeasureofthetotalindirectilluminancereceivedbyan object or surface within the space, so that for a surface S, the total surfaceilluminancemaybeapproximatelyestimatedbytheformula:

where Es(d) is the direct illuminance of surface S. Procedures for predicting directilluminationareexplainedinChapter6.

BeforeweexaminehowMRSEmaybeappliedinthedesignprocess,Iamconsciousthatsomereadersmaybefindingtheexitancetermunfamiliar,asitofteniscustomarytoreferto illuminance as themetric for incident light, and luminance for reflected light. To seewhere exitance fits in, take a step back. Illuminance is a simple concept. It refers to thedensityof luminous flux incidenton a surface, either at apoint or over an area, in lux,where1luxequals1lumenpersquaremetre(lm/m2).Exitanceisalsoasimpleconcept.Itrefers to thedensityof fluxexiting,or emerging from,a surface in lm/m2. (It shouldbenotedthattheluxunitisdefinedastheunitofilluminance,andsoshouldnotbeusedforexitance.Actually,keepingtheseunitsdistinctforincidentandexitingfluxhelpstoavoidconfusion.) Now consider luminance. This is not a simple concept. As simply as I canexpressit,itistheluminousfluxduetoasmallelementinagivendirection,relativetothe

area of the element projected in that direction and the solid angle subtending the flux,measured in candelas per squaremetre (cd/m2). It needs to be recognised that there aretimeswhenit isnecessarytousethe luminancemetric,asforvisualtaskanalysiswherethecontrastofthecriticaldetailhastobedefined,buttorefertotheaverageluminanceofawalloraceilingreallyismeaninglesswithoutadefinedviewpoint.Afterall,whatistheaverage projected area of one of these elements?Readerswho are not familiarwith theexitancetermarestronglyadvisedtomakethemselvesacquaintedwithit.Notonlyisitamuchmoresimpleconceptthanluminance,butwhenweareconcernedhowilluminationaffects the appearance of room surfaces, it is the correct term to use. Seen in thisway,MRSEisthemeasureoftheoveralldensityofinter-reflectedlightwithinthevolumeofanenclosedspace.

Applyingtheambientilluminationconceptindesign

Roomsurfacereflectancesaresoinfluentialuponboththeappearanceofindoorspacesandthe distribution of illumination within them that, in an ideal world, lighting designerswould take control of them. The reality is that generally someone elsewillmake thosedecisions,butlightingdesignersmustpersistinmakingthesedecisionmakersawareoftheinfluence they exert over ambient illumination and the overall appearance of theilluminatedspace.

The creativityof a lightingdesigner is largelydeterminedby the ability toperceive aspaceand itsobjects in light, and aswehave seen, the perceived light is reflected (notdirect)light.Aroominwhichhighreflectancesurfacesfaceotherhighreflectancesurfacesisoneinwhichinter-reflectedfluxpersists,anditisthisinter-reflectedfluxthatprovidesforoursenseofhowbrightlyordimlylitthespaceappears.

To initiate this inter-reflected flux,direct light,which travels fromsource to receivingsurfacewithoutvisibleeffect,has tobeapplied.Theessential skillofa lightingdesignermay be seen as the ability to devise an invisible distribution of direct flux that willgenerateanenvisageddistributionofreflectedflux.

Large, high reflectance surfaces enable the direct light to be applied efficiently andunobtrusively, andwhere highMRSE levels are to be provided, the availability of large,light-colouredsurfacesthatcanbewashedwithlightbecomesanimportantconsiderationforbothappearanceandenergyefficiency.Conversely,wheretheaimistokeepMRSElow,perhaps to provide high contrasts for display lighting, dark-coloured room surfacesreinforce the visual effect by absorbing both spill light (display lighting thatmisses thedisplay)and‘firstbouncelumens’reflectedfromthedisplays.

Figure2.7ReflectanceplottedagainstMunsellValue,whereasurfaceofMV0wouldbeassessedasaperfectblackand

MV10asaperfectwhite.PerceptuallyMV5ismid-waybetweentheseextremesandmightbeexpectedtohavea

reflectanceof0.5,butactually,ithasareflectanceofapproximately0.2.

Estimating surface reflectance values is not straightforward. TheMunsell Value (MV)scaleorders surface colourson a 10-step scale according to lightness assessments,whereMV0appearstobeaperfectblack,andMV10aperfectwhite.Unlikereflectance,lightnessisasubjectivescale,andwhileitrelatestoreflectance,therelationshipisfarfromlinear.Avalue of MV5 is perceptually mid-way between black and white and so it might beexpected to have a reflectance around 0.5, but as Figure 2.7 shows, its actual value isapproximately0.2.Furthermore,itcanbeseenthatasurfacehavingareflectanceof0.5hasaMVofapproximately7.5,andthatputsitperceptuallythree-quartersofthewaytowardsperfect white. The practical implication of this pronounced non-linearity is thatinexperienced designers are inclined to substantially overestimate reflectance values. Areasonably reliable procedure is to fit an internally blackened tube over an illuminancemeterasshowninFigure2.8andtotaketworeadings,oneforthesurface,RSandoneforasheetofgoodqualitywhitepaperwhichhasbeenslidoverthesurface,RP.Itisreasonableto assume that the paper has a reflectance of 0.9, so that for a measure of surfacereflectance,ρS=0.9RS/RP.Patternedaswellasplainsurfacescanbedealtwithinthisway,but care needs to be taken to avoid specular reflections, particularly for glossy surfaces.Also, it shouldnotbe assumed that shiny surfaceshavehigh reflectance.These surfaces

simplyreflectwithoutdiffusion,sothatifthemeterisexposedtospecularreflection,whatisbeingmeasured isan imageofa lightsourcerather thantheoverallreflectionof lightfromthesurface.

Figure2.8Usinganinternallyblackenedtubemountedontoalightmetertoobtainameasurementofsurfacereflectance.

Twomeasurementsaremadewithoutmovingthemeter,oneofthesurfaceasshown,andacomparisonreadingwitha

sheetofwhitepaperinthemeasurementzone.

Theeffectsof this tendencytooverestimatereflectancevaluesarecompoundedbytheimpactofsurfacereflectancevaluesonMRSE.ItcanbeseenfromFormula2.1thatMRSEisproportionaltotheratioofroomsurfacereflectancetoabsorptance,ρ/α.Figure2.9plotsthevalueof this ratiorelative toreflectance,andagain itcanbeseen that the impactofroomsurfacereflectanceincreasesexponentiallywithreflectance,andcouldleadtogrosslyinflatedMRSEvaluesbeingpredictedwhere reflectancevalueshavebeenoverestimated.Wecanseeheretheeffectsofreflectancethatwereobservedinthethoughtexperiment,andwhile these effects are real, theywill not be realizedunless reflectance values havebeenaccuratelyassessed.

Theseconsiderationssuggestaninitialsequenceforapplyingtheseconcepts:

1. Decide upon the level of MRSE, taking account of design considerationsconcerning the perceived brightness or dimness of ambient illumination, andreferring toTable2.1 and the discussion in the section entitled ‘The amount of

light’.2. Calculatetheroomabsorption,Aα,referringtoFormula2.4.3. Determine the levelof first reflected flux, turningFormula2.2around toFRF=

MRSE×Aα4. DetermineadistributionofdirectfluxtoprovidetheFRFvalue.Atthispoint,we

cometoacentraldesignissue:howtodistributethedirectflux,Fs(d),orinotherwords, how to choose the surfaces ontowhich fluxwill be directed.To explainthisissuewewillconsidertwocases.

Figure2.9Thevalueofthereflectance/absorptanceratioisproportionaltomeanroomsurfaceexitance,MRSE.Notehow

valuesincreaseexponentiallyathigherreflectancevalues.

Throughout thisbookwewillbemakinguseof spreadsheets to facilitate calculations,and their outputs are shown in the Boxes alongside the text. Readers are stronglyencouragedto followthe instructions fordownloadingthespreadsheetsso theycanthenfollow the applications described. Boxes 2.1 and 2.2 show two outputs of the AmbientIlluminationSpreadsheet,buttherealbenefitofdoingcalculationsinthiswayisnotthatitallhappenssoquicklyandeasily(althoughthatundoubtedlyisabenefit)butthat,onceasituation has been set up, the user is able to explore alternative solutions with instantfeedback.Readersarestronglyencouragedtofollowtheseexamples,andthentogobeyondthembyasking,‘Whatif…?’

Box2.1

AmbientIllumination

140117

Project Case1MRSE 150lm/m2

RoomDimensions

Length Width Height12 9 3m

RoomabsorptionAα 160.2m2

Firstreflectedflux(FRF) 24030lmTotalluminaireflux(F) 32583lm

Key

Notes

Enterdataonlyincellsshowninred–allotherdataarecalculatedautomatically.

DirectFlux(%)isthedirectfluxincidentonSasapercentageoftotalluminousflux.

Envisage an indoor space measuring 12m long, 9m wide, and 3m high. To keep lifesimple,wewill not get too specific about the function of this room. ForCase 1wewillwork on the basis that the aim is to provide a fairly bright overall appearance, whereeverythingappearsadequatelylitbutnoobjectsaretobeselectedforparticularattention,andwhatiscalledforisawell-diffused,overallillumination.Decisionshavebeenmadeforsurfacefinishes,andithasbeenagreedthatceilingreflectance,ρclg, istobe0.85,ρwall tohave a value of 0.5, and ρflr will be 0.25, and Box 2.1 shows the dimensions and thereflectancesenteredontheAmbientIlluminationSpreadsheet.

After giving due consideration to the points discussed in ‘The amount of light’, wedecideuponaMRSElevelof150lm/m2.Thisvalue isenteredonthespreadsheet,notingthatdataaretobeenteredonlyintocellsmarkedinred.Tofullyunderstandtheprocedure,thereaderisadvisedtocheckthecalculationonpaperusingtheaforementionedformulae.

TheFRFvalueshowninBox2.1isthenumberoflumensreflectedfromalloftheroomsurfacesrequiredtoprovidethemoderatelybrightoverallappearancethatwehavesetasour goal. Nowwe address the first really important design issue: how to distribute thedirect flux? The aim is to achieve a well-diffused illumination, and to do this withoutcreatingdistinctlybrightzonessuggestsalightinginstallationthatdistributesilluminationevenly over large surfaces. The only remaining red values are in the Direct Flux (%)column, and this is the column where the designer experiments with direct fluxdistributions.Twovalueshavebeenentered:15percentoftotalluminaireoutputistobedirectedonto thewalls,and10percentonto the floor.Asnoobjectshavebeenentered,that leaves 75 per cent onto the ceiling. The next column, Fs(d), shows the number oflumens of direct flux required on each room surface; next, the Es column shows theilluminance(includingindirectflux)oneachsurface;andinthefinalcolumn,theratiosofsurfaceilluminancetoambientilluminance,Es/MRSE.BelowthesecolumnsarethevaluesofAα,FRFandthetotalluminousflux,F,tobeemittedbytheluminaires.

Ways of predicting luminaire layouts for direct light distributions are explained inChapter6,butbefore that, this spreadsheetgives thedesigneropportunity toexplore theimplicationsof fluxdistribution.To experience this, download theAmbient IlluminationspreadsheetandclicktheBox2.1 tag.Trychanging thewallsandfloor fluxpercentages,andifyoulike,youcanaddafewobjects,suchasfurnitureitems.Youwillseethateverytimeyouaddmoreroomabsorptionordirectmorefluxontosurfacesoflowerreflectance,upgoes the luminaire flux.Foroptimumenergyefficiency, set thewallsandfloordirectfluxpercentagestozerosothatthedirectceilingfluxbecomes100percent,andyouwillseetheluminairefluxdroptojustover28,000lm.ThiswouldbethemostenergyefficientsolutionforachievingtheMRSEtargetinthislocation,butwhenthishappens,thevalueofEs/MRSEclimbsto2.7,andthismaybeacauseforconcern.

If the aim is to achieve the ambient illumination without any surface appearingnoticeablymorestronglylitthananyothersurface,thenasindicatedinTable2.2,theaimshouldbetokeepvaluesofEs/MRSEbelow1.5.Avalueof2.7fortheceilingindicatesthatthissurfacewillappeardistinctlymorestronglylitthananyothersurfaceorobjectinthisspace,andinfact,forthecaseshowninBox2.1,wheresomefluxisdirectedontothewallsandfloor,theEs/MRSEvalueisonlyslightlyreducedto2.5,sotheappearanceofthedirectilluminationontotheceilingwouldcertainlybe‘noticeable’,evenifnot‘distinct’.Wecouldtryadjustingthepercentagevaluesonthespreadsheettoachievealesspronouncedeffect,butwatchthevalueofthetotalluminaireflux,F.Asmoreluminairefluxisdirectedontolower reflectance surfaces, so the flux required to provide the MRSE value goes up. Itshouldnotpassnotice that this flies in the faceof conventionalpractice.Allaround theworld, lightingstandardsforilluminationsufficiencyforindooractivitiesarespecifiedintermsofilluminanceappliedontothehorizontalworkingplane,fromwhichitfollowsthat‘efficient’ lighting takes the form of a grid layout of luminaires that directs its outputdirectly onto that plane. While it is widely acknowledged that indirect ceiling lightinginstallationscanachievepleasanteffects,thewaythestandardsarespecifiedcausesthemtobeclassifiedasinefficient.Whenadesignerissatisfiedthatasatisfactorydistributionofdirect fluxhasbeenachieved,acopyof thespreadsheetwouldbesavedonto thedesignprojectfile.

NowturnattentiontoCase2,forwhichwehaveaquitedifferentaim.Again,wewillnotgettoospecificaboutthesituation,butthistimetheaimisthatafewselectedobjectsaretobepresentedfordisplay,andthesearetobecomethe‘targets’forthelightingwiththeintentionthattheywillcatchattentionbyappearingbrightlylitinadimsetting.TherevisedoutputfortheAmbientIlluminationspreadsheetisshowninBox2.2,anditshowsthatmostofthedirectfluxistobedirectedontothesetargets.Evenso,thisisaspacethatpeoplewouldneedtobeabletofindtheirwaythrough,soabackgroundofinkyblacknesswouldnotbeacceptable.Thisbringsusface-to-facewithatrickydesigndecision.Ononehand we aim to achieve a luminous environment that is dark enough to provide foreffectivedisplaycontrasts,whileontheotherhanditneedstobelightenoughforpeopletofindtheirwaythroughsafely,and,atleastasimportant,weneedtocreateanentrytothe space that people findwelcoming.We should keep inmind that in order to attractpeopletoenterthisdimspace,atleastpartofthedisplayedmaterialshouldbepositionedsothatitisvisibletosomeoneapproachingtheentrancetothespace.

AsshowninBox2.2,wehaveoptedforaMRSElevelof10lm/m2,andatthisstageweenterintodiscussionwiththedesignteam.Itisagreedthatbothρclgandρflraretobekeptdowntoalevelof0.15,althoughtoprovideaslightlylighterbackgroundtothedisplays,awall finishwith a reflectancevalueof 0.25 is chosen.Thedisplayedobjectshave a totalsurface area of 20m2 with an average reflectance of 0.35, but it would be unrealistic tosupposethatwewillbeabletodirect100percentoftheluminairefluxontothem.Ithasbeenassumed that therewillbe10percent spill light,halfof itonto thewallsandhalf

ontothe floor,andbasedonall these inputs, thespreadsheetshowsthatweneeda totalluminaire fluxof8690 lumens.That luminousflux,appropriatelydirected,willprovideadisplay illuminanceof401 lux,and, referringagain toTable2.2, thevisual effectwill be‘emphatic’, as itwill provide a Es/MRSE value of 40. Note that in order to achieve thisdramatic effectwedidnot start by setting the target illuminance, but rather,we set theambientilluminanceandthendeterminedthefluxdistribution.Toprovideahigherleveloftarget illuminance would have the effect of raising the ambient illumination above thedesignvaluewithoutaddingtotheEs/MRSEratio.

Fromthese twocases itcanbeseen that inorder for lighting toexert itspotential forinfluencing the appearance of everythingwe see, control over room surface reflectancevaluesisasimportantasbeingabletocontroldirectfluxdistributions.Betweenthesetwoquiteextremecases,manyoptionsexistfordesignerstocontrolambientilluminationlevelto support chosen lighting design objectives. TheAmbient Illumination Spreadsheet is ausefultoolforachievingthiscontrol.

Box2.2

AmbientIlluminationSpreadsheet

140117

Project Case2MRSE 10lm/m2

RoomDimensions

Length Width Height12 9 3m

RoomabsorptionAα 291.1m2

Firstreflectedflux(FRF) 2911lmTotalluminaireflux(F) 8690lm

Key

Notes

Enterdataonlyincellsshowninred–allotherdataarecalculatedautomatically.DirectFlux(%)isthedirectfluxincidentonSasapercentageoftotalluminousflux.

3IlluminationHierarchies

Chaptersummary

Whereambientilluminationissufficientforilluminanceandlightness(whichisrelatedtoreflectance)tobeperceivedseparately,astypicallyoccursforconventionalindoorlightingpractice, lighting may be planned in terms of illuminance (rather than luminance)distributions. Local concentrations of illumination can be applied to direct attention, togiveemphasisandidentifyobjectsthatthedesignerdeemstobevisuallysignificant.Thenotion of ordered distributions of illumination leads to the concept of illuminationhierarchy,wherebyilluminationdistributionsarestructuredasaprincipalmeansbywhichthedesignermay express his or her design intentions. Suchdistributions are planned aschanging balances of direct and indirect illumination, and are achieved by specifyingtarget/ambient illuminance ratio (TAIR) values. The Illumination Hierarchyspreadsheetfacilitatesapplicationofthisconcept.

Orderedilluminationdistributions

Mostformsoflifeareattractedtowardslight,andhumansarenoexception.Phototropismistheprocessbywhichattentionisdrawntowardthebrightestpartofthefieldofview.Itcan be detrimental, aswhen a glare source creates a conflict between itself andwhat apersonwants tosee,and ingeneral lightingpracticemuchattention isgiventoavoidingsuch effects. However, for lighting designers it is a powerful tool, enabling us to drawattentiontowhatwewantpeopletonoticeandawayfromthingsofsecondaryortertiarysignificance.Anorderedilluminationdistributionistheunderpinningbasisforstructuringalightingdesignconcept.

Itisimportanttospendsometimelookingcarefullyathowourperceptionsofspaceandobjectsareinfluencedbyselectiveillumination.ItwasnotedinChapter1thatcoloursthatmake up an overall scene are generally perceived as related colours, and as long asillumination is sufficient to ensure photopic adaptation, we have no difficulty inrecognisingallthesurroundingsurfacesandobjectsthatmakeupourenvironments.Theprocess of recognising the multitude of ‘things’ that may, at any time, comprise oursurroundingsfallswithinthetopicofperceptualpsychology,butwithoutgettinginvolvedin that fieldof learning it is sufficienthere toacknowledge that this recognitionprocessinvolves discriminating differences of object attributes such as lightness, hue andsaturation,fromwhichweformperceptionsofspaces,people,andobjects.Weachievethiswithout conscious effort throughout our waking hours over a very wide range of‘adequate’lightingconditions.Inthiscontext,theonsetofdimnessmaybethoughtofastheborderlineofreliablerecognitionofobjectattributes.

However,withorderedilluminationdistributionswecangobeyondsimplyprovidingforobjectrecognition.Retailerslongagoworkedoutthatifanobjectthatissmallinrelationtoitssurroundingsreceivesselectiveillumination,particularlywithoutthesourceoflightbeingevident,people’sperceptionsofthatobject’sattributescanbesignificantlyaffected.Whether or not it appears more brightly lit, it is likely to appear more colourful, andperhapsmoretexturedormoreglossy,thanitwouldappearwithoutselectiveillumination.Lighting designers have at their disposal the means to establish hierarchies of visualsignificance in illuminatedscenes,andmeans forachieving this inanorderedmanner isthecontentofthischapter.

Illuminanceratios

Whenweplaceanattractiveobject,suchasavaseofflowers,besideawindowto‘catchthe light’,weare exploiting thepotential forapoolof local illumination to identify thisobjectashavingbeenselectedforspecialattention.Similarly,electriclightingcanprovidea planned gradation of illumination that expresses the designer’s concept of layers ofdifference. Hard-edged contrasts can give emphasis to such effects, but alternatively, adifferent but equally striking effect may be achieved by a build-up of illuminance thatleads the eye progressively towards the designer’s objective. High drama requires thatsurroundings are cast into gloom, but in architectural situations, safety requirementsgenerally require surroundings to remain visible, although perhaps distinctly dim, at alltimes. Planning such distributions is more than simply selecting a few objects forspotlighting. It involves devising an ordered distribution of lighting to achieve anilluminationhierarchy.

The concept of a structured illumination distributionwas pioneered by J.M.Waldram(1954).Working froma perspective sketch of the location, hewould assign an ‘apparentbrightness’valuetoeachsignificantelementoftheview,andthenhewouldconvertthosesubjective values into luminance values so that he could apply illumination engineeringprocedures to determine a suitable flux distribution. Waldram’s notion of creating anordered brightness distribution related to luminance would seem to be valid for lowadaptation situations, such as occur in outdoor lighting, but not for situations wheresurfacelightnessisreadilyrecognised,suchasinadequatelyilluminatedindoorscenes.Ashas been noted, for these situations our perceptions distinguish illumination differencesmoreorlessindependentlyofsurfacereflectancevalues.

J.A.Lynes(1987)hasproposedadesignapproachbasedonWaldram’smethodwiththedifference that the designer develops a structured distribution of surface illuminancevalues. Lynes introduces his students to the topic through an exercise in perceiveddifferenceofillumination,andhissimpleprocedureisillustratedinFigure3.1.Hestandsinfrontofhisclasswithaspotlightshiningontoawhitescreen.Point0isthebrightestspot,andfromthispointanumberedscaleextendsacrossthescreen.Eachstudentcompletesascore card, and starts by indicating the scale value that, in his or her assessment,corresponds to the point along the scale atwhich a ‘noticeable difference of brightness’occurs.Thisisthestudent’s‘N’value,andwouldbefollowedbya‘D’valueforadistinctdifference,an‘S’valueforastrongdifferenceandan‘E’valueforanemphaticdifference.Thecardsare thengathered, averagevalues calculatedandconsensusvalues forN,D,SandEaremarkedonthescreen.Afterthat,Lynesmeasurestheilluminancelevelateachpoint,fromwhichilluminanceratiosarecalculatedforeachperceiveddifference.

Theauthorhasconducted thisexercisewith studentsonnumerousoccasions.Perhapsthefirstsurpriseistofindhoweasyitistoobtainconsensus,andthesecondishowwell

theresultsarerepeatedyearafteryear.ThedatapresentedinTable2.2istypical,andwhilethissimpleproceduremaynotqualifyas‘goodscience’,itiswellworthgoingthroughtheprocedure.Itcallsforthoughtfulobservation,and,perhapssurprisingly,itprovidesusefulguidanceforlightingdesign.Notonlystudents,butanyoneinterestedindesigninglightingshouldgothroughtheprocessofmakingtheseilluminationdifferenceassessmentsatleastonceduringtheirlifetime.

Figure3.1Demonstrationset-upforgainingassessmentsofnoticeable,distinct,strongandemphaticillumination

differences.

WhereasinChapter2wediscussedhowinitialresponsestoaspacemaybeinfluencedbyambientillumination,nowweturnattentiontotheperceivedeffectsthatcanbecreatedbycontrollingthedistributionofilluminationwithinaspace.FromTable2.2itcanbeseenthatwheretheaimistoachieveadifferencethatissufficienttobenoticed,youcanforgetabout10or20percentdifferences.Unlessadifferenceofatleast1.5:1isprovided,peoplewill not notice the illumination to be anything different from uniform. To achievedifferences that are likely to be described as ‘distinct’ or ‘strong’, it is necessary for thedesigner to be purposeful and deliberate in how they achieve such pronounced visualeffects. Illumination distributions will have to be carefully controlled and, preferably,surrounding reflectances kept low. An ‘emphatic’ difference is quite difficult to achieveotherthaninatheatreorsimilarsetting,andaswasnotedtowardstheendofChapter2,raisingthetargetilluminationunavoidablyraisestheambientillumination.Wheretheaim

is to achieve high illuminance differences, target objects need to be small in relation totheir surrounding space, ormore specifically, to the roomabsorptionof the surroundingspace.

Wewill return to this lastpoint,butbeforemovingon, let itbe repeated thatmakingassessments of the appearance of illumination differences is a revealing exercise inobservation. Actually doing it, and measuring one’s own assessments of perceiveddifference, is instructive. Then following up with observation and measurement in reallocationsisenormouslyvaluable.Themetertellsyounothingusefuluntilyouhaverelateditsreadingstoyourownexperience.ThedatainTable2.2istypical,butadesignerneedstobeabletovisualisetheseilluminanceratios.Itisbyhavinginmindtheperceivedeffectofilluminance ratios that adesigner is able to specifyvalues that reflectobservation-basedexperience.

Target/ambientilluminanceratios

While theperceivedadequacyof illumination (PAI) criterion is concernedwithensuringadequateinter-reflectedflux(MRSE)withinaspace,theilluminationhierarchycriterionisconcernedwithhowthedirectfluxfromtheluminairesmaybedistributedtocreateanordered pattern of illumination that supports selected lighting design objectives, whichmay range from directing attention to the functional activities of the space to creatingaesthetic or artistic effects. For all of this, we make use of the target/ambientilluminance ratio, TAIR, where target illuminance is the sum of direct and indirectcomponents, and TAIR relates target illuminance to the ambient illumination level. Thedesigner selects target surfaces anddesignatesvalues according to the level of perceiveddifference of illumination brightness to be achieved both between room surfaces, andbetweenobjectsandthesurroundingsagainstwhichtheyareseen.Asthepointhasbeenmade that illumination isnotvisibleuntil ithasundergone its first reflection, itmaybewondered why we are now dealing with incident target illumination, which comprisesboth direct and indirect illumination. The answer is that as both components undergoreflection at the same surface, it makes no difference whether we take the ratio of theincidentorreflectedvalues.

MRSEprovides themeasureofambient illuminationwithinaspace,andexceptwherethere are obvious reasons to the contrary, it is reasonable to assume that the incidentilluminationoneachtargetsurfacetgtwillbethesumofdirectilluminanceandMRSE,sothetotalilluminanceonatargetsurface:

andthetarget/ambientilluminanceratio:

The TAIR concept provides a basis for planning a distribution of direct flux from theluminaires that will achieve an envisioned illumination distribution within a space. Itfollowsthatforanychosentargetsurface,thedirectilluminance:

DesigninganilluminationhierarchyinvolvesdesignatingTAIRvaluesforselectedsurfacesor objects to signal noticeable, distinct, or strong perceived differences of illumination,againreferringbacktoTable2.2,andtherereallyisnolimittothesituationsforwhichthisproceduremaybeapplied.Adesignermaychoosetotargetasubstantialproportionofthetotal roomsurfacearea, andexamplesof thiswould include lightingamural coveringawholewall,oranarchitectural icon,ora libraryreadingarea,orperhaps, thehorizontalworkplaneofan industrialassemblyshop.Alternatively, thetargetareamaybeasingleobject that comprises a small proportion of the total surface area, such as a solitary

sculpture,ora featuredretaildisplay,or thepreacher inhispulpit;or itmaycompriseanumber of even smaller items, such as display of coins, or individually lit items ofglassware.Whateverthesituation,thedesignerfirstneedstodecideupontheMRSElevelto achieve the required ambient illumination for the space, and then to decideupon theTAIR for each target surface for the differences of illumination brightness. This enablesFormula3.3tobeappliedtodrawupthedistributionofdirecttargetilluminancevalues.

Thisputsthedesignerinthepositionofbeingabletodeterminethedistributionofdirectlighttobeappliedthroughoutthespaceinordertoachievetheenvisioneddistributionofreflected light. The total indirect flux provided by first reflections from all surfacesreceivingselectivetargetlighting:

Note that the suffix tgt indicates an individual target surface, and ts refers to all targetsurfaces within the space. This value of Fts(i) indicates the extent to which all of theselectivetargetlightingwillcontributetowardsthefirstreflectedfluxrequiredtoachievetheambient illuminationMRSE.Theusefulnessof this formulabecomesapparent in thefollowingsection.

Itmaybenotedinpassingthat,unlikeMRSE,TAIRisnotproposedasasuitablemetricfor lighting standards. TAIR is a tool that enables pursuit of chosen lighting designobjectives,whichmayrangefromverysimplethroughtodistinctlycomplexinnature,anditsapplicationinvolvesobjectivesthatarebeyondthescopeofstandards,whetheradvisoryormandatory.

Illuminationhierarchydesignprocedure

Withoutwishing togive the impression that creative lightingdesigncanbeachievedbyfollowingastep-by-stepprocedure,theconceptspreviouslydescribedimplyasequenceforlogical decisionmaking. The flowchart shown in Figure 3.2 should be referred towhilefollowingthisprocedure.

Figure3.2Flowchartforachievingmeanroomsurfaceexitance,MRSE,andtask/ambientillumination,TAIR,design

values.

1. For a design location, consider a level of MRSE that would provide for anappropriate appearance of overall brightness or dimness. Codes or standardsspecified in taskplane illuminance areunlikely tobehelpful. Should therebe apublishedMRSEvaluerelevanttothelocation,itprobablyrelatestotheperceivedadequacy of illumination (PAI) criterion and specifies the minimum value ofMRSE to be provided. Consider whether a higher level to give a brighterappearancewouldbeappropriate,referringtoTable2.1 forguidance,and takingintoaccounttheimmediatelypreviousbrightnessexperienceofapersonenteringthisspace.Considerwhetheritistoappearbrighterordimmerthanthepreviousspace,andifso,byhowmuch,thistimereferringtoTable2.2forguidance.Where

nominimumlevelsarespecified,designingforanappearanceofdimnessbecomesanoptionprovidingsafetyconcernsarekeptinmind.

2. Decide upon the design value ofMRSE, this being the overall density of inter-reflectedfluxtobeprovidedwithinthevolumeofthespace,andenterthisvalueinto the Illumination Hierarchy spreadsheet (see Box 3.1, and use your owndownloadedcopyofthespreadsheet).

3. Estimate theareaandreflectancevalue foreachsignificantsurfaceSwithin theroom,making sure to include any surfaces or objects that youmight decide tohighlight with selective lighting, and enter these onto the spreadsheet. Thespreadsheet calculates the room absorption value, Aα(rs), and the total firstreflectedflux,FRFrs,requiredtoprovidetheMRSEvalue.

4. Considertheilluminationhierarchythatthelightdistributionistocreateinthisspace. Think about which objects or surface areas you want to highlight withselectivelighting,andbyhowmuch.Youwillprovidedirectlightontothesetargetsurfaces, while surrounding areas will be lit mainly, or perhaps entirely, byreflectedlight.

5. EnteryourdesignvalueofTAIRforeachtargetarea,takingaccountofhowtheappearanceof theselectedobjectsorsurfaceswillbeaffectedbylocaliseddirectillumination. This listing of TAIR in Column 5 of the spreadsheet becomes therecordofyourilluminationhierarchyforthespace.

6. The spreadsheet completes the calculations, giving the first reflected flux to beprovidedbylightreflectedfromthetargets,FRFts,andthedifferencebetweenthisvalueandthetotalFRFrequiredtoprovidetheMRSEvalue,FRFrs–FRFts.

Then:

If the first reflected flux from the targets is less than the total first reflected fluxrequired,thatistosay,ifFRFts<FRFrs,theninadditiontothelightdirectedontothetargetareas,thesurroundingroomsurfaceswillneedsomedirectilluminationto make up for the difference, FRFrs – FRFts. This is needed to ensure that theMRSEdesignvaluewillbeachieved.Thedirectilluminationontotheroomsurfacesdoesnotneedtobeapplieduniformly,andoftenthemosteffectivewaywillbetospread light over large,high-reflectance surrounding surfaces suchas ceiling andwalls. Concentrating this light onto small areas may cause them to competevisuallywiththetargetareas,ashasbeendiscussedinChapter2.Thereisplentyofscopeforingenuityindevisingwaysofraisingtheoverallilluminationbrightnesswithoutdetractingfromtheselectedtargets.IfFRFts≈FRFrs,thetargetilluminationalonewillprovideforthedesignvaluesforbothMRSEandTAIR.Thisisbecausereflectedlightfromthetargetsurfaceswillboth provide the design level of ambient illumination and achieve the intendedbalanceoftarget/ambientlevels.Aserendipitousoutcome.

If FRFts > FRFrs, the proposed balance of MRSE and TAIR values cannot beachieved in this situation. The reason is that if the direct target illuminance isapplied,thereflectedfluxwillraiseMRSEabovethedesignlevel,andreduceTAIRvaluesbelowthedesignlevels.Usuallythemosteffectiveremedialactionwillbetoreducethe total targetarea,suchasbyconcentratingtheobjects toreceivedirectlightintomorerestrictedareas.Otherwise,itwillbenecessarytoreduceeither,orboth,ρtsandρrs,butunfortunately,lightingdesignersseldomhavemuchinfluenceoverreflectancevalues.Acompromisemaybeinevitable,butatleasttheoutcomewillnotcomeasanunwelcomesurprise.

Example:abankingpremises

Box3.1showsaworksheetfromtheIlluminationHierarchyspreadsheet,andagain,readersarestronglyrecommendedtoexperiencetheuseofthesedesigntools.Roomsurfacedatahavebeenenteredforabankingpremises, so takeamoment to familiariseyourselfwiththelocation.

A bright and business-like appearance is wanted, and a MRSE level of 200 lm/m2 isproposed.Thisvaluehasbeenentered,andaspreviously,datashowninredareinputbytheuserandallothervaluesare calculatedautomatically.Column4gives thecomputedroom absorption values, and the bottom line shows that 39,096 lumens of first reflectedfluxfromtheroomsurfacesisrequiredtoprovidetheMRSElevel.NextthedesignerentersaTAIRvalueforselectedtargetsurfaces.This is thevitalcomponentof thisstageof thedesign process, and Column 5 forms the statement of the designer’s initial intent forilluminationhierarchy.Atthebottomofthefinalcolumnitisshownthat20,899lmoftherequiredFRFwillbeprovidedfromthetargetsurfaces,sothatthedifferenceof18,197lmwillneedtobemadeupbyapplyingadditionaldirectlightontoroomsurfaces.

This is the information that thedesignerneeds todetermine thebalanceofdirectandindirectillumination.VariousoptionsforprovidingthedeficitFRFmaycometomind,butasimpleandefficientsolutionwouldbeuplighting.Therequireddirectceilingilluminanceis:

ThisdirectilluminanceaddedtotheMRSEvalueof200lm/m2wouldgiveatotalceilingilluminanceEclgof414lux,givingaTAIRvalueofjustovertwo.Table2.2 indicates thatthiswould correspond to a perceived difference thatwould appear somewhere betweennoticeableanddistinct,andsowouldcreateavisibleeffect thatmightcompetewith theplanned distribution of TAIR values. This effect could be reduced by applying lessilluminationontotheceilingandmakingupforthedeficiencybyaddingsomedirectlightontoothersurfaces,particularlythewalls.

Box3.1

IlluminationHierarchySpreadsheet

Date:140119

Symbols

It is at this point that the attraction of using the spreadsheet becomes evident. Bytreatingselectedroomsurfacesastargets,alternativestrategiesmaybereadilyexamined.Asthewallsurfaceshavelowerreflectancevaluesthantheceiling,itwilltakemoredirectlumenstobringtheFRFrsvalueup to therequired level,but the light-colouredblinds inwalls 3 and 4 could receive selective wallwashing, and this might create an attractiveappearance. However, the effectiveness of this solution would depend upon the staffpulling down the blinds during hours of darkness. It would be necessary to enquire

whether this could be relied upon, and after all, this is the way that lighting designhappens. It is part of the reason why no two designers would come up with identicalschemes.

Box3.2

IlluminationHierarchySpreadsheet

Date:140119

Symbols

Box3.2showsadesignproposal.TheTAIRvalues inColumn5havebeenadjustedtoprovidevariouslevelsofunnoticeable,noticeable,distinctandstrongperceiveddifferences,and by addingmore target surfaces in this way, the FRFrs – FRFts difference has beenreducedtoanegligiblevalue.Thismeansthatthefirstreflectedfluxfromthetargetswillprovidetherequired200lm/m2ofmeanroomsurfaceexitance,andwiththeexceptionoftheblinds,thevisibleeffectofthisadditionalilluminationwillnotbebrightenoughtobenoticed. In thisway, theoriginaldesign intentwillbemaintained. It canbe seennotallsurfacesaretoreceivedirectlight.

Column6showsthedirectilluminancetobeprovidedontoeachtargetsurface.Allthatisleftnowistoapplysomestraightforwardilluminationengineering,andproceduresfordetermining luminaire layouts to distribute direct flux to achieve specific illuminancevaluesareexplainedinChapter6.

References

Lynes,J.A.(1987).Patternsoflightandshade.LightinginAustralia,7(4):16–20.

Waldram, J.M. (1954). Studies in interior lighting. Transactions of the IlluminatingEngineeringSociety(London);19:95–133.

4SpectralIlluminationDistributions

Chaptersummary

Variouswaysinwhichhumanperceptionofalitspaceisinfluencedbythespectralpowerdistribution(SPD)ofilluminationarereviewed.Distinctionismadebetweenassessmentoflightforvisibilityandforbrightness,andalternativeresponsefunctionsforindoorspacesare examined. The effects of SPD upon the perception of illumination colour (colourappearance) and colouredmaterials (colour rendering) are examined, alongwith variousproposalsforidentifyinghowbothSPDandilluminationlevelinfluencetheappearanceoflit spaces. These include perceived attributes of illumination, such as the whiteness,naturalness and colourfulness of illumination, as well as some non-visual effects. It isconcluded thatpeoplehavedifferentdaytimeandnight timeexpectationsandneeds forlighting.

Luminoussensitivityfunctions

Before 1924, the onlyway ofmeasuring lightwas tomake comparisonswith a familiarlightsource,whichledtometricssuchasthecandlepowerandthefootcandle,butinthatyear the CIE (International Commission on Illumination) introduced the V(λ) luminoussensitivity function which defines the relative visual response, V, as a function of thewavelength of radiant power, λ, as shown in Figure 4.1. This was a significantbreakthrough that required innovative research, and it enabled luminous flux, F, to bedefinedintermsoflumensfromameasurementofspectralpowerdistribution:

where:

P(λ)=spectralpower,inwatts,ofthesourceatthewavelengthλV(λ)=photopicluminousefficiencyfunctionvalueatλΔλ=intervaloverwhichthevaluesofspectralpowerweremeasured

ItcanbeseenfromFigure4.1thatV(λ)hasitsmaximumvalueof1.0at555nm,andsotheluminousefficiencyofradiantfluxatthiswavelengthisequaltothevalueoftheconstantin Formula 4.1, 683 lm/W.At 610nm,where the value of V(λ) is approximately 0.5, theluminousefficiencyreducestohalfthatvalue.

SobydefiningtheV(λ)function,theCIEmadeitpossiblefortheoutputofalightsourceto be specified in terms of the lumen,while at the same time enabling light itself to bedefinedintermsofradiantpowerwithinthewaveband380–780nanometres(nm).Tothisday,lightingstandardsandrecommendedpracticedocuments,aswellasthecalibrationofall lightmeters, are based on V(λ), and in fact, it continues to be quite appropriate formeasuringilluminationinsituationswherephotopically-adaptedviewersarefixatinguponvisual tasks.Examplesrangefroma libraryreadingroomtoahospitaloperatingtheatre,and for these, as well as for most task-based applications in between, this luminoussensitivityfunctioncontinuestoserveuswell.Thereis,however,moretohumanresponsetolightthanthis,andfordesignerstobeabletoapplylightingknowinglyandeffectivelyintherangeofsituationsencounteredingeneral lightingpractice,wecouldbenefitfrommetricsthattakeaccountofawiderrangeofhumaninteractionswithradiantflux.

Figure4.1RelativesensitivityfunctionsforV(λ),andthethreeconetypes;long-,medium-andshort-wavelength;L(λ),

M(λ)andS(λ).ItcanbeseenhowcloselyV(λ)representstheresponsesoftheLandMcones,andignorestheScone

response.

Formula4.1assumesahumanobserveroperatingwithin the rangeofphotopicvision,and thismeans that error is incurredwhenever V(λ) is applied formesopic or scotopicconditions. Also, the researchers who established the V(λ) function had their subjectsobservingaquitesmallluminouspatchthatsubtendedjust2degreesattheeye,sothatitwas illuminating only the foveal regions of the subjects’ retinas. The photoreceptors inthesecentralregionsareonlylong-andmedium-wavelengthresponsivecones,whichareoften (but inaccurately) referred to as the red and green cones, and their luminoussensitivity functions are shown in Figure4.1 as L(λ) andM(λ) respectively. It should benotedhowsimilarare the responsesof these twocones,particularlywhen it isborne inmindthatitisthedifferenceinresponseofthispairoftwoconetypesthatenablescolourdiscriminationonthered–greenaxis,andalso,howcloselysimilar theyaretoV(λ).Theresponsesoftheshort-wavelength(blue)cones,shownastheS(λ)function,aswellasalloftherods,aresimplynottakenintoaccountbytheV(λ)function.

Foraphotopically-adaptedviewer,theS(λ)functiondoesnotaffectacuityforafixatedtask, but it does affect assessments of the brightness of the surrounding field, and thisoccurs to an extent that changes with field luminance. The Bezold-Brücke hue shiftdescribes the effect of perceived colour differences on the blue–yellow axis increasingrelative to those on the red–green axis with increasing luminance, and this affectsbrightnessassessments.Reaetal.(2011)haveproposedaluminoussensitivityfunctionforbrightness:

wherethevalueofgisrelatedtofieldluminance.Inthisway,avariableallowancefortheresponseoftheshort-wavelengthconescanbeaddedtothelong-andmedium-wavelengthcones dominated V(λ), and Mark Rea has tentatively suggested that for the range ofluminousenvironmentsdiscussedinthisbook,forwhich10<MRSE<1000lm/m2,agvalueof 3.0 would be appropriate. The resulting luminous sensitivity function, indicated asVB3(λ), isshowninFigure4.2. It isproposedthatapplyingthisfunctionforpredictingormeasuringMRSEwouldgivemorereliableresults,intermsofbettermatchingmetricstoassessments,thanusingconventionallumen-basedmetrics.

Meanwhile the CIE has given attention to other deficiencies of V(λ) by definingadditional luminous sensitivity functions, the most notable being the V′(λ) functionintroduced in1951,whichdefines therelativeresponseof therodphotoreceptors,andsorelatestoscotopically-adaptedvision(Figure4.3).Thisfunctionshowssubstantiallygreatersensitivityforshorterwavelength(blue)radiantflux,butwhileresearchscientistsareabletorecalculateluminousfluxaccordingtotheviewingconditions,thisdoesnothappeningenerallightingpractice.Thenotionthatthelumenoutputofalampmightdependonthecircumstancesof itsuse isa complication that the lighting industrywouldnotwelcome,and so the 1924 V(λ) function persists. Until lighting practice comes to termswith thisdiscrepancy,somelevelofmismatchbetweenmeasuredorpredictedlightingperformanceandhumanresponse is inevitable.Fordesigners, itbecomesamatterofhowwebalancesimplicityandconvenienceagainstactuallyprovidingwhatwehavepromised.

It may be noted that the visual field has to become distinctly dark, with adaptationluminancelessthan0.001cd/m2,forvisiontobecomeentirelyduetotherodphotosensors.Whenthisoccurs,scotopicconditionsprevailandtheV′(λ) luminoussensitivity functionapplies,sothatscotopicluminousflux:

Figure4.2TheVB3(λ)spectralsensitivityofbrightnessfunctionfordaytimelightlevels,wherethecontributionoftheS

conesrelativetoV(λ)ishigh(g=3).AfterRea(2013).

Figure4.3TheV(λ)andV′(λ)relativeluminousefficiencyfunctionsrelatetophotopicandscotopicadaptation

respectively.

In this way, while the photopic luminous flux, F, for a given source is determined byapplicationofFormula4.1, itsscotopic lumens,F′,couldbedeterminedbyapplicationofFormula4.3.Note the increasedvalueof the constant in this formula to reflect thehighsensitivityofdark-adaptedrods. It followsthat if thevalueofF′/F,referredtoastheS/P

(scotopic/photopic) ratio, is high, then at low light levels,where the rods are active, thevisual response will be underrated. Sources rich at shorter wavelengths, such as metalhalidelamps,will,forthesamelumens,generatestrongervisualresponsesthanlampsrichatlongerwavelengths,suchassodiumlamps.

Someothervisualandnon-visualresponses

While it would seem quite straightforward that F′ should be used as the measure forluminous flux for scotopic conditions, these conditions are in fact so dim that nobodyactuallyprovidesilluminationtoachievethem.Lightingpracticeforoutdoorspaces,suchas car parks, roadways and airport runways, aims to provide conditions in themesopicrange,whichextends from0.001cd/m2 up to the lower limit of thephotopic range, at 3cd/m2.Within this substantialadaptation luminance range, spectral sensitivityundergoestransition between scotopic and photopic adaptation, andwherewe are concernedwithbrightnessassessments,thismeanstransitionbetweentheverydissimilarV′(λ)andVB3(λ)functions,whichmakesaccurateassessmentofthelikelyvisualresponseproblematic(Rea,2013).Thisisarealissueforprovidingilluminationatoutdoorlightinglevels.

For indoor lighting at photopic levels, there are some different issues that concernresearchers.Ithasbeenestablishedthat,atthesameluminancelevels,pupilsizeissmallerforhigherS/Pillumination,andthisledtotheassumptionthatpupilsizeisdeterminedbythe response of the rod photoreceptors, even at photopic levels. Berman et al. (1993)conductedaseriesoflaboratorystudiesfortasksclosetothevisualthreshold(thepointatwhichthereisa50/50probabilityofaccuratedetection)andshowedthatperformancewasbetterforhigherS/Psources.Itmightseemoddthatreducedpupilsize,whichmustreducethe amount of light reaching the retina, should give increased performance, but theexplanationofferedwasthatreducingthelensaperturewouldimprovethequalityoftheretinalimage.Aswithacamera,smallerlensaperturegivesincreaseddepthoffield,whichisanadvantageforanyonewhoserefractivecorrectionis lessthanperfect. Italsooccursthat rays passing through the peripheral zones of the eye’s lens tend to undergoaberrations,asthelensoftheeyeis,infact,ofnomorethanmoderateopticalquality,sothatreducingobservers’pupilsizesis likelytocausethemtoexperienceimprovedimageresolution. It was claimed that these advantages would more than compensate for thereducedretinalilluminance.

ApplicationoftheS/Pfindingstolightingpracticehasrecentlybeenthesubjectofbothresearch and debate. The notion that visual performance could be maintained at lowerilluminancelevelsoffersopportunitiesforsignificantenergysavings,andthiscertainlyhasarousedinterest,butithasbeenpointedoutthatthehigherperformancedemonstratedforthreshold visual taskswould be unlikely to apply for themuchmore usual condition ofsuprathreshold tasks. General lighting practice aims to ensure that tasks are performedwithhigh ratesof accuracy,meaning that theyare tobe illuminated towell above theirthresholdlevels,sothatadvantagesthatmayoccurinanexperimentwheretheprobabilityof error is high probablywouldnot occur in practical situations (Boyce, 2003).A recentfield study byWei et al. (2014) of office workers found not only that any advantagesattributabletohighS/Psourcesweretoosmalltobeworthwhile,butalsothatthepeople

workinginthoseconditionsdislikedthehighS/Plighting.Amongtheresearchcommunitytherenowseems tobea lackof interest inpursuing this topic,but thathasnot stoppedsome unscrupulous suppliers from making claims that are exaggerated, and evendownright false, for high S/P lamps. It may be noted in passing that since the originalinvestigations, researchers have become aware that pupil size response ismore complexthan simply responding to the level of rod cells stimulation, and seems to involve therecentlydiscoveredipRGCresponse(seefollowingparagraph).

Humansexhibitvariousnon-visualresponsestolight,andthemost important,at leastfrom our point of view, is the circadian response, being the 24-hour cycle that weexperiencealongwithmostlivingthingsonthisplanet.Withtheonsetofcircadiannight,ahormonenamedmelatonin isreleasedfromthepinealgland into thebloodstream,andthis isassociatedwiththesleep/wakecycle that issaidtoberegulatedbyahypotheticalbiological clock that each one of us carries inside us. Researchers had noted that themelatonin response to light exposure displays a spectral sensitivity that does notmatchthatofanyof the retinalphotoreceptors,but itwasnotuntil 2002 that themysterywassolved.Theanswer lies in thecomplexpatternofconnectionswithintheretinathat linkthephotoreceptorstotheopticnerveforcommunicationtothebrain.Retinalganglioncellswereknowntoplaymajorrolesinthisprocess,butwhathadnotbeensuspectedwasthatsomeof thesecellsactuallycontainaphotopigment,whichhasbeennamedmelanopsin,andthelightresponseoftheseintrinsicallyphotosensitiveretinalganglioncells(ipRGCs)connectsnottothevisualcortex,buttotheendocrinegland,andontothepinealgland.Thepeaksensitivityofthesecellsduetothemelanopsinphotopigmentoccursat460nm,whichissubstantiallyshorterthanthepeakresponsesofanyoftheretinalphotocells.

Figure4.4Rea’sproposedVC(λ)functionfortherelativecircadianresponse(AfterRea,2013).

Rea(2013)hasproposedaspectralsensitivityfunction,VC(λ), for thehumancircadianresponse,which is shown inFigure4.4.This is rather different from the other functionsdiscussedsofarinthatitisnottheresponseofacell,butofasystem.Alargepartoftheresponse is additive, meaning that light at these wavelengths will have the effect ofdispersing melatonin from the blood, and part is subadditive, which means that for abroad-spectrumsource,energyatthesewavelengthswillhaveanegativeeffect,butifthetotalsumforthewholespectrumisnegative,theresponseshouldbeassumedtobezero.

Taking account of this function calls for a quite different way of thinking about theimpactoflightexposure.Beforetheinventionofelectriclighting,illuminationaftersunsetwaseitherabsent,oritwasoflowintensityandbiasedtowardlongerwavelengths,sothatcircadian cycles were largely undisturbed by after-dusk light exposure. While we allapplaud thebenefitsof electric lighting, a consequencehasbeena substantial growth innocturnallightexposure,andwhilemanyfindthislifestylechoiceattractive,healthstudiesofpeoplewhoengageinitoverlongperiods,suchasshiftworkersandairlinestaff,areacause forconcern.There is reason tosuppose thatdaytimeexposure to illumination thatscores highly on the circadian spectral sensitivity function, followed by night timeexposure toreduced levelsof lowscoring illumination,wouldbeconducive to long-termhealth.

While these human responses to light represent concerns that lighting designers canneverignore,therearetwoprincipalconcernsforthespectraldistributionofilluminationthat must always be at the forefront of a lighting designer’s mind. These are how thecolourappearanceoftheilluminationrelatestothedesignconceptofthespace,andhowthecoloursofilluminatedobjectswithinthespacewillberenderedbytheillumination.

Colourappearanceofillumination

Itiscommonexperiencethatsomematerialscanbeheatedtothepointwheretheybecomeincandescent,startingfromadullred,increasingwithtemperaturethroughbrightcrimsontobrilliantwhite-hot.Mostmaterialswouldmeltorevaporateifthetemperaturewastobefurther increased, but the theoretical ‘black-body’ does not have this limitation and itstemperaturecanberaiseduntilitbecomes‘blue-hot’.

Whenlampmakersdiscoveredhowtostepbeyondtherestrictionsofproducinglightbyincandescence, that opened up opportunities to produce light with different spectra,including light thatwas not far removed fromwhite butwhichwas distinctly differentfromthewarm,yellowishlightemittedbyahotmetalfilament.Infact,itbecamepossibletoproducelightthatmatchedtheappearanceofdifferentphasesofdaylightillumination,butthisraisedthequestionofhowtodescribethesevariationof‘white’lightinawaythatwould make sense to people choosing, or specifying, these new-fangled discharge andfluorescentlamps.

The answer they came up with was to define the colour appearance of all types ofnominally‘white’lightsourcesintermsofcorrelatedcolourtemperature(CCT),thisbeingthetemperatureofablack-body,specifiedindegreesKelvin,thatmostcloselymatchedtheappearance of the source in question. In Figure 4.5, the ‘black-body locus’ defines thechange inchromaticityofemitted light fromtheblack-bodyas its temperature isvaried,anditcanbeseenthatthiscorrespondstothecommonlyexperiencedchangeofcolourofemitted light when materials are heated. The invention of the halogen cycle enabledincandescentfilamentstobemaintainedattemperaturesofup3300K,andCCTdescribedtheappearanceoftheemittedlightquitereliably.However,therealneedforbeingabletoindicatethecolourappearanceof illuminationwasthedevelopingmarketforfluorescentlamps,wherespectraldistributionhasnothingtodowithtemperature.Therewasdemandforlightsourcesthatcouldprovide‘whitehot’,andeven‘blue-hot’,illuminationcolours,aswellasthecoloursofdaylightillumination,andfluorescentlampsmadeallofthispossible.Figure4.6 showsCCTvalues for some familiar lamp types related to colour appearance.TheconfusingwaysinwhichtheCCTscaleassociateslowcolourtemperatureswithwarmcolour appearance and high colour temperatures with cool colour appearance, and thatintervals on this scale are quite out of step with perceived differences, are both neatlyovercomebythereciprocalmegaKelvinscale(MK-1).Whilelampmakershaverecognisedtheusefulnessof this scale, ithasnot come intogeneraluseand, inanycase, ithas thedisadvantage that itassociates thechromaticityofablack-bodywithwhiteness,and thishashadunfortunateconsequencesthathavebeenshownupbyrecentresearch.

Figure4.5Theblack-bodylocus(solidline)plottedontheCIE1931(x,y)chromaticitychartwithintersectinglinesof

constantcorrelatedcolourtemperatureindicatedindegreesKelvin.AlsoshownarethechromaticitycoordinatesofCIE

StandardIlluminants,A,C,andD65(fromIESNA2000).

Rea and Freyssinier (2013) have reported a study in which subjects described theappearanceofdifferentlightingchromaticities,anditwasfoundthatthereisanextendedrange of chromaticities thatmay appear ‘white’, orwithminimum perceived ‘tint’, andimportantly,thesechromaticitiesdonotfollowthelineoftheblack-bodylocus.Figure4.7showsasectionoftheblack-bodylocuscrossedbylinesofconstantcolour

Figure4.6ThereciprocalmegaKelvinscale(MK-1)comparedwiththeKelvin(K)scale,andwithtypicalassessmentsof

colourappearanceandCCTsofsomefamiliarlightsources.

Figure4.7Contoursofperceivedleveloftint.Thesolidlineistheblack-bodylocusplottedontheCIE1931chromaticity

chart.Thelineof0%tintisthecontourofsourcechromaticitiesperceivedtohaveminimumtintatthatcolour

temperature,andthesearereferredtoas‘white’sources,withotherlinesshowingincreasinglevelsofperceivedtint.See

textformoreexplanation(fromRea,2013).

temperature (seeFigure4.5), and superimposedover these are lines of perceived level oftint. The 0% line is the experimentally-derived contour of ‘white’ sources. This doesnotmeanthatsourcechromaticitiesonthiscontourappearidentical,butratherthatatagivencolour temperature, any source chromaticity on this contour is perceived to be withminimumtint.WhilesourcesA,B,andCallhavethesamecolourtemperatureof4100K,theywillbeperceivedquitedifferently.Infact,sourceCwillappearmoresimilartosource1thantoeitherAorB,asbothCand1appeartobewithminimumtint.Departuresabove(+ive)orbelow(-ive)thiscontourincurincreasingperceivedtint,wherefordifferentpointsalong this contour, positive tint may appear slightly yellow, chartreuse or green, andnegativemayappearslightlypink,purpleorblue(Rea,2013).Itshouldbenotedthatthis‘white’source locusdepartssignificantlyfromtheblack-bodycontour,beingaboveit forCCTsabove4000K,andbelowitforCCTsbelow4000K.

Seeninthisway,itbecomesobviouswhyconventionallightsourcesaround4000Khavebeendescribedas ‘white’,andlowercolourtemperaturelightsourcesareperceivedtobeyellowish-whiteandaresaidtoappear‘warm’,andhighercolourtemperaturesourcesareperceivedtobebluish-whiteandaresaidtoappear‘cool’.Thenotionofthe‘black-body’beingthestandardreferencesourceisingrainedtothepointthatasthelightingindustryhas developed newer technologies, such as compact fluorescent lamps and now LEDsources, repeatedlytheiraimis tomatchthecharacteristicsof traditionalsources.At thetimeofwriting,examplesareoccurringoflightingcompaniesadvertisingnewLEDsourcesby claiming that the illumination is indistinguishable from halogen lighting. There is,however, at leastoneLEDmanufacturer that ispromoting itsproductasdeparting fromtheblack-bodylocus,butevenso, itmaybesomewhilebeforewehaveopportunitiestoexperience tint-free ‘white’ illumination of different colour temperatures in spaces thatenableustoproperlyassesstheirappearance.

Illuminanceandilluminationcolourpreference

Itwasway back in 1941 thatA.A. Kruithof, a lamp development engineerwith PhilipsLighting in theNetherlands,wrote an article describing the fluorescent lamp.This lamphad been introduced in theUSAonly three years earlier, and despite the turmoil of theSecondWorldWar,itwasfindingitswayintoEurope.AmongthemanyunfamiliaraspectsofthisnewtechnologythatKruithofdescribedwasthatitwouldbepossibletoselecttheCCToflighting.Thishadnotbeenpossiblepreviously,andtoprovideguidanceonhowtodothis,heincludedthediagramreproducedinFigure4.8.Thisfigureispossiblythemostreproduced diagram in the history of lighting. The white zone indicates acceptablecombinationsofilluminanceandCCT,andwithinthelowershadedzone,whichincludescombinationsoflowilluminanceandhighCCT,Kruithofdescribedtheeffectas‘coldand

harsh’,whileintheuppershadedzone,whichincludescombinationsofhighilluminanceandlowCCT,hedescribedtheeffectas‘unnatural’(Kruithof1941).

Thearticlegives little informationonhowthisdiagramwasderived,butKruithofhastoldtheauthorthatitwasa‘pilotstudy’basedentirelyontheobservationsbyhimselfandhisassistant.Forlowcolourtemperatures,incandescentlampswereswitchedfromseriestoparallel,butasthehalogenlamphadnotbeeninvented,thoseconditionswouldhavebeenlimited to 2800K.ForhigherCCTs, theyused some ‘special fluorescent lamps’ thatwerecurrently under development, but even with the resources of the Philips researchlaboratoriesatthattime,therangeofphosphorsavailablewouldhavebeenrestricting.Forsomepartsofthediagram,Kruithofreliedonacommonsenseapproach.ItisobviousthatoutdoordaylightwithaCCTof5000Katanilluminanceof50,000luxisveryacceptable,soheextrapolatedtothatpoint. Itwas inthiswaythatthediagramofthe ‘Kruithofeffect’wasputtogether.

Figure4.8Kruithof’schartrelatingcorrelatedcolourtemperature(TC)andilluminance(E)tocolourappearance.The

whitezoneisdescribedas‘preferred’,whilethelowershadedzoneappears‘coldandharsh’andtheupperzoneappears

‘unnatural’(fromKruithof,1941).

Sincethattime,severalresearchershavesoughttoapplyscientificmethodtodefiningasound basis for this phenomenon, but this has proved an elusive goal. However, the‘Kruithof effect’ lives on. Lighting designers continue to refer to it with reverence, andperhapsmoreconvincingly,youareunlikelytofindopportunitiestocarryoutobservationsoflightinginstallationthatoccurintheshadedareasofthediagram.Youwillfindthatthe

higher lighting levels provided in commercial and industrial locations, whether byfluorescentorhighintensitydischargelamps,tendtomakeuseofCCTscorrespondingtotheintermediateorcoolrangesshowninFigure4.6.EvenwhereCCTshigherthan5000Kareused,iftheilluminancealsoishigh(saymorethan1500lux),theeffectismoreinclinedtowardsabrightandcolourfulappearancereminiscentofdaylight,ratherthananoticeably‘cool’effect.Conversely,wherelightingisdeliberatelydim,thelowCCTsofincandescentlamps,orevencandles,arelikelytobethechosenlightsources.Ifyoupractiseobservationcoupled with measurement, you are likely to find ample confirmation of the Kruithofeffect.

Illuminationcolourand‘flow’oflight

There isan interestingdimensionofcolourcontrast thathasbeenroutinelyexploitedbystage lighting designers, and which has the potential to be influential in architecturallightingdesign.Peoplearesometimessurprisedbytheappearanceofcolourphotographstakenoutdoors in sunnyconditions.Areas in sunlightappear tohaveayellowcast, andparticularly for snowscenes, shadowsappearnoticeablyblue.Whileourvisual responsetendstoobscurethisnaturallyoccurringcolourdifference,ifyoulookforityoucanseeit,andmanyartists,particularlythe impressionists,haverecordedtheirobservationsofthis‘sunandsky’lightingeffect.

Stanley McCandless incorporated the effect into his method for stage lighting(McCandless1958).Anessentialfeatureofhisapproachisthatallobjectsonstagearetobeilluminatedfromoppositesides,withthelightfromonesidehavinglowerCCTtogiveasunlight effect, and the light from the other side having higher CCT, perhaps of lowerintensity, to give a skylight effect. In thisway, a distinct and coherent ‘flow’ of light isachievedwithoutstrongshadowsbeingcast.Thismeansthatanactorcanremainclearlyvisiblewhilehavinghisfaceintheshadow.

Whenyouareawareofthis‘sunandsky’lightingeffect,itissurprisinghowoftenyoucanfindexamplesofitinretaildisplaylighting.Carshowroomscanachieveveryeffectivedisplaysbyfloodingthespacewithdiffuselightusingrelativelyefficient‘daylight’lampswhichmighthaveaCCTofmorethan5000K,whileprovidinghighlightingfromspotlightshaving CCT close to 3000K. Clothing stores often use lower CCT spotlights to stronglyhighlightselected items thatarearrangedasverticaleye-catchingdisplays,whilerelyingonthecoolerappearanceofgeneral fluorescent lighting toreveal thedaylightcoloursofthemerchandisethatthecustomershandle.Blueisafrequentlyusedcolourfortheinternalsurfaces of display cabinets that have internal spotlights, and of course, it gives the skyeffect to the shadows. Everybody sees ‘flow’ of light effects of this sort, but it takes alightingdesignertoobservethevisualeffectandtounderstandhowitcanbeprovided.

Colourrenderingofillumination

Among themore spectacular developmentswithin the lighting industry during the pasthalf century has been progressive improvement in colour rendering, being the influencethat lighting has on the perceived colours of objects and materials. In fact, for mosteveryday lightingapplications,colourrenderingreallyhasceased tobeaproblem.Usershaveachoiceof lightsourcesthatarequitesatisfactoryfor industrialandoffice lightingapplications,aswellasforgeneral lightingforretail, recreational,andsocialactivities. Ithasnotalwaysbeenso,andwhentheColourRenderingIndex(usuallyabbreviatedtoCRI,butnotealsotheuseofscientificsymbolRabelow)wasintroducedin1965,itwasausefultoolforsortingoutthegood,theindifferentandtheplainugly.

CRIcontinuestoappearincodes,standardsandspecifications,wherestatementssuchas‘CRIshallbenotlessthan85’isasimpleformulaforavoidinglamptypesthatwouldcauseunsatisfactoryuserresponses.However,fortheapplicationswherecolourrenderingisanimportantfactor,CRIfailstoprovidereliableguidance.Artgalleryandmuseumdirectorshave learned the hard way that simply specifying a high CRI value does not ensureexcellent,orevenacceptable,appearanceofdisplays.

TherehavebeenseveralproposalsovertheyearstomakeCRImoreuseful,ortoreplaceitwithsomethingbetter.Thefollowingsectionsreviewsomeoftheseproposalsandoffersguidanceoncomingtotermswithcolourrendering.

TheCIEColourRenderingIndex

TheInternationalCommissiononIllumination(CIE)definescolourrenderingasthe‘effectof an illuminant on the colour appearance of objects by conscious or subconsciouscomparisonwiththeircolourappearanceunderareferenceilluminant’(CIE,1987).

The supposition here is that the observer is fully adapted to the same lighting thatilluminatestheobjects,andthatthecolourappearanceoftheobjectswouldbenatural,andtherefore optimal, if the lighting was provided by a reference source. The concept of areferencesourceiscentraltoanydiscussionofcolourrenderingasitprovidesthebasisforthe comparison that is contained in the definition. It is an inherent assumption that theperceivedcoloursofobjects litbytheappropriatereferencesourcewouldappearentirelyacceptable,andthatanydeparturefromthisappearancewouldbedetrimental.

As the brightness and the colour of the ambient illumination in our environmentchanges, theresponseofourvisualsystemadaptstotheambientcondition.CRIassumesphotopicadaptationandmakesnoadjustmentforbrightness,whiletheobserver’sstateofchromaticadaptationisassumedtobedeterminedbythechromaticityoftheactuallightsource.Thecorrespondingreferencesourceisaccordedacolourtemperaturethatmatches

thecorrelatedcolourtemperature(CCT)ofthelightsource.ForCCTslessthan5000Kthereferencesourceistheblack-body,andfor5000KandaboveitisaCIEstandarddaylightdistribution defined by its CCT. Getting these assumptions in mind is essential forunderstandingCRI.

TheCRIvaluesforatestsourcearedeterminedbytheTestColourMethod(CIE,1994).Fourteen testcoloursamples (TCS), listed inTable4.1, aredefinedby individual spectralreflectancecurves.ForeachTCS,u,vchromaticitycoordinatesonthe1960UCS(UniformChromaticityScale)chartarecalculatedforboththe testsourceand itsreferencesource,andacolouradaptationtransformisappliedtoallowforchromaticadaptationdifferencesbetween the two sources. After that, colour differences in UCS space are calculated foreach TCS under both sources. Each difference is defined by a vector that specifies thecolourshiftforviewingtheTCSalternativelyunderthereferencesourceandunderthetestsource,allowingforadaptationtoeachsource.ThemagnitudeofeachvectorΔEienablestheSpecialColourRenderingIndexRiforeachTCStobecalculated:

FromonlythefirsteightTCSvalues,theGeneralColourRenderingIndexRaiscalculated:

Thismayseemcomplicated,buttheCIEdocumentationincludesacomputerprogramthatperformsthetaskeffortlessly.Whilethistakesawaythepainforthelampmanufacturer,itisnecessaryforustounderstandwhatisbeingdonesowecanseehowitmightbedonebetter.TheprogramoutputforastandardWarmWhitehalophosphatefluorescentlampisshowninFigure4.9.

Table4.1The14CIETCS(Testcoloursamples).TCS1–8comprisetheoriginalsetofmoderatelysaturatedcolours

representingthewholehuecircle,andthesearetheonlysamplesusedfordeterminingCRI.Theothersixhavebeen

addedforadditionalinformation,andcomprisefoursaturatedcolours,TCS9–12,andtwosurfacesofparticularinterest.

Regrettably,detailsofcolourshiftsfortheseTCSareseldommadeavailable

There is plenty to ponder here. The lamp is, of course, an old-fashioned fluorescentlamp, and it is sobering to realise that when CRIwas introduced in 1965, this was thestandard lamp for general lighting practice. The program gives the x,y chromaticitycoordinates,theCCT(Tc),andameasureofhowfarthechromaticityisofftheblack-bodylocus (dC).TheCRI (Ra) is theaverageofRivalues forTCS1–8,and itcanbeseen thatthese vary substantially. Referring to Table 4.1, colour shifts are relatively small for theyellow-green and violetTCSs, but become large in other zones. Then look at the strongcolours,particularlythestrongred,representedasTCS9,forwhichthechromaticityshiftismassive,butthevalueforthisTCSwasnot,andstillisnot,takenintoaccountbyCRI.Human complexion (TCS 13) has a poor score, so it is no wonder that everybody waspleasedtoseethebackofthislamp,andreally,thathasbeentheforemostachievementofCRI.Nobodywouldnowdreamof lighting an indoor space inwhich the appearance ofpeoplemightbeofsomeconsequencewithsuchanutterlydismallamp.

Figure4.9OutputfromCIE133W.execomputerprogramtocalculateCRIs,foraWarmWhitehalophosphatefluorescent

lamp.Whilethisisanold-fashionedlamp,thisexampleillustrateswellthecolourrenderingissuesthatCRIwasdevised

tocopewith.

ProblemswithCRI

Despitethislevelofsuccess,CRIhasseveralproblems,someofwhichmaybeevidentfromtheprevioussection.TheCIEspecifies14TCSs,andcalculatesCRIfromjusteightofthem,ignoringtheothersix.ThereasonforthisisthatoriginallyonlyTCS1–8werespecified,and theyareallmediumsaturationcolours,butpeoplehadnoted that lamps thatmightperform reasonablywell for these colours could fail badly for rendering strong colours.Also,theappearancesofhumancomplexionandfoliagehavespecialsignificanceaspeoplehaveclearnotionsofhowtheyshouldappear,andsoitwasdecidedthatthesetooshouldbeadded.This ledtotheadditionofsixmoreTCSs,butthen,ratherthanchangeCRI, itwas decided that they should be listed separately to provide users with additionalinformation. However, while the program output gives these values, most users arecompletelyunawareof them.Manufacturersclaimthatpeoplewouldbeconfusedbytheadditional data, but nonetheless, it needs to be recognised that colour rendering is toocomplicatedanissuetobeadequatelydefinedbyasinglenumber.

To illustrate this point, if data for the additional six TCSswere to be provided,what

interpretationshouldbeplaceduponthem?AlowvalueofRiindicatesthattheappearanceof this TCS will be distinctly different under the test and reference sources, but noindicationisgivenofthenatureofthatdifference.Forexample,thenegativeRivaluenotedforthestrongredTCSmightindicatethatthetestsourceshiftsittowardsyellow,givinganorange tint, or towards blue, giving a mauve tint. Alternatively, it might appear lesssaturated,givingapinktint,oritmightappearmoresaturated,appearingasavividred.NotonlydoesCRIgivenoindicationofwhichofthesedifferencesoccurs,butittreatsallof them as being equally detrimental. There is good evidence to indicate that, withinreason,people like lamps thatmake their surroundings appearmore colourful, that is tosay,whichcause increasedsaturation.Thischallenges thecentralnotion thatareferencesourceprovidesoptimalcolourrendering.

Anotherissueisthatithasforsomewhilebeenacknowledgedthatthe1960UCSchartisfar from uniform in its spacing of chromaticity values, and since then there have beenseveralproposalsformoreuniformdefinitionsofcolourspace.TochangethecolourspacewouldaffectCRIvalues,sothishasnotbeendone,withtheresultthatCRIcontinuestobecalculatedusingaprocedurethatisknowntoevaluatecolourdifferencesunequally.

Thereareotherproblems.TheCRIscalecausesconfusion,someuserssupposingittobeapercentagescale,sothefactthatsomelampsareshowntohavenegativevaluescomesasa surprise.Also, becauseCRIhasbeen sowidelyusedby specifiers,manufacturershavedeveloped lamps to achieve high CRI values, so that they have incorporated theshortcomingsofCRIintotheirnewproducts.Ithasbecomeincreasinglyapparentthatthisapproach has led to lamps being promoted for good colour rendering but which havedistinctly less than optimal performance. These shortcomings of CRI became clearlyevidentwiththedevelopmentoftri-phosphorfluorescentlampsinthe1970s,andtheyarenowseentobeasubstantialhindrancetoprogressbycompaniesworkingondevelopmentofwhiteLEDsources.ItishightimeforchangestobemadetoCRI.

WhatisbeingdoneaboutCRI?

Therehasbeennoshortageofsuggestionsovertheyears,withpastproposalsforaColourDiscriminationIndex,andevenaFlatteryIndex.Whilethesemayhaveattractedattentionatthetimeswhentheywereproposed,theCIEhassetupaTechnicalCommitteetoreviseCRI and this project has gained support from theUSNational Institute for Science andTechnology. Ithas led to thedevelopmentof theColourQualityScale (CQS) (DavisandOhno,2004),whichisasubstantialrevisionofCRIandinvolvesanewsetof15testcoloursamples of high chromatic saturation spanning the entire hue circle, and it makes theswitch to 1976CIELAB colour space,which assesses different types of colour differencemorecloselytohowtheyappear.Shiftsofhueorshiftstolowersaturationaretreatedasbeingequallydetrimental,butshiftstohighersaturationincurnopenalty.Aweightingis

placedonCCT,sothatforCCTslessthan3500Kormorethan6500K,scoresaremodifiedbyascaledmultiplicationfactor.Thiswouldhavetheeffect,forexample,ofreducingthedomesticincandescentlamp’sratingfrom100to97.Thescaleitselfismodifiedtoeliminatenegative values, with the effect that all of the very poorly performing lamps will haveratingsbetween0and20.

The single rating indicatorwith amaximumvalue of 100 is retained, and the overallweightingofCQSbetween20and100isnottoodifferentfromCRI,althoughratingsforsome lamp types do undergo significant changes. In particular, itmay be expected thatlampswithmultiplenarrowwavebandemissions,suchasLEDcombinations,willachievemore favourableCQSratings than the ratings theygainunderCRI.Finally, toovercometheeffectofaveraging,bywhichalampmaygainamoderatelyhighscorewhileoneortwo test colours show large colour differences, individual scores are calculated as rootmeansquare(RMS)values.

Theretentionofasinglescaleindicatorofcolourrenderingsuitsspecifiers,whowouldcontinue to be able to prescribe aminimum value for a given application, andwhile itshouldreduceanomalies,itwillnotprovidelightingdesignerswithguidanceonhowthecolourappearanceofilluminatedobjectswillbeaffectedbythelightsource.SowhileCQSfallsshortofprovidinglightingdesignerswithalltheinformationtheyneed,itdoesgoalong way towards overcoming the anomalies incorporated into CRI. It is, however,importanttoappreciatethatwhileCQShasbeenpublishedanddiscussioninvited,atthetimeofwritingithadnotbeenendorsedbytheCIE.

Whatisthecurrentstateofknowledgeoncolourrendering?

Researchers in thecolour science fieldhaveachieved remarkable successduring thepastdecade, which has led to the development of colour appearance models (CAMs). Twoscientists had independently developed models for predicting how a typical observerperceives colours in the environment, each taking account of a range of variables andknownvisual phenomena.DrR.W.G.Hunt,with theKodakCorporation in theUK, hadspenta lifetimeworkingoncoloured imagesandDrY.Nayataniof Japandevelopedhismodeltoaddressconcerns in illuminationengineeringandcolourrendering. In1997, thetwomodelsweremergedtoproduceasingleColourAppearanceModel,CIECAM97s.Thiswastakenupwithenthusiasminarangeofindustrieswherecolourisacriticalaspectofquality control, particularly where imaging is involved, and soon the CIE TechnicalCommitteeconcernedhadavailableawealthoffeedbackgainedfrompracticalapplication.This ledtoCIECAM02,whichwasactuallypublishedin2004,andisconsideredlikelytoremainunalteredforsomewhileasitisbelievedtobeasgoodamodelascanbeproducedfromcurrentknowledge.ForareviewofCIECAM02,seeFairchild(2004).

The input data required to apply CIECAM02 to predict the colour appearance of an

elementinthefieldofviewincludecolorimetricdatafortheobject(stimulus)andthelightsource (adaptingstimulus), theabsolute luminanceandcolorimetricdataof theproximalfield,includingthebackgroundandsurroundtothestimulus.Thesuccessofthismodelliesinthevarietyofpotentiallyinfluentialfactorsthatmaybetakenintoaccount.InChapter1wenotedhowthecolourappearanceofanobjectcanbeaffectedbywhethercoloursareperceived as related or unrelated, and in CIECAM02, the effect of surrounding surfacesupontheperceptionofsurfacecoloursispredictable.Thisisjustoneofarangeofcolourappearancephenomena thathavebeenobservedandreportedover theyears,andwhichhave subsequently been researched andquantified, andnowhavebeen combined into asinglecomprehensivemodel.

Thespectralpowerdistributionofthelightsourceisoneoftheinputvariables,andsoaspects suchashowbright and colourful a specific objectwill appear in a given settingcould be examined for alternative lamps. In terms of applied scientific knowledge, thisundoubtedlywouldbealeapforward.However,wecannotuseCIECAM02inthewaythatweuseCRI,thatistosay,wecannotuseittodescribethecolourpropertiesofalamp,asithastobeappliedtoaspecificviewingsituation.Perhapsthiswillbecomepossibleoneday.Thespectacularadvancesincomputervisualisationsoftwarethathaveoccurredduringthepast decade might enable us to model the effect of different light sources upon colourappearanceofarealorsimulatedscene,butmeanwhile,weneedtothinkaboutwhattheinformationisthatwouldbeusefultousnow.

Whatdowewanttoknowaboutcolourrendering?

Whenwe get down tomeeting actual needs for presenting coloured objects for criticalexamination and assessment, it becomes apparent that those who put such objects ondisplay have learned a lot about people’s preferences for colour appearance. Forconfirmation of this, you need look no further than your local supermarket. The freshproducedisplaysusedifferentlamptypesforthemeat,fish,fruitandvegetables,aswellasfor the ‘deli’displays,allofwhichhavebeenchosen forhowtheyrender thecoloursofthat particular type ofmerchandise, and quite obviously, those choices have beenmadewithout reference toCRI. Theway inwhich colour rendering is understood by theCIEexpertsisclearlyindicatedbythedefinitiongivenatthebeginningofthissection,butitisapparent that thepreferencesshownbypeoplemakingvisualselectionsof freshproducehavenothingtodowithmakingcomparisons,consciousorsubconscious,withappearanceunderareferencesource.

It can be seen that the lamp type chosen for each of the fresh produce applicationsimparts a particular type of colour shift, and the store operators havemade themselvesaware of which type of colour shift suits each type of merchandise. For the lightingdesigner who encounters a situation that calls for a certain type of colour shift, the

available lampdata fail to provide thenecessary guidance.Manufacturers give theCCTandCRIvalues,andalsotheymayshowthespectralpowerdistributioncurve,butnobodyshouldassumethatthereisasimplerelationshipbetweenthiscurveandcolourrenderingproperties. Even for an experienced lighting designer, an SPD curve comprising acombinationoflinespectraandbroad-bandemissionsgiveslittleornousefulguidanceoncolourrendering.

Whatisneededisastraightforwardwayofshowingwhatalampwilldotothecolourappearanceoftheobjectsthatitilluminates.Alightingdesignerdoesnotneedtobetoldwhat is good andwhat is bad. The information that the designer needs is to enable aninformed decision onwhich lamp typewill best suit his or her purpose for a particularapplication.Thisleadsustothecolour-mismatchvector.

Thecolour-mismatchvector(CMV)method

In1988,twolampengineersatPhilipsLightingintheNetherlandsproposedanovelwayofpresentingcolourrenderinginformation(vanKemanadeandvanderBurgt,1988).Figure4.10(a) shows thechromaticityshifts fora setof215coloursmoreor lessequallyspacedover the chromaticity chart,when illuminated by a reference source and then by a testsource.Theindividualcolour-mismatchvectorsareplottedontotheCIELABchart,andinthiscase,thetestlampisahalophosphatefluorescentlampnotverydifferentfromtheonerepresented inFigure4.9.Eachvector indicates theextentanddirectionof themismatchbetweenthereferencesourceandthetestlamp.Avectorpointingtowardsthecentreofthechartindicatesachromareduction,andaradialdirectionindicatesahueshift.Itshouldbenotedthatthevectorsarenotrandomlyscatteredbutshowadistinctflowpattern,anditshouldnotcomeasasurprisethatmismatchesincreaseforhigherchroma,thatistosay,forTCSpointsfurtherfromthecentre.

Themain featuresof the flowpatternareexpressed inFigure4.10(b)and (c). Thehueangle on these graphs ismeasured froma* anticlockwise, so relationship to the uniquehuescanbereadfromFigure4.10(a).Itiscleartoseewhereaboutsonthehuecircleatestlamp introduces hue shifts or changes in chroma. The authors includedmore charts forfluorescentlampswithdifferentcolourrenderingproperties,andaninterestingcomparisonofwhiteSONandmetalhalide.

Atfirstthistypeofchartmayappearintimidating,butwithalittlepractice,thewealthof information that it provides is easily extracted. There is still the comparison with areference source, but instead of the system deciding what is good or bad, it is for thelightingdesigner tochoose thecolour renderingcharacteristics thatwill suitaparticularapplication.Quite apart from those freshproduce supermarketdisplays, howelsedoes adesignerselectthemostsuitablelampforanindoorswimmingpool,oramake-upmirror,oranorchiddisplay,oranexhibitionofantiquemanuscripts,oranice-coldvodkabar?The

CMV method enables designers to make informed lamp selections based on colourrenderingcharacteristics.Unfortunately,lampmanufacturersareshowingthemselvestobereluctanttoprovidethisinformation,particularlyforthenewergenerationoflightsources.

Colourgamutarea

Itmight seem that an ideal light sourcewouldproduce aCMVdiagram inwhich everyvector radiates outwards from the central point, creating a colourfulworld inwhich allcoloursappearmoresaturated,andthereisevidencetoindicatethatpeopledopreferlightsourcesthattendtoincreasecoloursaturation,atleasttosomeextent.

AcolourgamutisthepolygonformedwhentheeightTCSs,illuminatedbyagivenlightsource, are plotted onto the CIE UCS diagram. Equal distances between points on thisdiagram correspond approximately to equal perceived colour differences, so that therelative areas of the polygons formed by connecting the TCS points for a given sourceprovideanindicationofthe‘colourfulness’associatedwiththatsource.Figure4.11showsthecolourgamutsforarangeofwidelyusedlightsources,andthegeneraltrendof

Figure4.10Colour-mismatchvectordataforahalophosphateCoolWhitecolour33fluorescentlamp(Fromvan

KemenadeandvanderBurgt,1988).

A)CMVsontheCIELABchartforOpstelten’ssetof215testcoloursamples.

B)HuecomponentofCMV,where+iveΔHueindicatesshifttohigherhueangles.

C)RelativechromacontentofCMV,whereΔC*=Δchroma/chroma,and+iveΔC*indicatesincreaseinsaturationwith

respecttoreferencesource.

increasinggamutareaswithincreasingCCTisclearlyevident.NotethelargeareaoftheDaylightsource,actuallytheCIED65daylightstandard,anditbecomesevidentwhythissourceisoftenregardedasthelightsourceagainstwhichallothersshouldbejudged.

Figure4.11GamutareasforsomefamiliarlightsourcesplottedontheCIE1976UCS(uniformchromaticityscale)

diagram.Gamutarearelatestotheperceived‘colourfulness’associatedwithalightsource(fromBoyce,2014).

Boyce (2003) has noted a correspondence between gamut areas and findings fromresearchstudies into thephenomenonof ‘visualclarity’ (BellchambersandGodby,1972).Although this concept has never been precisely defined, a variety of studies have foundthatwhensubjectscompareadjacentscenesandareinstructedtoadjustthelightlevelinone‘sothattheoverallclarityofthesceneisthesame’asintheother,alowerilluminanceis set in the scene with greater colour gamut area. Boyce’s formula for predicting theilluminanceratioformatchingappearancefromthegamutarearatiois:

whereE1andE2aretheilluminancevaluesandG1andG2arethegamutareasforthetwolightsources.

Quite separately, Rea (2013) has reported that CRI does not reliably predict people’scolourpreferences for fruit,vegetables, skinandotheroften-encounterednaturalobjects,and has proposed that light source gamut areas should also be taken into account. Thegamut areas calculated from the u’, v’ values of the UCS diagram produce very smallvalues,leadinghimtoproposeagamutareaindex:

whereGS is the gamut area of light source S, and Gees is the gamut area for an equalenergysource, forwhich thevaluehasbeencalculated tobe0.007354.ThevalueofGAImaybemoreor lessthan100accordingtothegamutareaofS,andReaadvisesthatforpreferred appearance of natural objects, which, of course, includes other people, lightsources should be ‘high in CRI and high (but not too high) in GAI’. This leads to hisproposal for ‘Class A’ colour for general illumination light sources, for which thechromaticityshouldlieonthe‘white’sourcelocus(Figure4.7),andCRIshouldbeequaltoormore than 80,and GAI should be between 80 and 100 (Rea, 2013). For specificationwriters,astatementalongthelinesof‘AlllightsourcesshallbeofClassAcolour’couldbeexpectedtoimprovereliability.

Sourcespectrumandhumanresponse

At first sight, this review of how the spectral properties of illumination may influencepeople’s responses to a lit scene might seem to comprise a bewildering array ofdisconnectedfactors,someofwhichhavebackgroundsof intensiveresearchwhileothersare based on not much more than casual observation. However, some introspectionsuggests an underlying pattern that gives some insight into how these factors areconnected.

Itisclearthatwhenweareconcernedwithabrightnessresponseratherthanvisibility,V(λ)tendstounderratesourcesthatarerichintheshortervisiblewavelengths,thatistosay,sourcesthatarehighinS/PratioandCCT.Whileithaslongbeenrecognisedthatthisoccurs for scotopic conditions, the B(λ) function (Formula 4.2) applies for photopicconditionsaswell.TheVB3(λ) function (Figure4.2)hasbeenproposedas theappropriateilluminationmetricforindoorgenerallightingpractice,buthasyettogainacceptance.

IlluminationthathashighluminousefficiencyontheB3metricwouldalsoprovidewellfor circadian response, measured on the VC(λ) function (Figure 4.4), making it anappropriatesourcefordaytimeillumination.LightsourceswithCCTvaluesaround4000Karecommonlydescribedas‘white’lightsources,anditmaybenotedthatthisistheCCTvalueatwhichthe‘white’sourcelocuscrossestheblack-bodylocus(Figure4.7),suggestingthathigherCCTsourceswithchromaticitiesonthe‘white’sourcelocusmightnotattractthenegativeassessmentsaccordedtothehighS/Psourcesusedinrecentresearchstudies.Theusefulnessof thehigh retinal image resolutionassociatedwithhighS/P sourceshasbeen questioned, but itwould seem reasonable to suppose that ‘white’ high S/P sourceswouldgainanysuchadvantageswithoutincurringnegativeassessmentsforappearance.

McCandless’notionof ‘sunlightandskylight’ suggestsoptions forattractiveeffectsbyaddinglowS/PhighlightstooverallhighS/Pillumination,andtheKruithofeffectpointstohigh S/P (or CCT) illumination gaining preference at high illuminance levels, in otherwords, high CCTs for daytime and low CCTs for night time. All of this fits in withprovidingilluminationtocoincidewiththecircadiancycle.

Rea’sproposalthat,forgenerallightingpractice,theshortcomingsofCRImaybelargelyovercome by specifying ‘Class A’ colour defines a basis for generally preferred colourrendering. Figure 4.11 shows clearly how, for high CRI sources, gamut area (related tocolourfulness)tendstoincreasewithCCT.WhileCRIrelatestothe‘naturalness’ofcolourappearance,Rea’sproposaladdsanewnotionof‘whiteness’,and,throughincludingGAIinthecriteria,theappearanceof‘colourfulness’.Thismay,inturn,beseentobeconsistentwiththe‘visualclarity’concept,andfurthermore,withtheothermoreanecdotalconceptsobservedbyMcCandlessandKruithof. In thisway, the rangeof factors reviewed in thischaptermaybeseenascontributingtowardsareasonableandconsistentunderstandingofhumanresponsetolightsourcespectrum.

Even so, a designer who wishes to have control over the appearance of a space, orselected targets within the space, is left in a difficult situation. He or she cannot avoidfeelingpoorlysupportedbytheinformationcurrentlyavailablefromthelightingindustry,and it is perhaps ironic that efforts to improve this situation tend to be resisted by theindustryonthegroundsthatsuchchangeswouldcauseconfusion.

Figure4.12TheGretagMacbethColorCheckercolourrenditionchartbeingexaminedunderdaylight.Aviewerwho

formsaclearmemoryimageofthechartinthissituationcanthenmakecomparisonswithitsappearanceunderother

sourcesofillumination.

MyownapproachhasbeentoequipmyselfwithaGretagMacbethColorChecker,andtousethistomakeobjectiveassessmentsofthecolourcharacteristicsoflightsources.TheColorChecker comprises24matt-surfacedcoloursamplesmountedonastiffboard,andsometimeneedstobespentexaminingitundermid-daydaylight,asshowninFigure4.12.Thebottomrowisagreyscale,fromfull-whitetofull-black,andinthisviewingconditionallthesamplesappearneutral(nohintofhue),andthestepsbetweenthemappearequallyspaced.Thenextrowupcomprisesprimarycolours,withtheadditiveprimariestotheleftandthesubtractiveprimariestotheright,andallofthemappearasfullysaturated,clearcolours. The two rows above are moderate colours, some with special significance. Forexample,startingfromthe left-handendof thetoprow,thesamplesrepresentdarkskin,lightskin,bluesky,foliageandsoon.Explanationsaregivenonthereverseside.

StartbygainingexperienceoftheappearanceoftheColorCheckerunderdaylight.Thisgivesyoua tool that enablesyou toobjectivelyassess thecolourcharacteristicsofotherlightsourcesandilluminationconditions,whetheryouareevaluatingasampleofnewtype

of light source or visiting a recent lighting installation. The appearance of theColorCheckerwillquicklyrevealtoyouhowyourperceptionofcoloursisinfluencedbytheillumination.Itisworthnotingthatunderlowlightlevels,allthecolourswillappeardullandtheintervalsbetweenthegreysampleswillappearcompressedtowardsthedarkerend.Providingthatilluminationissufficienttoensurephotopicadaptation,theappearanceoftheprimariescanbeparticularlyrevealing.Whileyouwillbeaccustomedtoallofthesesamples appearing saturated, certain light sources can cause some of them to appearunexpectedly bright. To understand this, think back to the discussion of lamps used toenhancetheappearanceofvarioustypesoffooddisplays.Moregenerally,lookcarefullyattheappearancesofthemoderatecolours,notingthatpeopleareparticularlysensitiveaboutskincolours.Whenpeoplecomplainaboutcolourrendering,themostcommonlyoccurringcommentsareofthe‘Theymakeyoulookawful!’type.

Itisinthiswaythatalightingdesignermayselectlamptypesforvariousapplicationswithconfidencethattheeffectontheappearancesofcolouredroomsurfacesandobjectswill be in accordwith the overall design objectives. From the foregoing discussion, it isclearthatpeoplehavedifferentexpectationsfordaytimeandnighttimeillumination,andwheretheaimistosatisfythoseexpectations,thedesignershouldprovideforcoincidencewith the circadian cycle. Of course, circumstances will occur where the intention is toachieve alertness and visual stimulation when people would naturally be inclined torestfulness,and for theseapplications the intensityanddurationofbright lightexposureneedstobegivenconsideration.Meanwhileitistobeexpectedthatdevelopmentsinlightsourcetechnologywillprovidedesignerswithincreasedoptions,anditistobehopedthatthelightingindustrywillrespondwithmoreusefulproductinformation.Inparticular,thatitwillrecognisethatwhiletheneedsofspecifiersmaybebestmetbyfamiliar,singlefigurevalues, designers’ needs are more complex. They need information that addresses theforegoing issues, and this is not met by catalogue pages presenting brightly colouredspectralpowerdistributioncurves.

References

Bellchambers, H.E. and A.C. Godby (1972). Illumination, colour rendering and visualclarity.LightingResearchandTechnology,4:104–106.

Berman,S.M.,G.Fein,D.I. JewettandF.Ashford(1993).Luminance-controlledpupilsizeaffects Landolt C task performance. Journal of the Illuminating EngineeringSociety,22:150–165.

Boyce, P.R. (2014).Human Factors in Lighting, Third edition. Boca Raton, FL: CRCPress.

Commission Internationale de l’Eclairage (CIE) (1987). International LightingVocabulary,item845-02-59,CIE17.4-1987.

——(1994).MethodofMeasuringandSpecifyingColourRendering,CIE13.3-1994.

Davis,W.andY.Ohno(2004).Towardsanimprovedcolorrenderingmetric.Proceedingsofthe5thInternationalConferenceonSolidStateLighting,SPIE5941.

Fairchild, M.D. (2004). Color Appearance Models, Second edition. Chichester: JohnWiley&Sons.

IlluminatingEngineeringSocietyofNorthAmerica(IESNA)(2000).TheIESNALightingHandbook,NinthEdition.

Kruithof, A.A. (1941). Tubular luminescence lamps for general illumination. PhilipsTechnicalReview,6(2):65–73.

McCandless, S.R. (1958).A Method of Lighting for the Stage (4th edn). New York:TheatreArtBooks.

Rea,M.S.(2013).ValueMetricsforBetterLighting.Bellingham,WA:SPIEPress.

Rea,M.S.andJ.P.Freyssinier(2013).White lightingforresidentialapplications.LightingResearchandTechnology,45(3):331–344.

Rea, M.S., L.C. Radetsky and J.D. Bullough (2011). Towards a model of outdoor scenebrightness.LightingResearchandTechnology,43(1):7–30.

VanKemanade,J.T.C.andP.J.M.vanderBurgt(1988).Lightsourcesandcolourrendering:AdditionalinformationtotheRaindex.ProceedingsoftheCIBSENationalLightingConference,Cambridge,UK,pp.133–143.

Wei,M.,K.W.Houser,B.Orland,D.H.Lang,N.Ram,M.J.SliwinskiandM.Bose(2014).FieldstudyofofficeworkerresponsestofluorescentlightingofdifferentCCTandlumenoutput.J.Environ.Psychol.,http://dx.doi.org/10.1016//j.envp,2014.04.009

5SpatialIlluminationDistributions

Chaptersummary

Theappearancesofthree-dimensionalobjectsareinfluencedbythelightingpatternsthatare generated through interactions between the objects and the spatial distribution ofillumination.AsnotedinChapter1,therearethreetypesoftheseobjectlightingpatterns;shading,highlightandshadowpatterns;andtheyappearsuperimposedovereachobject’ssurface in response to the optical characteristics of the objects and the photometriccharacteristicsof thesurrounding light field.The light field isalsoexamined in termsofperceived characteristics, and the concepts of the ‘flow’ and the ‘sharpness’ ofillumination are discussed. These characteristics may have the effects of revealing, orenhancingorsubduingtheappearanceofselectedobjectattributes.Theperceivedstrengthof the ‘flow’ of light relates to the vector/scalar ratio (VSR) and its perceived directioncorresponds with the vector direction. The highlight contrast potential (HCP) gives anindicationoftheextenttowhichlightingmayprovideforperceived‘sharpness’.

Three-dimensionaldistributionsofillumination

InChapter1,wenotedthatthegreencylinderinteractedwiththeilluminationdistributiontoproduce lightingpatterns thatappearedsuperimposedonto thesurfaceof thecylinderand the checker board (Figure 1.1), and these patterns not only affected the cylinder’sappearance, but they influenced ourwhole understanding of the surrounding light fieldand the objects within it. In this chapterwe are going to look closely at these lightingpatterns,andwearegoingtoidentifythethree-dimensionalcharacteristicsofilluminationthatcausethem.

Thethreeobjectlightingpatterns

ThethreeobjectsshowninFigure5.1areall interactingwiththesamesurrounding lightfield, but the object lighting patterns produced by those interactions are strikinglydifferent.Themattwhitespherehasformedagradedshadingpatternofvaryingsurfaceilluminancerelatedtosurfaceorientation,andinthisrespect,thepatternfollowsfromthecosine lawof illumination.Completelydifferent in appearance is thehighlight patterngenerated by the glossy black sphere,which is formedby specular images of the higherluminanceelementsinthisspacethatarealsothesourcesofillumination.Differentagainistheshadowpatternproducedbythepeg-on-a-disc,wheretheshadowcastbythepegisclearlyrevealedonthedisc’ssurface(Cuttle,1971).

Figure5.1Thetripleobjectlightingpatternsdevice.Thisdeviceseparatesasfaraspossiblethethreeobjectlighting

patterns.Themattwhitesphereshowstheshadingpattern;theglossyblacksphereshowsthehighlightpattern;andthe

peg-on-a-discshowstheshadowpattern.

Eachoftheselightingpatternstellsussomethingdifferentaboutthethree-dimensionallightfieldsurroundingtheseobjects.Lookcarefullyatthemattwhitesphere.Nopartofitssurfaceisunlit,butthereisadistinctbias.IfIcouldhandyouasmallarrow,youwouldbeabletoplaceitontheimagetocoincidewithyourperceptionofthedirectionofthe‘flow’

of light. Itwouldnotmatter howmany sources of illumination are present, alwaysyouwouldperceivejustone‘flow’direction.Youmightalsodescribetheapparentstrengthofthe ‘flow’ as being distinct, but not strong. Now turn to the glossy black sphere. Itsappearanceisdominatedbyasingle‘highlight’image,andifyoulookcarefullyyouwillbeabletomakeouttheshapeofthislightsource’soutlineandrecognisethatitisawindow.Nootherlightsourceisbrightenoughtoregisteranoticeable‘highlight’,andsoyoumayconclude that the window is the sole source of direct illumination. Finally, look at theshadowpatternformedonthepeg-on-a-disc.Itsdirectioncoincideswiththeappearanceofthe ‘flow’direction,and like theshadingpattern, it isonlymoderatelystrong.Also, it isquitesoftydefined,asthislightinglacks‘sharpness’.

You will have worked out by now that you have been looking at lighting patternsgenerated by the light field in a small, or moderately sized, room with fairly light(reflective)roomsurfaces,litbyasingleside-window.Thisisaprettydetaileddescriptionofthelocationandthelightfieldwithinit.Whatwouldbetheeffectifweleavethetriple-objectinitspresentpositionbutchangethelighting?

Figure5.1reappearsasFigure5.2(a),andbelowit,youseetheeffectofblankingoffthewindowandintroducingaspotlightinFigure5.2(b),andthenturningoffthespotlightandaddingsix smalldisplay lights inFigure5.2(c).The twocolumnsofphotos across to theright show the effects of these lighting conditions on the appearance of two groups ofobjects. The first column shows a group of familiar domestic items, and we shouldappreciatethat,evenforobjectsthatweareunlikelytoselectfordisplaytreatment,theirappearancecanbesubstantiallyaffectedbylighting.Infact,theappearanceofeverythingthatwesee,pickupandmakeuseofisaffectedbytheobjectlightingpatternsformedbyits surrounding light field. Figure 5.2(d) shows the garlic pot and two capsicums in thedaylightsituation,whichhasadistinct‘flow’oflightwithout‘sharpness’.Theeffectofthesingle spotlight in Figure 5.2(e) is to reproduce quite closely the ‘flow’ direction whilesomewhatincreasingthe‘flow’strength,butthereallynoticeablechangeisthepresenceof‘sharpness’,revealedbythehighlight(evenmorenoticeableinreallifethanitappearsinthis image) and shading patterns. In Figure 5.2(f), the ‘flow’ revealed by the shadingpattern on the garlic pot has almost vanished, but the effect of ‘sharpness’ due to thelightingisstillhighlyevident.

Thesecondcolumnofphotosshowssomeobjectsthatwemightputondisplaytoattractinterest,andhere theobjectattributes include transparency, iridescence,anddichromaticcolours.These three figures,5.2(g), (h)and (i), call forcarefulattention.Howwouldyouapproachthetaskofprovidingdisplaylightingforthisgroupofobjects?Atfirstitmightseemthatalmostanythingcouldwork,butitshouldbenotedthatthethreeobjectlightingpatterns,beingtheshading,highlightandshadowpatterns,areseparatelyidentifiableandit is the different balances of these patterns that determine the spatial lighting effects.Carefulobservationoftheseimagesshowshowtheattributesofanyoneoftheseobjectsmayberevealed,enhancedorsubduedbythebalanceofobjectlightingpatternscreatedby

theirinteractionswiththelightfield.

Tosummarise,wehaveidentifiedthreeobjectlightingpatterns:

TheShadingPattern:Duetotheinteractionofanobject’sthree-dimensionalformwitha‘flow’oflight.Thepatternisavariationofsurfaceilluminanceduetochangingincidenceoflightwithsurfaceorientationwhichinfluencestheappearanceofobjectformandtexture.Thelightingmetricsthatrelatetothe‘flow’arethevector/scalar

Figure5.2Forthethreelightingconditionsdescribedinthetext;thefirstcolumnofphotosshowsthelightingpatterns

formedonthetriplelightingpatternsdevice,andthenexttwocolumnsshowthelightingpatternsonagroupofdomestic

objectsandagroupofdisplayobjects.

ratio (VSR),which correspondswith the perceived strength of ‘flow’, and the vector direction,which correspondswiththeperceiveddirectionof‘flow’.

TheHighlightPattern:Duetospecularreflectionsofrelativelyhigh luminanceobjects,particularly lightsources,thatappearsuperimposedonanobject’ssurface.Therehastobesomelevelofsurfaceglossforahighlightpatterntobeevident,andeitherpolishedmetalsorshiny,darkcolouredsurfacesgivemaximumeffect.Thispatterninfluencestheappearanceofgloss,sheenorlustre,andmaybedescribedasanaspectofthe‘sharpness’oflighting.Themetricthatrelatestotheseeffectsisthehighlightcontrastpotential(HCP).

TheShadowPattern:Duetoashadowcasterprojectingashadowontoareceivingsurfaceinadirectionallightfield.The appearance of this lighting pattern may be described in terms of both the strength and ‘sharpness’ of castshadows,anditmayinfluencetheperceptionofobjectform,textureand/orlocation.PerceivedshadowstrengthisassociatedwiththeVSR,and‘sharpness’withtheHCP.

Theconceptofobject lightingpatterns is readilyunderstoodbynon-lightingpeople andcanformausefulbasisfordiscussionwhenlightingdesignersaretalkingabouthowtheirproposals will affect the appearance of the various objects that will form significantcomponents of the design. Designers are usually able to communicate their ideas usingtheseconceptswithoutgoingintodetails,suchasexplainingtheprecisedifferencebetweenashadingpatternandashadowpattern.However,whilenon-lightingpeoplewillperceivethe lighting patterns entirely as a visual effect, for designers, there is a deeper insight.Every lighting pattern is recognised to be a three-dimensional interaction between aparticular typeof surfaceandaparticular typeof incident light.Theunderstanding thattherearejustthreetypesofobjectlightingpatterns–shading,highlightandshadow–andtwo lighting characteristics of concern – ‘flow’ and ‘sharpness’ – provides powerfulconceptsfordevisingdistributionsoflightthatrespondtospace,formandmaterial.

Toappreciatehowtheseconceptsmightbeappliedintherealworld,wewilltakealookat the lighting for an up-market retail store. QELA offers high couture fashion in thesettingofanexclusiveartgallery,and is located inDoha,on thePearl,which isaman-madearchipelagooffthecoastofQatar.Theentrancefromashoppingmallgivesnoviewtotheinterior,givingasenseofenteringintoaprivatezone.Theinitialviewofthecentralatrium,showninFigure5.3,withitsfreestandingstaircaseconnectingthetwofloors,hasbeen designed to create a strong visual impact. Here, selected displays of beautifulaccessoriesarepresentedwithinthesettingofanartgallery,andallof thiscontainedbythecurvedformsofthearchitectureandtheoverarchingdomedceiling.

Thedesignbriefhadstatedthat“merchandisewastostandoutfromtheambienteffectwith highly controlled accent lighting”. The lighting designers, Gary Campbell andTommaso Gimigliano of dpa lighting design, proceeded to devise separate lightingsolutionsspecificallyforeachaspectoftheoveralldesign.Intheinterestsofcontrollabilityand energy efficiency, itwasdecided that lighting throughout the storewas to beLED-basedanddimmable,althoughsomeexceptionsweremadeforthejewellerydisplaysanddecorativefittings.

Figure5.3ThestrikingfirstviewoftheinterioroftheQELAboutique,Doha,wherehighqualityaccessoriesare

presentedinthesettingofanartgallery,callingforavarietyoflightingcharacteristics.InteriordesignbyUXUSDesign,

Amsterdam;PhotographybyAdrianHaddad;Lightingbydpalightingdesign.

Theimmediateimpressionisoneofabrightandlivelyspace.AMRSElevelofatleast300lm/m2 is required to give this sense of a distinctly bright space, and it can be seen thatwhilethereareareasofwhiteornear-whitesurfaces,overallreflectancevaluesarevariedand includesomequitedarksurfaces.Noteparticularly the floor,whichalthoughhighlypolished, is nonetheless highly absorptive, which will have the effect of increasing theperceived strength of the downward ‘flow’ of light. However, the high MRSE valuerequiresahighleveloffirstreflectedflux(FRF),andtoachievethiswithoutwastinglightcallsforluminairefluxtobedirectedontohighreflectance(lowabsorptance)surfaces.

Taking a closer look at the central area, Figure5.4 shows how direct flux is stronglyconcentratedontothedisplays.Thiscentralzoneis litfromtheceilingabovetheatrium,andthisinvolvesthrowsofnineortenmetres.The‘flow’oflightisstronglydownwardsand its ‘sharpness’ createsglitteringhighlightpatternson thepolishedmetalsand richlyglossy surfaces of the luxury goods ondisplay, aswell as crisp, sharply defined shadowpatterns.Theselightingpatternsareset intocontrastbythedisplaypodiums,whichlackany surface features that respond to ‘sharpness’. Their smooth, matt surfaces revealshadingpatterns,butnothighlightpatterns.

Aquitedifferentlightingdistributionisprovidedforthebackgroundtothesedisplays,which is formedby the perimeterwalls and the artworks supported on them.These arewashedbyangledoverheadlighting,whichdeliversmuchoftheFRFforthespace.Asfor

thedomedceiling,thesearesurfacesforwhichdistinctlightingpatternsarenotwanted.

Moving into the smaller surrounding areas, even more strongly accentuated displaylightingeffectsareachievedonthemannequinsasaresultofthemuchreducedambientillumination, as shown in Figures 5.5 and 5.6. ‘Flow’ directions are still verticallydownwards,andthislightingcreatesparticularlystrongshadingandhighlightpatterns.

Figure5.4QELA–Thedisplaylightinginthecentralareahasstrongdownward‘flow’,with‘sharpness’creatingcrisp

shadowandhighlightpatterns,setagainstabackgroundofartworkdisplays.InteriordesignbyUXUSDesign,

Amsterdam;PhotographybyAdrianHaddad;Lightingbydpalightingdesign.

Somesubtlechangesarerevealeduponascendingthestaircasetotheupperlevel.Warmwhiteilluminationisusedthroughoutthestore,and,asshowninFigure5.7,thissenseofwarmth is reinforced by the flames of simulated open fire. To the left of this view, thejewellerydisplaysreceivespecial treatment.Thefreestandingpodiumsincludeintegratedfibreopticdownlightsintheslimpolishedchromeringatthetop,andthesearepoweredbymetalhalideprojectorsadjustableforbothintensityandcolour.‘Sharpness’isessentialforthestronghighlightpatternsthatgivejewelleryitssparkle,andcoolwhiteilluminationisbestforviewingsilveranddiamondpieces.

LEDsources,ceilingrecessedandtrackmounted,areusedextensively,andallmaybedimmedby remotedevices. Inaddition, staff canadjustdisplay luminaires fordirection,bothpanandtilt,aswellas for intensity, froman iPad,givingthemfreerein toachievecreativelightingeffects.

Clearly theability toenvisage lighting in threedimensions iscrucial tounderstandinghow to evolve design proposals to create light fields to interact with the surfaces andobjectsthatmakeupoursurroundings.Thethreelightingpatternsprovideausefulbasisnotonlyfordescribingvisualeffectsthataproposedlightingdistributionwillachieve,butalso for thinking through the characteristic of lighting that will do the job. For thoseproposalstobeeffective,theyneedtohavephotometricvalidity.

Figure5.5QELA–Inthisdisplayarea,whichisadjacenttothecentralarea,thelowermeanroomsurfaceexitance

(MRSE)levelhastheeffectofstrengtheningtheshadingpatterns.InteriordesignbyUXUSDesign,Amsterdam;

PhotographybyAdrianHaddad;Lightingbydpalightingdesign.

Figure5.6QELA–Inthisdisplayarea,themannequinappearsisolatedbythestrongshadingpatterngeneratedbythe

selectivelighting.InteriordesignbyUXUSDesign,Amsterdam;PhotographybyAdrianHaddad;Lightingbydpalighting

design.

Figure5.7QELA–Ontheupperfloor,the‘fire’ontherightmatchesthewarmwhiteilluminationusedthroughoutthe

boutique,exceptforthejewellerydisplayareaontheleft,wherethecolourtemperatureaswellastheintensityofthe

displaylightingcanbeadjustedtosuittheitemsondisplay.InteriordesignbyUXUSDesign,Amsterdam;Photography

byAdrianHaddad;Lightingbydpalightingdesign.

Three-dimensionalilluminationdistributions

Therearedistinctdifferencesbetweenmeasuringilluminationatapointonasurfaceandat a point in space. The CIE (International Commission on Illumination) definesilluminanceintermsofincidenceatapointonasurface,andthefamiliarcosine-correctedilluminationmeterisdesignedspecificallyformeasuringthatquantity.TheCIEdefinitionsimplifies illumination into a two-dimensional concept, but this has not been achievedwithoutconsequence.

It isconventionalfor illuminationtobemeasured,calculatedandspecifiedintermsofilluminance on two-dimensional planes, such as visual task planes and wall-to-wallhorizontal working planes, and this is severely limiting for design options (Lam, 1977).Conversely, the ‘flow’ of light is a three-dimensional concept, and it involves quitedifferent thinking about lighting. Instead of planes, think of the volume of a spacecomprisingalightfieldthatfullyoccupiesthespace,andthree-dimensionalobjectswithinthe space interactingwith the light field to generate object lightingpatterns that appearsuperimposed on their surfaces. The appearance of ‘flow’ is made evident by shadingpatterns and by the strength of shadowpatterns, andmay be perceived to vary in bothstrengthanddirectionthroughoutthespace.Leavingaside‘sharpness’forthemoment,weneed to be able tomeasure the spatial distribution of illumination at any chosen pointwithinthespaceinordertoexaminethiseffect.

ConsiderthepointPasapointinspacewithitslocationdefinedbythethreemutuallyperpendicularaxes,x,y,andz,asshowninFigure5.8.Figure5.9showsasectionthroughPintheplaneofthezaxisandthepointsourceS1,whichisthesolesourceofilluminationatP.AsolidplanepassingthroughPisrotatedformaximumilluminanceonsurfaceA,andfor this condition, the distance shown at P to the perimeter of the illumination solid isproportional toEA(max).Rotationof theplane from this direction causesEA to reduce inproportiontothecosineoftherotationangle,sothatwhentheangleexceeds90º,EA=0.

Inthisway,thecircularforminFigure5.9canbeenvisagedasanilluminationsolidthatformsathree-dimensionalrepresentationofthedistributionofEA,andforthisspecialcaseof the illumination distribution due to a single point source, the illumination solid is athree-dimensional cosine distribution, represented by a sphere whose surface passesthroughthereferencepoint,andforwhichadiameterfromthereferencepointcoincideswiththedirectionofthesource.ItcanbeseenthattheilluminationdistributionaboutPistotallyasymmetric,sothatifasmallthree-dimensionalobjectisplacedatP,onesidewillbeilluminatedandtheothersidewillbeintotaldarkness.Thisilluminationdifferenceonoppositesidesofanobjectisofinterest.If,insteadofrecordingthedistributionofEA,werecordthedistributionof(EA–EB), that is tosay, thedifferenceonoppositesidesoftheplane,thesolidwouldbeunchangedbecausewhensurfaceAisfacingawayfromS1,(EA–EB)wouldhaveanegativevalue.

Figure5.8ThepointPislocatedattheintersectionofthex,yandzorthogonalaxes.Thexandyaxesareinthe

horizontalplane,andthezaxisisvertical.Unlessotherwisespecified,itisconvenienttoassumeadirectionofviewfrom

they-direction(‘eye’direction),sothatxis‘a-cross’.Whileanyotherviewdirectionmaybepossible,thissimple

conventiontendstoavoiderrors.

Figure5.9Thethree-dimensionalilluminationdistributionaboutpointPduetothesmallsourceS1isdefinedbya

sphericalilluminationsolid,wherethelengthofEA(max)isproportionaltotheilluminanceonsurfaceAwhennormalto

thedirectionofS1.

Figure 5.10 shows the effect of adding a second point source S2. In this case, the bluecontours show parts of the illumination solids for the individual sources, butwhere thesolidscoincide,thevalueof(EA–EB) isshownbytheredcontour.ThevaluesofEAmax

andforS1andS2areshownasvectors,andthevalueoftheresultantvector,(EA–EB)max,is given by completing the vector parallelogram. It can be seen that the red contour issimilartotheilluminationsolidforthesinglepointsource,meaningthatthedistributionofilluminance difference on opposite sides of the plane is identical to that produced by apointsource.Ifthishappenswhenweaddasecondsource,itwillhappenwhenweaddathird,orfourth…oraninfinitenumberofsources.Wehaveestablishedthepointthatatany illuminated point in space, the distribution of illuminance difference in oppositedirections (EA – EB) may be represented as an illumination vector. This concept isattributed to Professor A.A. Gershun, whose book, The Light Field, was published (inRussian)in1936.

InFigure5.11,wemovefromhypotheticalsituationstoarealsituation.Thebluecontouris typical of an illumination solid that might occur in an indoor location illuminatedpredominantly from overhead, but with a sideways bias as might occur near a darkcolouredwall.Thecontourissmoothbecauseitisthesumofsphericalsolidsduetoeveryluminouselementsurroundingthemeasurementpoint.Illuminationsolidcontourscannotdisplaysharppeaksor troughs.Theredcontour is thedistributionof (EA–EB), and theplanepassingthroughPhasbeenrotatedaspreviously,butthistimetheaimhasbeentofindthedirectionthatgivesmaximumilluminancedifferenceonoppositesidesofthe

Figure5.10TheilluminationsolidisnowthesumofcomponentsolidsduetosourcesS1andS2,butthedistributionof

EA–EBisasphericalsolid,identicalinform,butnotmagnitudeordirection,totheilluminationsolidduetoS1alone.

Figure5.11Theilluminationsolidatapointinaspacewherelightarrivesfromeverydirection,butpredominatelyfrom

overheadalthoughwithasidewaysbias.Despitetheirregularityoftheilluminationsolid,thedistributionEA–EBis

definedbyasphericalsolididenticalinformtotheilluminationsolidproducedpreviouslybyS1.

plane,whichmaynotcoincidewiththemaximumvalueoftheilluminationsolidcontour.Rotationoftheplanefromthisdirectionwouldcause(EA–EB)toreduceinproportiontothecosineoftherotationangle,sothatwhentherotationangleequals90º,EA=EB.Thedistributionof(EA–EB)isathree-dimensionalcosinedistributionidenticalinformtotheilluminationsolidduetoasinglepointsource.AsshowninFigure5.8,thex,yaxisliesinthehorizontalplane,andthezaxisisvertical.

Wearenowinapositiontoanalysetheilluminationdistributionaboutapointinspaceintoitstwocomponents.InFigure5.12themaximumvalueof(EA–EB)andthedirectioninwhichthisvalueoccursdefinetheilluminationvectorE.ForanyplanepassingthroughP,theilluminancedifferenceonoppositesurfacesequalsthevectorcomponentontheaxisnormaltotheplane.ForthehorizontalplanethroughP,E(z)=E(z+)–E(z-),andsimilarly,foraverticalplanethroughPnormaltothexaxis,themagnitudeoftheilluminationvectorcomponentisE(x).Notethatthesymbolforavectorisshowninboldtype,andwhilethisdistinctionisclearlyindicatedinprint,inmanuscriptitismadebyasmallarrowovertheE.Notethatavectorisdefinedintermsofbothmagnitudeanddirection.Thedistributionof the vector component is defined by the three-dimensional vector solid,which, aswehavenoted,isalwaysasphericalcosinedistributionwithitssurfacepassingthroughPanditsdiameterequaltothevectormagnitude.

Figure5.12Themagnitudeanddirectionof(EA–EB)maxdefinestheilluminationvector,whichisdepictedasanarrow

actingtowardsthepoint.Thevectorinturnisdefinedbyitscomponentsonthex,y,andzaxes.Thevectorsolidaccounts

entirelyfortheasymmetryoftheilluminationsolid.

InFigure5.13,thevectorsolidhasbeensubtractedfromtheilluminationsolidandwhatremains is a three-dimensional solid that is divided by the axis normal to the vectordirection.ThissolidhaspointsymmetryaboutP,thatistosay,foranyaxisthroughP,thedistance to the contour of this solid in one direction equals the distance in the oppositedirection. This is the symmetric solid, and while it may depart from uniformity, theilluminancedue to the symmetric solid ~E inanydirection fromP is equal to ~E in theopposite direction. If a plane passing through P is rotated, for every orientation, theilluminancevaluesonoppositesidesoftheplaneduetothesymmetricsolidwillbeequal.Inotherwords,itisthesolidforwhichEA–EB=0foreveryorientation.

Figure5.13Ifthevectorsolidissubtractedfromtheilluminationsolid,whatisleftisasolidthatissymmetricalinevery

directionaboutthepoint.Thisisthesymmetricsolid.

Inthisway,wearriveatthefollowingconclusions:

1. That at any illuminated point in space, the three-dimensional distribution ofilluminancemaybedefinedbyanilluminationsolid.

2. The illumination solid is the sumof twocomponent solids: thevector solidandthesymmetricsolid.

3. Thevectorsolidisasphericalcosinedistribution,andisdefinedbythemagnitude

and direction of the illumination vector E. The illuminance distribution at thereference point P represented by the vector solid is identical to the distributionthatwouldbeproducedbyasinglecompactsourcelocatedinthevectordirection.

4. The symmetric solid has the property that, for any plane passing through P, itproducesequalilluminance~Eonoppositesides.

5. The visible characteristics of the illuminance distribution over the unobstructedsurfaceofathree-dimensionalobjectthat issmall inrelationtothesurroundinglight field may be analysed as the sum of distributions due to the vector andsymmetriccomponents,onebeingentirelyasymmetricalaboutthemeasurementpoint,andtheotherentirelysymmetrical.

6. Twospecialcasesmaybenoted:

For a single point source, the illumination solid is coincident with thevectorsolid,andthesymmetriccomponent~E=0.For an integrating sphere, the illumination solid is coincident with thesymmetric solid, comprising ideally a spherical distribution centred at P,andtheilluminationvectorE=0.

To all of the above, I wish to add a personal observation. The concept that the spatialdistributionofilluminationatanyilluminatedpointinspacecomprisestwocomponents–oneofwhichisentirelyasymmetricaboutthepointandcouldbeproducedbyacompactsourceinthedirectionofthevector,whiletheotherisentirelysymmetricaboutthepoint–isnotintuitive.Itemergesfromamathematicalanalysis,andis,inmyopinion,themostremarkable finding to emerge from the study of illumination engineering. It provides auniquedesigninsight,andifyoulookforit,youcanseeit.

Illuminationsolidsandthe‘flow’oflight

LookbacktoFigure5.2,andnoteparticularlythechangingappearanceofthemattwhitesphereinthethreelightingconditions.Thisobjectformsadifferentshadingpatternwitheachvariationofthelightfieldand,everytime,theappearanceoftheshadingpatterncanbedescribedintermsoftheapparentstrengthanddirectionofthe‘flow’oflight.Equally,itmaybedescribed in termsofdifferent relationshipsof theasymmetricand symmetriccomponents of the illumination solid.We have here the basis of a means for assessinglightingaccordingtoitspotentialtoinfluencetheappearanceofthree-dimensionalobjectsthroughthecreationofshadingpatterns,whichinturn,maybedescribedintermsofthe‘flow’oflight.

Ifanobjectissmallenoughinrelationtoitssurroundingenvironmentforustobeabletoexamineitsilluminationbyconsideringtheilluminationdistributionatapoint,thenwecanthinkofeveryelementvisiblefromthepointtobecontributingitsownmini-vectoratthepoint.Wehavetwoalternativewaysofsummingthesemini-vectors.Ifwesumthemindividually,wehavetheilluminationsolid.Ifwesumtheiroppositedifferences,thesumof these individual vectors is always a single vector that defines the magnitude anddirection of the (asymmetric) vector solid. Thedifference between the illumination solidandthevectorsolidisthesymmetricsolid.Itshouldbeapparentthatifthevectorsolidislargeinrelationtothesymmetricsolid,thenthe‘flow’oflightwillappeartobestrong.Itmight seem, therefore, that the ratioofasymmetric to symmetric solidswouldprovideauseful index of this effect, but it would have a range from zero to infinity, which isinconvenient. In mathematics, a vector quantity is one that has both magnitude anddirection,whileaquantitythathasmagnitudeonlyistermedascalar.ItwaswiththisinmindthatJ.A.Lynesproposedtheconceptofscalarilluminance,whichisdefinedintermsoftheaverageilluminanceoverthewholesurfaceofasmallspherecentredatareferencepoint(Lynesetal.,1966).Itfollowsthatforanyilluminationsolid,thescalarilluminancewillbe the sumof contributions from thevectorand symmetric solids.Thecontributionfromthesymmetricsolidwillbeequaltothesymmetricilluminance~E,andfromFigure5.14 is can be seen that the asymmetric solid will contribute one-quarter of the vectormagnitude,sothatscalarilluminance:

This enables us to specify the apparent strength of the ‘flow’ of light in terms of thevector/scalarratio:

VSRhasascalefromzero(theintegratingspherecondition)tofour(thepointsourceinablack environment). Research studies in a face-to-face situation indicated preference for

VSRwithintherange1.2to1.8(Cuttleetal.,1967)andthisfindinghasmorerecentlybeencorroboratedbyProtzmanandHouser(2005).Table5.1givesanapproximateindicationofhowassessmentsoftheperceivedstrengthof‘flow’arelikelytovarywiththeVSR.

Figure5.14In(a),asmallsourceSprojectsluminousfluxofFlmontoadiscofradiusr,producingasurfaceilluminance

E=F/(π.r2).In(b),thediscisreplacedbyasphereofradiusr,givingasurfaceilluminanceE=F/(4π.r2).

Regardlessof thenumberof lightsourcespresent, theasymmetriccomponentresolvesintoacosinedistributiondefinedbytheilluminationvector,andprovidingthattheVSRissufficient for the ‘flow’ direction to be apparent, its direction coincideswith the vectordirection.Therearetwoalternativewaysofdefiningthevectordirection.Oneistospecifyvectoraltitude(α)andazimuth(φ)angles,whereFigure5.15(a)shows:

Thereismorethanonewayofspecifyingtheazimuthangle,andFigure5.15(b)showsφmeasuredanticlockwisefromthey-axis,asthisisoftentakentorepresentthedirectionofview.Careneedstobetakentocopewiththefull360degreesofrotation.

Table5.1Vector/scalarratioandtheperceived‘flow’oflight

Figure5.15(a):VerticalsectionthroughPshowingilluminationvectoraltitudeangleα,and(b):Horizontalsection

throughPshowingazimuthangleφofthehorizontalvectorcomponent.

Anotherwayistospecifythevectordirectionintermsofaunitvector,whichassumesthevectortohaveunityvalueandexpressesthedirectionintheform(e(x),e(y),e(z)),wheretheunitvectorcomponentonthexaxis:

andsimilarlyfore(y)ande(z). Itcanbeseenthateachoftheseunitvectorvaluesdoesinfactspecifythecosineoftheanglethatthevectorformswiththeaxis.Whilecareneedstobe taken over the signs of the unit vector components, this concise form of notation isrecommendedaslargelyavoidingtheconfusionsthatarelikelytooccurwhendealingwithanglesgreaterthan2π.

The previously cited research into preferences for face-to-face viewing (Cuttle et al.,1967)founddistinctpreferenceforvectoraltitudesintherange15to45degrees,or0.25<e(z)<0.7.Evenmoredistinctwas the identificationofadownward ‘flow’of lightas theleastpreferredcondition,forwhichα=90degreesande=(0,0,1).Forface-to-faceviewingsituationswhereoverheadlightingisunavoidable,VSRshouldbekepttoalowvalue.

Inthisway,thecharacteristicsofathree-dimensionaldistributionofillumination,asitmayaffect theperceivedstrengthanddirectionof the ‘flow’of light,canbespecified interms of simple photometric quantities. Procedures for predicting and measuring thesequantitiesaredescribedinthefollowingchapter.

The‘sharpness’ofillumination

While ‘flow’ relates to the appearance of the shading patterns and the density of theshadow patterns, ‘sharpness’ relates to the appearance of the highlight patterns and thecrispnessoftheshadowpatterns.LookbacktoFigure5.2,andappreciatehowdifferentlythesetwolightingconceptsappear.

Athoughtexperiment

Once again, you must clear your mind of what you expect to experience and let yourimagination takecontrol. InFigure5.16,a surface is illuminatedbyadiffusingdisc lightsource.TheilluminanceatPisgivenbythediscsourceformula(SimonsandBean,2001):

Thesubtenceangleofthesource,α,mayhaveanyvaluefromadegreeortwo,forwhichthe sourcewould be close to the hypothetical point source, up to 180 degrees, atwhichpoint the source becomes a uniform diffusing hemisphere.At the point P,we place thecomparisonpanelshowninFigure5.17,whichcomparestwomaterials,asampleofblackglassandamattwhitesurface.ThispanelwasoriginallyproposedbyJ.A.Worthy(1990)toexplainhisownresearchintothisaspectoflighting.

Imaginethataswevaryα,thesourceluminanceadjuststomaintaintheilluminanceEP

ataconstantvalueof100 lux.Theappearanceofthemattwhitesurfacewillnotchangewhilewemakethisvariationbecauseitsluminanceremainsconstant,buttheappearanceof the black glass sample will undergo radical changes. If we start from the luminoushemispherecondition(α=180degrees),theglassappearstohaveagreycastoverit.Asαisreduced,thiscastshrinkstobecomeanimageofthediscsource,andaroundtheimagewe see the blackness of the glass. As the image continues to shrink, it increases inbrightness until it becomes an intensely bright, small dot that appears sharply definedagainsttheblacknessoftheglass.Thelightingnowhas‘sharpness’,andthisisrevealedbythe appearance of the glass, not the white disc. This effect is similar to the highlightpatternsonthewhiteandblackspheresshowninFigure5.2(a)to(c).

Figure5.16ThepointPisonasurface,andisilluminatedbyadisc-shapedsourcethatisnormaltothesurfaceandof

angularsubtenceα.

Figure5.17Thiscomparisonsurfacehastwomountedsamplesthatresponddifferentlytothediscsource.AfterWorthy

(1990).

Ifwerearrangethediscsourceformulato ,Figure5.18showshowthevalueof

LShastobevariedtomaintainEP=100lx.Itcanbeseenthatarelativelylowluminancevalue satisfies over a substantial angular range, but as the source becomes smaller thanabout30degrees,itsluminancehastobeincreasedquitesharply.However,itiswhenwegetdowntothereallysmallsourcesthatthesourceluminancehastoescalateinordertoprovidetherequiredilluminance,asshowninFigure5.19.

It can be seen that when α = 180 degrees, LS = 100/π, or approximately 31 cd/m2.Reducingαto90degreesrequiresLStobedoubled,butthisisstillalargesource.Whenαcomesdownto30degrees,LShastobeincreasedto475cd/m2,andat10degreesithastobeover4000cd/m2.However,itiswhenwegettoasourcesubtendinglessthan10degrees

thattheluminancevaluereallyclimbs.Asourcesubtendingjust1.0degreehastohavealuminanceofnearlyhalfamillioncd/m2todeliverjust100lux.

Figure5.18Asthesubtenceofalargediscsourceisreduced,thesourceluminancerequiredtomaintainanilluminance

valueof100luxincreasesrapidlyassubtencefallsbelow30degrees.

Figure5.19Forsmallsources,theincreaseinluminancerequiredtomaintain100luxincreasesdramaticallyforsubtence

angleslessthan3degrees.

Imaginenowthatyouwant todeliveragiven illuminanceEtgt ontoa three-dimensionaltarget, and you have selected a location for the luminaire at distance D. The required

sourceluminousintensityIS=Etgt×D2cd(foratwo-dimensionaltargetyouwillneedtotakeaccountoftheangleofincidence),soyouscrollthroughtheluminairemanufacturers’websiteslookingforaspotlightwithsuitableperformance.Whilemostmanufacturerswillgiveyouintensitydata,theyareunlikelytogivesourceluminancevalues,butclearlythiswill affect substantially the perceived ‘sharpness’, so you will need to check this foryourself. From the sourcedimensions,workout the luminousareaASprojected towardstheobject,andthenthesourceluminanceLS=IS/AScd/m2.Thisisyourfirststeptowardsassessingthepotentialfor‘sharpness’.

Highlightcontrastpotential(HCP)

Generally, smoothdielectric (non-electroconducting)materialshave specularcomponentsof their total reflectance of around 4 per cent (although for electroconductingmaterials,such as polished metals, it can be much higher). Typically then, the luminance of thereflectedhighlightseenonaglossysurfaceLhl=0.04LS,whereLSisthesourceluminance.Thevisibilityofthehighlightdependsprimarilyontheluminancecontrastofthehighlightagainstthebackgroundonwhichitisseen.

Thehighlightcontrastpotential(HCP)isameasureoftheextenttowhichalightsourcemayprovide a visiblehighlight. For thiswe ignore light thatmaybe reflected from thesurrounding environment (the highest possible highlight contrast will occur in a blackenvironmentwhereSistheonlysourceoflight)andconsideradielectrictargetsurfacetgtthat has a reflectance ρtgt (which includes the 0.04 specular component) illuminated bysourceS,thenthehighlightluminance:

Lhl=0.04LS

Andtheluminanceofthetargetsurface:

Applyingthediscsourceformulaaspreviously,

So,

ThenHighlightContrastPotential:

NotethatalthoughLSdoesnotappearinthisformula,thesourcesubtenceangleisinthere.

Iftherearenoothersourcesoflight,thentheonlyotherfactordeterminingHCPistargetreflectance.Inreality,othersourcesoflightwillbepresentandtheywillhavetheeffectofreducingthehighlightcontrast,sothatforagivenlightsource,thisisanexpressionforitsmaximum potential to provide highlight contrast. Figure 5.20 shows how theconspicuousness of highlights is dependent on lowdiffuse reflectance, such as the blackglasssampleinthecomparisonpanel,aswellassmallangularsizeofthelightsource.

Figure5.20HighlightcontrastpotentialHLCforthreevaluesoftargetreflectance,representinglow,mediumandhigh

surfacelightness,andarangeofsourceangularsubtenceangles.

Generallightingpracticeseekstoavoidspecularreflections,identifyingthemas‘veilingreflections’, but designers should distinguish between highlights and veiling reflections.Whenweconsideredtheappearanceofthecomparisonpanel(Figure5.17)inthethoughtexercise, the effect of the large source was to create a grey cast over the glossy blacksurface, reducing its blackness but giving no hint of its glossiness. This was a veilingreflection.However,whenthesourcesubtencewasreduced,thespecularreflectionbecameahighlightpattern,seenincontrastagainst theundiminishedblacknessof theglass. If ithad theeffectof reducing thevisibilityof surfacedetail, this couldeasilybeavoidedbyheadmovement, andmeanwhile, the smooth, shiny surface of the glasswould be givenvisual emphasis. The ability to create highlight patternswhen andwhere required is animportantskillinthelightingdesigner’stoolkit.Again,mathematicalanalysisofareadilyobservedcharacteristicoflightinggivesinsightintoitsoccurrencethatisnotintuitive.

ThatwehaveanexpressionforHCPdoesnotmeanthattargetvaluesshouldbeset,or

thatwehaveanother factor tobecalculatedandmeasured.Whatmatters is thatweareabletoidentifythephysicalparametersonwhichHCPdepends,andthisenablesdesignerstoexercisecontrolover theaspectsof lightingthatare influential. In thesectionentitled‘The three object lighting patterns’ (page 66) itwas noted how the appearance of someobjectscanbebroughtalivebyhighlightpatterns,whileothersbenefitfromtheircompleteabsence. The usefulness of HCP lies in enabling design decisions to be guided byunderstandingoftheconditionsthatgovernthe‘sharpness’oflighting.

Theappearanceofshadowpatterns

Shadowpatternsmight seemtobe thesimplestof the three typesof lightingpatterns tocome to termswith, but this is not so.While the perceived strength and direction of ashadowpattern relates to theVSRand the vector direction, its ‘sharpness’ relates to theHCP.Inthisway,theappearanceoftheshadowpatternswithinaspacevarywithboththeoverallimpressionofthe‘flow’oflightandtheperceived‘sharpness’oftheillumination.

LookoncemoreatFigure5.2,andnotetheroleof ‘sharp’shadowpatterns,boththosecastontoobjectsandthosecastontothebackground,increatingtheappearanceofdepthandasenseof‘crispness’withintheoverallscene.Ashasbeendiscussed,achievingtheseeffectsmaysupportdesignobjectives,as inFigure5.4,or thesituationmaycall for theiravoidance, as inFigure5.7. It is all amatter of being able to visualise the space and itsobjects in light, and being able to control lighting patterns to reveal, or to subdue or toemphasisesurfaceattributes.

Figure 5.21 shows the formation of the penumbra, the extent of which is inverselyrelatedtotheperceived‘sharpness’ofashadowpattern.Meanwhile,theapparentdensityoftheumbraisdeterminedbytheVSR,andinthiswaywecanseehowtheconceptsofthethreeobjectlightingpatterns–theshading,highlightandshadowpatterns–areconceptsthatcanreadilybevisualisedanddiscussedwithclientsandotherdesignprofessionals.Ontheotherhand, theconceptsof ‘flow’and ‘sharpness’of illuminationprovidemeans fordescribing illumination in termsof itspotential to createobject lightingpatterns, andassuch,theycanenablemembersofadesignteamtobuildasharedunderstandingofthree-dimensionallightfieldsthatfillspacesandinfluenceappearancesofeverythingwithinit.That both of these concepts – ‘flow’ and ‘sharpness’ – can be specified in terms ofphotometric concepts – vector/scalar ratio, vector direction, and highlight contrastpotential–enablesthemtobedescribedwithconfidencethattheywillbeprovided.Themeansfordoingthisaredescribedinthenextchapter.

Figure5.21Lightsourcesofsmallersubtenceangleproducelesspenumbra,increasingthe‘sharpness’ofthelighting.The

perceiveddensityoftheumbraisdeterminedbythestrengthofthe‘flow’oflight.

References

Cuttle, C. (1971). Lighting patterns and the flow of light. Lighting Research andTechnology,3(3):171–189.

Cuttle, C., W.B. Valentine, J.A. Lynes and W. Burt (1967). Beyond the working plane.Proceedingsofthe16thSessionoftheCIE.WashingtonDC,PaperP67–12.

Gershun,A.A.(1939).Thelightfield.(TranslationbyMoonP.G.Timoshenko.)JournalofMathsandPhysics,18:51–151.

Lam,W.M.C. (1977).Perception and Lighting as Formgivers for Architecture. NewYork:McGraw-Hill,Inc.

Lynes, J.A.,W.Burt,G.K. JacksonandC.Cuttle (1966).The flowof light intobuildings.TransactionsoftheIlluminatingEngineeringSociety(London),31(3):65–91.

Protzman,J.B.andK.W.Houser(2005).Ontherelationshipbetweenobjectmodelingandthesubjectiveresponse.Leukos,2(1):13–28.

Simons, R.H. and A.R. Bean (2001). Lighting Engineering: Applied Calculations.Oxford:ArchitecturalPress.

Worthy, J.A. (1990).Lightingqualityand light source size.JournaloftheIlluminatingEngineeringSociety(NewYork),19(2):142–148.

6DeliveringtheLumens

Chaptersummary

Throughoutthisbook,theaimhasbeenthatalightingdesignershoulddeveloptheskilltovisualise thedistributionof lightwithinthevolumeofaspace intermsofhowitaffectspeople’s perceptions of both the space and the objects within it. These envisioneddistributions of light comprise reflected light, while the distributions that the designercontrolsarethedirect lighttobeprovidedbythelightinginstallation.Furthermore,theyare, for themostpart, three-dimensionalvariationsof the light field, and the conceptofcubic illumination is introduced toprovide abasis forunderstanding themand enablingpredictive calculations. Procedures are explained for specifying lamps and luminaires ofcorrect performance, as well as controls to enable installations to respond to daylightavailability. Measurement procedures are described for ensuring that design objectiveshavebeenachieved,andtwoCubicIlluminationspreadsheetsareintroducedthatperformthecalculationsautomatically.

Lightingcalculations

Lighting calculations do not solve problems. Their purpose is to enable a designer tospecify a layout of lamps, luminaires and control circuits with a reasonable level ofconfidence that it will create an envisioned appearance. No matter how well thoughtthrough the envisioned appearance might be, it will not be achieved by guesswork or‘hoping for the best’. Lamp wattages, luminaire spacings and beam spreads need to becorrectforthedistributionandbalanceofthelightingtolookright.

Even so, some common sense needs to be applied. Photometric laboratories do, veryproperly,worktohighlevelsofprecisiontospecifytheperformancesoflightingproducts,butwhilelightingdesignersneedtohaveconfidenceinthereliabilityofthedatathattheyareworkingwith, precision inwhat theyprovideneeds to benobetter thandifferencesthatusers(perhapscriticalusers)arelikelytonotice.

Theaimhastobethataclientwhohashadalightingdesignproposaldescribedintermsof perception-based objectives will be satisfied that their expectations have been met.Specifyingandpredictingperformanceintermsoflightingmetrics,followedbycheckingtheactualperformancelevelsachieved,arenecessarypartsofthedesignprocess,butneednotundulyconcernclients.

Lighting design should not be thought of as a linear process, but nonetheless, thischapterfollowsasequencethatrelates,quitesensibly,toarationaldesignprocedure.

Meanroomsurfaceexitance,MRSE

Starting from how brightly lit, or dimly lit, the space is to appear, the designer decidesupon a level of ambient illumination and specifies this in terms of mean room surfaceexitance,asexplainedinChapter2andreferringtoTable2.1.Thegeneralexpressionis:

whereFRFisfirstreflectedflux:

andAαistheroomabsorption:

where:

Es(d)=directilluminanceofsurfaces(lux)As=areaofsurfaces(m2)ρs=reflectanceofsurfaces

Estimatingthereflectanceofasurfaceisnotassimpleasitmightseem.Patternedsurfacesareparticularlydifficult,butreasonablyreliablemeasurementscanbemadebyattachinganinternallyblackenedcardboardtubetoalightmeter,asshowninFigure6.1,andtakinga reading of the surface in question, taking care to avoid specular reflections. Then,without moving the meter, slide a sheet of white paper over the surface and take acomparative reading.Good qualitywriting paper typically has a reflectance around 0.9.Alternatively,paintmanufacturersoftenquotereflectancevaluesfortheirproducts,andapaintcolourswatchcanbeusedtomakematchestosurfacecolours.

It is sometimes useful to be able to determine the equivalent reflectance of a cavityplane,ρeq,suchasthatofaluminaireplane,inwhichcasetheupperwallsandtheceilingformthecavity.StartbycalculatingtheratiooftheareaofthecavityplaneAcptotheareaof thecavity surfacesAcs, and theaverage reflectanceof surfaceswithin thecavity,ρav,then:

Whenusingformula6.3tocalculatetheroomabsorptionvalue,Aα,itisoftenconvenienttouseAcpandρeqvalues.Afterthat,thetotalfirstreflectedfluxrequiredtoprovidetheMRSEvalueiscalculated:

ThisFRFvalueisthenumberof ‘firstbounce’ lumensthathastobeprovidedtoachievethe design value of MRSE, and ways of accounting for this value are explained in‘Illuminationhierarchydesign’(page32).ItneedstobenotedthatroomsurfacereflectancevaluesaremuchmorestronglyinfluentialinMRSEcalculationsthaninconventionalHWPcalculations, and it is important that designers work with realistic values. The bad oldpracticeofassumingroomsurfacereflectancevaluesisarecipefordisaster.

Figure6.1Measuringsurfacereflectance,usinganinternallyblackenedcardboardtubefittedoveranilluminancemeter.

Comparativereadingsaretakenofthesurface,avoidingspecularreflections,andofasheetofwhitepaper.

Illuminationhierarchyandtargetilluminancevalues

Afterselectingthedesignvaluefortheambientillumination,thelightingdesignerdecidesupontheilluminationhierarchy,whichdeterminesthedistributionofilluminationwithinthe space. This is achieved by providing direct illumination selectively onto specificsurfaces and objects, as demonstrated in Boxes 3.1 and 3.2, using the IlluminationHierarchyspreadsheet.ThescheduleofTAIRvaluesisthefirstcrucialstatementofdesignobjectives.The principal tool for devising this distribution is the classic inversesquarecosinelaw(sometimesreferredtoasthe‘point-to-point’formula),whichisstatedas:

Where

EP=illuminanceatpointofincidenceP(lux)IS=luminousintensityduetolightsourceS(candelas)θ=angleofincidenceD=distancefromsourceStopointP(metres)

ThisstatementofthelawisoftenaccompaniedbyadiagramofthesortshowninFigure6.2inwhichthewholeissueofprovidingdirectilluminationisreducedtotwodimensionsand,bydefault,itoftenhappensthattheplaneofincidenceisassumedtobethehorizontalworkingplane.

Oncewebecomeconcernedwithprovidingilluminationontoplanesthatpeopleactuallylookat,wearelikelytofindourselvesdealingwithsituationsthatarefarmorelikeFigure6.3.HerepointPisonaverticalsurfacewhichmaybeofanyorientation,andilluminatedbyadirectionallightsourceS.ThelocationofSrelativetoPisdeterminedbydimensionsX,YandZ(whichmaybepositiveornegativeaccordingtodirection),andthen,dependingoncircumstance,itcanbeconvenienttothinkofYasbeinginthe‘eye’direction,sothatXis‘a-cross’andmeasurespositivetotherightandnegativetotheleft,whileontheverticalaxis,Zmeasurespositiveupandnegativedown. It canbehandy tokeep thispicture inmindasitcanavoidalotofconfusionwhenitcomestoanalysingmeasureddata.

The performance of the light source S is indicated by its distribution of luminousintensity,IS, specified incandelas (cd),andoftenthiscanbesimplified intotwoessentialitems of performance data. These are the intensity value on the beam axis, which forhistoricalreasonsisstilloftenreferredtoasthecentrebeamcandlepower(CBCP),andthebeamangle,B,forwhichthebeamedgeisdefinedbytheangleatwhichintensitydropsto50percentoftheCBCPvalue.Forexample,ifCBCP=3000cdandB=12º,thenat6ºtoeach side of the beam axis, intensity IS = 1500cd. Any light emitted outside the beamshouldberegardedas‘spill’,andblockedbylouvresorbaffles.

FirstwewillconsiderhowtousethesedatatocalculatetheilluminanceEPatpointP.Thenwetakenotethat,providingthebeamisconical,itformsanellipticalpatternonthesurface,withminor andmajor axes q′ and q″. These tooweneed to be able to predict,particularly as we often need to use several light sources to build up a pattern ofoverlappingellipsestoprovidecoverageoveratargetsurface.Also,itneedstobekeptinmindthat,followingtheproceduresdescribedintheforegoingchapters,thepurposemay

Figure6.2Applicationofthepoint-to-pointformula,Ep=IScosθ/D2,fordeterminingtheilluminanceatpointPona

horizontalplane.ISisthebeamcentreluminousintensity.

Figure6.3DeterminingtheilluminanceatpointPonaverticalplane,andthebeampatternformedontheplane.Bisthe

beamangle,whichdefinestheconeoverwhichluminousintensityequalsmorethan50%ofIS,theCBCPvalue.

be to provide the direct illumination required to provide the target/ambient illuminanceratios(TAIRs)determinedfortheilluminationhierarchy.Keepinmindthatthismeansthatweneedtostartoffknowingtheilluminancethatisrequired,andtheaimistodeterminethe luminous intensity to be provided. As we look though suppliers’ data for suitableluminaires,we check the suitability of potential luminaires bynoting theirCBCPandBvalues.

ApplicationoftheinversesquarelawtothesituationshowninFigure6.3callsforsomecarefulexaminationofthesituation.Pythagoras’theoremtellsusthatthedistanceDofPfromSisgivenby ,butwhatseemsalittlemoretrickyinthisthree-dimensional situation is finding thecosineof theangleof incidence,θ.This is theanglethatthebeamaxisformswiththeyaxis,whichisthenormaltothesurfaceatP,sothatcosθ=Y/D.LookbacktoFormula6.6,anditcanbeseenthatwecanrewritetheformulaforcalculatingtheilluminanceatPas:

Take good note of this ‘D to the 3’ expression. By eliminating cosθ we have greatlysimplifiedthe‘point-to-point’calculations,andwewillmakeuseofthisformula.Itmaybe

rearrangedtogivetherequiredsourceintensitytoachieveEP:

Now we turn our attention to the elliptical beam pattern formed on the surface. Thispatternbecomescrucialwhenweareselectivelyilluminatingachosensurfacefromsomedistance.Accordingtotheshapeofthesurface,itmaybenecessarytobuildupcoverageofoverlappingellipsesusingseverallightsources.

ReferringagaintoFigure6.3,itcanbeseenthat:

qˊ/2=D.tan(B/2)

andunlessBislarge,inwhichcasethebeamfluxmethoddescribedinSection6.7islikelytobemoresuitable,thisexpressionmaybeapproximatedto:

Notealsothatcosθ=Y/D,sothat:

Thesehandyexpressions,whichenableilluminationtobeprovidedontoverticalsurfacesevenlyandwithminimalspill,aresummarisedonFigure6.3.InclinedsurfacescanalsobedealtwithbykeepinginmindthatYisthedimensiononthesurfacenormalatP.Forthemathematically agile, an even more versatile approach employing vector algebra isavailable(Cuttle,2008).

Astheperimeterofthebeampatternellipseisdefinedbythecontourwhereluminousintensity drops to half the beam axis value, even coverage of a surface is achieved bybutting ellipsesup edge to edge. It is reasonable to assume that the average illuminancewithinanindividualellipseis75percentofthecalculatedEPvalue.

TheD/rcorrection

Thereisalingeringconcern.Theinversesquarecosinelaw isreferredtoasthe‘point-to-point’ formula for a good reason. It is based on the concept of a point sourceilluminatingapointonasurface,andofcourse,pointsourcesarehypotheticalastheyhavenoarea.Itmaybeshownthaterrorwillbenotmorethan1percentifthedistanceDisatleastfivetimesthemaximumdimensionoftheluminaired,andonthisbasisitisgenerallyrecommendedthatuseof‘point-to-point’formulaeisrestrictedtosituationswhereD>5d.Inpractice,many situationsmayoccurwhereDwillbe less than5d,particularlywherethereisaneedtoget inclosewiththelightingor large,diffusinglightsourcesarebeingused. For these situations various ‘area source’ formulae have been published, but theytend to be cumbersome.Amore simple solution is to staywith the prediction formulaebasedontheinversesquarecosinelawandtoapplytheD/rcorrection.

Figure6.4ThepointPisilluminatedbytwoalternativesources,S1beingapointsourceandS2aluminousdiscofradius

r.BothsourcesareatdistanceD.

Figure 6.4 shows a point P illuminated alternatively by two light sources, both atdistanceDandnormal to the surfaceatP.S1 isahypotheticalpoint source,andS2 isadiffusingdiscsourceofradiusrandnormaltothedirectionofP.ForS1,theilluminanceatPis:

Es1=Is1/D2

ForS2,weapplythediscsourceformula(SimonsandBean,2001):

whereMS2istheexitanceofsourceS2.

Foradiffusingsource,theluminousintensitynormaltothesurfaceequalstheluminousfluxoutputdividedbyπ,sothat:

So:

IfweassumethatIS1=IS2,andweapplythemoresimpleES1expression tocalculate theilluminanceduetoanareasource,theilluminancevaluewillbeoverestimated.Thiscouldbeovercomebyapplyinga(D/r)correction:

NotethatDisthedistanceStoP,andristheradius,orhalfthemaximumdimension,ofthelightsourcenormaltothedirectionofP.

ThevalueofC(D/r)canbereadfromFigure6.5andapplieddirectlytocalculationsusingFormulae6.6or6.7.ItmaybenotedthatthevalueofD/rneedstoreducetoalowvaluebeforethecorrectionmakesmuchdifference,infact,thesourceradiushastoapproachthedistancebefore theerrorbecomes really significant.Added to that, itmaybenoted thatFormula6.11assumesthatthelightsourceisaluminousdiscofdiameter2rthatisnormaltothedirectionofP,andthisdefinesthe ‘worstcase’situation. Ifa linearsource isusedinsteadofadiscsource,theC(D/r)correctionwilloverestimatetheilluminancereductionbyabout one-third. The reality is that when we assume a point source we are tending tooverestimate illuminance, and when we assume a disc source, we are tending tounderestimate.Unlessthesourceislargeinrelationtothedistance,sensiblejudgmentwillsuffice.

Inthisway,Formula6.11enablessimple‘point-to-point’expressionstobeappliedforawiderangeofpractical lightingsituations,and theC(D/r) correctionmaybeappliedwithreasonableconfidencewherevertheaimistoilluminateatwo-dimensionalsurfacewithasource that is other than small in relation to its distance from the surface. Practicalexamples might include selected room surfaces, or pictures displayed on them, orfreestandingpanels, aswell as any surfaces forwhich target/ambient illuminance ratios,TAIRs,havebeenspecifiedaspartoftheilluminationhierarchyplanning.

Figure6.5ThecorrectionfactorC(D/r)tobeappliedtopointsourceilluminationformulaetoallowfortheratioof

distanceDtosourceradiusr.

Itmaybenotedthattheclassic‘point-to-point’formula(Formula6.6)couldberestatedinagenerallyapplicableform:

Wewillnowmoveontoconsiderthree-dimensionalapplicationsoftheseformulae,whereexamples are given that dealwith both two-dimensional surfaces and three-dimensionalobjects.

Cubicillumination

Theprincipleofcubic illumination (Cuttle,1997) is illustrated inFigure6.6.Ashasbeenexplained inChapter5, the illumination distribution about a point in three-dimensionalspace,howeverirregular,mayberepresentedasthesumoftwosimpledistributions.Oneof these is defined by a vector solid that accounts for the entire asymmetry of theilluminationdistributionaboutapoint,whiletheotherisasymmetricsolidthatis,asitsnamesuggests,entirelysymmetricaboutthepoint.Toanalyseanilluminationdistributionintothesecomponents,wecalculate(ormeasure)theilluminancevaluesonthesixfacesofasmallcubecentredatthepoint,orientatedsothatitsfacetsarenormaltothex,yandzaxes,asshowninFigure6.6.Thexandyaxeslieinthehorizontalplaneandthezaxisisvertical,andthesixfacetilluminancevaluesaredesignatedE(x+),E(x-),E(y+),E(y-),E(z+)andE(z-).Iftheseconventionsareadheredto,theprocedurefordealingwithsixilluminancesatapointbecomessurprisinglypainless.

DistancesofsourceSfromPontheaxesareindicatedbyX,Y,andZdimensionswhichmaybepositiveornegativeaccordingtodirection,asshowninFigure6.7.ThelocationofthecubeisX=Y=Z=0,whichwouldberecordedas(0,0,0),andifwereferbacktoFigure6.3,itcanbeseenthatifweweretoreplacethetwo-dimensionalsurfaceshowntherewiththethree-dimensionalcube,thelocationofthelightsourceSwouldbeindicatedby(X-,Y-,Z+)dimensions.

Figure6.6TheCubicIlluminationconcept.Thespatialdistributionofilluminationatapointischaracterisedbysix

illuminancevaluesonthefacetsofacubecentredatthepoint,withthefacetsalignednormaltothex,y,andzaxes.From

thesesixvalues,theilluminationvectormagnitudeanddirectioncanbedefined,andthescalarilluminancecanbe

estimated.

Figure6.7ThelocationofsourceSrelativetoathree-dimensionalobjectisdefinedintermsofX,Y,andZdimensions,

whichmaybepositiveornegativeaccordingtodirection.

LookingbacktoFormula6.7,wecandefinethecubicilluminancevaluesas:

According to the signs of the X, Y, and Z dimensions, the illuminance values may bepositiveornegative,sothatifE(x)hasapositivevalue,thatilluminanceisincidentonthe(x+)facetofthecube,andifnegative,itisincidentonthe(x–)facet.ItfollowsthatpositiveandnegativeE(x)valueshavetobesummedseparately,andnegativevaluesdonotcancelpositiveones.

Consider a 50 watt halogen reflector lamp, such as the MR16 EXT, which we willidentify as S1, and this source is aimed so that its peak beam candlepower (IS = 9150candelas) is directed towards P. The location of S1 relative to P is defined by the

dimensionsX=–1.9m,Y=–2.7m,andZ=3.2m.Then:

and

ISD3=9150/(4.6)3=94.1

Thenfromformulae(6.13–6.15):

E(x)=X(I/D3)=−1.9×94.1=−179lux

E(y)=Y(I/D3)=−2.7×94.1=−254lux

E(z)=Z(I/D3)=3.2×94.1=301lux

Yes,itreallyisassimpleasthat:noangles,nocosinesandthreeilluminancevaluesforthepriceofone.However,itisnecessarytokeepaneyeonthosesigns.NotethatE(x)=–179lux is simply anotherway ofwriting E(x-) = 179 lux.Aswe add the contributions fromdifferentsourcesoneachfacetofthecube,weaddseparatelythesumsofE(x+)valuesandE(x-)values,astheyaretheilluminancesonoppositesidesofthecube.

This example shows the underlying process for determining the six direct cubicilluminance values, but for practical calculations we again utilise the facilities of aspreadsheet. Box 6.1 shows the output of the Cubic Illumination spreadsheet, and aspreviously, theonlydata tobeenteredby theuserare those inred,asallotherdataarecalculatedautomatically.Inthebox,sourceS1fromtheforegoingexampleisshown,andthreemore sources have been added. Rather than have a separate spreadsheet for two-dimensional surfaces, it ismoresimple touse this spreadsheetand tokeep inmind that,following the view direction convention shown in Figure 6.3, the E(y-) value gives thesurfaceilluminance.

Box6.1

CubicIlluminationSpreadsheet

140121

Project: Box6.1MRSE 150lm/m2

DistancesS-P

Vectorcomponents SymmetriccomponentsEvr(x) –209.2 Esym(x) 513.2 Evr 791.6Evr(y) –446.3 Esym(y) 228.4 Esym 327.4Evr(z) 619.3 Esym(z) 240.8 Esr 525.3

Esr(d) 375.3Vector/scalarratio Unitvectorcomponents VectordirectionEvr/Esr 1.51 e(x) -0.264 α51°

e(y) -0.564e(z) 0.782

Notes

Enterdataonlyincellsshowninred.Is=luminousintensityofSindirectionofP.MRSEisthedesignlevelofambientilluminationwithinthespace.CheckDistancesS-P:eithera‘+’ora‘–’dimension;neverboth.

For a typical outdoor application it would be necessary to consider only the directilluminance values on the six faces of the cube, but otherwise, the effect of ambientilluminationneedstobeincluded.AsshowninBox6.1,theuserspecifiestheMRSEvaluetobeprovidedwithin the spaceby the entire lighting installation.Eachof the six cubicilluminancevalueswill be the sumofdirect and indirect values, and theMRSEvalue isaddedtoeachdirectcubicilluminancevaluetorepresenttheeffectofindirectlight.This

assumesthatthecontributionofindirectlightisuniformforallsixfacets,andwhilethisavoidsthetediousprocessofmakingapreciseevaluationoftheindirectilluminanceontoeach facet of the cube, some caution needs to be observed. For a situation where thedistributionofreflectedfluxislikelytobedistinctlyasymmetric,suchaswhereanobjectis located close to a dark wall surface, this simplification could lead to a misleadingoutcome,andusersneedtobealertforthis.Evenso,theassumptionisnotunreasonable.Inanindoorspacewheretheproportionof indirect illuminationis low,itwillhavelittlevisibleeffectandsoitwouldbeawasteoftimetoevaluateitsspatialdistribution.Wherethe proportion of indirect light is high, it is likely to be highly diffused by multiplereflections from light-colouredroomsurfaces so that itscontribution to thevisibleeffectwill be to soften thedirectional effectof thedirect light rather than to impart adistinctdirectionaleffect.Theusershouldbealertforsituationswhereindirectlightcouldbebothdominantanddirectional, and foramore rigorous treatmentof indirect illuminance, seeSimonsandBean(2001).

Thereasonforpredicting,ormeasuring,thecubicilluminancevaluesistoenablevectoranalysis of the illumination solid, and the Cubic Illumination spreadsheet performs theanalysis toproduceBox6.1byapplying the formulaegiven in theprevious section.Thegreatbenefitofusingspreadsheets isnotsimplythat theyautomate thecalculations,butthat they enable the user to explore alternative options, and the reader is stronglyencouraged to access the spreadsheet and to experiencehow this is done. It is simple tochangealightsource,ortomoveitfromonelocationtoanother,andinstantlytheeffectson the vector/scalar ratio (VSR) and the vector direction are given, so that the user canenvisagehowanarrangementofluminaireswillinfluencethe‘flow’oflight,andhowthismightaffecttheperceptionofaselectedthree-dimensionalobject.

Providinganilluminationhierarchy

An illumination hierarchy expresses a lighting designer’s concept for the overallappearanceofalitspace.Itspecifiestheambientilluminationlevelasameanroomsurfaceexitance (MRSE) value, and it expresses how the distribution of direct flux from theluminaireswillaffecttherelativeappearancesofspecifiedtargetsintermsofadistributionoftarget/ambientilluminanceratios(TAIR)values.

TheeffectofambientilluminationupontheimpressionofthebrightnessordimnessofilluminationwithinaspaceisatleastasmuchdeterminedbytheMRSElevelinadjacentspaces as by the actual level within the space, and both Tables 2.1 and 2.2 need to beconsidered formaking adesigndecision.Asdescribed inChapters2 and 3,Table 2.2. isused also for making decisions about TAIR values, and it can be seen that for targetillumination tobeevennoticeable, aTAIRvalueofat least 1.5 isnecessary,withhigherlevelsneededtoachievedistinctorstrongeffects.Emphaticdifferencescanbedifficulttoachieve,asunlessveryhightargetilluminancevaluesaretobeused,theycallfordistinctlylowlevelsofMRSE.

AscheduleofdirectilluminancelevelstobeprovidedontoeachselectedtargetcanbegeneratedfromtheMRSEandTAIRvalues:

The sumof individual target FRF values gives the total first reflected flux due to directilluminationofalltargetsurfaces:

For two-dimensional targets, Etgt(d) is the average direct illuminance, and for three-dimensionaltargets,thebestguideisthedirectcomponentofthescalarilluminance,whereEsr(d)=Esr–MRSE.ThisvaluecanbereadfromtheCubicIlluminationspreadsheet(Box6.1).

The level of first reflected flux (FRF) that is required to provide the design value ofambientillumination,specifiedintermsofMRSE,comprisesthesumofcomponentsduetodirectlightreflectedfromtargetsurfaces(FRFts)andfromroomsurfaces(FRFrs):

Refer back to Boxes 3.1 and 3.2 and note the distinction that was made between firstreflectedfluxduetoilluminationdirectedontotargetsurfaceswiththeaimofestablishingan illumination hierarchy, and first reflected flux that was then required to bring theambientilluminationuptotheMRSEdesignvalue.Whiletargetsneedsignificantlevelsofselective illumination directed onto them in order to achieve appreciable differences ofappearances,forprovidingilluminationontoothersurfacestobringuptheMRSElevel,the

aimshouldbetokeeptheEtgt(d)/MRSEwellbelow3.0,andpreferablybelow1.5(althoughthis can be difficult), so that the flux directed onto these surfaces will not noticeablydetractfromtheilluminationhierarchy.

From Formula 6.16, it follows that FRFrs = FRF – FRFts, and to provide this bothefficientlyandwithlowEtgt(d)/MRSEwillneedoneormore large,high-reflectanceroomsurfaces to receivedirect flux.The ceiling is often the obvious choice, but other optionsshouldbesought.Aseriesofilluminatedwhiteceilingscanhaveablandoveralleffect.

Beforelookingfurtheratcalculationalprocedures,itshouldbeacknowledgedthataveryeffectivewaytoexploredesignoptionsforprovidingFRFrsistouseaproprietarylightingdesign software package such as AGI32 or DIALUX. The trick is to set all surfacereflectance values to zero, so the program gives you direct surface illuminance values.Thesepackagesusuallygiveseriousattention toworkingplane (or floor)uniformityandprovide precise-looking illuminance contours, while giving only average illuminancevalues forwalls and ceiling, so it pays to give attention tohow the appearanceof thesesurfacesmaybeaffectedbyluminairespacing.However,usedinthisway,thesepackagescanprovideausefuldesignfacility.

Lighting techniques such as cove lighting onto ceilings, wallwashing and recessedlightingontofloorplanesarewidelyusedforprovidingroomsurfaceillumination,butdonotoverlookopportunitiesforsuspended(andvisible)pendantluminaires,orincorporatinglightingintofurnitureorhandrails.Whateverlightingtechniqueisemployed,selectionofsuitableluminairesinvolvescarefulexaminationoftheangularrelationshipsbetweenthesourcelocationsandthereceivingsurfaces.Wheremultiplesourcesaretobeused,choosesourceswithbeamanglesthataresmallerthanthesubtenceangleofthereceivingsurface,but large enough to provide full coverage from overlapping beams. The number ofluminairesrequiredis:

whereFBisthe‘beamflux’,orthequantityoflumenswithinthebeam(s).

Manufacturer’s data for lumen outputs of directional luminaires have to be examinedwith care. The only lumens that count are thosewithin the beam, as those outside thebeamare‘spill’andneedtobeblockedorshielded.Forluminaireswithconicalbeams(i.e.,not shaped beams as in wallwashers) it is generally recognised that beam width ismeasured to the direction in which the luminous intensity falls to 50 per cent of themaximumvalue (Figure6.3), and usually the quoted angle iswhole angle from edge toedgeofthebeam,althoughsometimesitisthehalf-beamangle,measuredfromthebeamaxistotheedge.Forthistext,thewholebeamangleisgiventhesymbolB,andthehalf-beamangleisb.

Becauseitisnotalwaysclearwhether‘lumenoutput’datarefertotheentireoutputoftheluminaireorjustthebeamlumens,themostreliablewayofdeterminingthevalueisto

workfromdatafortheluminousintensitydistribution,givenincandelas.Thebeamflux,inlumens,isgivenby:

here:

Imax=maximumbeamluminousintensity,orCBCP,incandelasb=halfbeamangleLD=lumendepreciationfactor

Consider theMR16 EXN halogen lamp,which has a beam angle of 36° and a CBCP of1800cd.Allowingforalumendepreciationof0.8,FB=1.5×1800×π(1−cos(18º))×0.8=332 lm. It should not escape notice that the luminous efficacy for beam lumens for thislampisjust6.6lm/W,andthatisnotallowingfortransformerlosses.Itisclearthatevenwith the precision focussing that these lamps achieve, a significant proportion of thefilamentlumensdonotfindtheirwayintothebeam.This‘spill’lightisnotonlyaconcernfrom the point of efficiency, but also for achieving a controlled distribution of light.Reflectorlampsshouldalwaysbehousedinluminairesthatareshroudedorhavebafflesorlouvres to intercept light spill. This becomes particularly important where spill ontosurfacesadjacenttotheluminairescouldproduceverybrightunwantedlightingpatterns.

It is in thiswaythata layoutof luminaires isdeveloped thatwilldeliver therequiredquantityoflumensontoeachselectedroomsurface.Thedesigner’saimistodeviseafluxdistributionthatwillprovidethefirstreflectedfluxfortherequiredambientillumination,together with the range of TAIR values that will achieve the envisaged illuminationhierarchy.

Daylightillumination

Lighting designers may, from time to time, become involved in fenestration design forspecial applications, such as thewindows for an observation tower or a picture galleryskylight,butregrettably,itismuchmoreusualpractice(inthisauthor’sexperience)thatbythetimeaprojectisintroducedtoalightingdesigner,othershavedeterminedthelayoutofwindows,clerestoriesandskylights,aswellasthetypeofglazing,sunshadingandblindstobeinstalled.Incasesomereadersshouldfindthemselvesconfrontedwithdemandsforadvicethatdifferfrommyexperience,itmaybenotedthatthereisnoshortageofbookson daylighting practice written for architects. However, this section is based on theassumption that the lighting designer’s task is restricted to devising electric lightinginstallations that may from time to time need to respond to significant presence ofdaylight.

Theprincipalmeansforassessingtheperformanceofadaylightinginstallationhasformanyyearsbeenthedaylightfactor,anddespiteafairamountofrecentactivityaimedatimprovingthemodellingofoutdoordaylightavailability,thedaylightfactorcontinuestobe concerned with provision of illumination onto indoor horizontal working planes(HWPs).Inthissection,aquitedifferentapproachisproposed.Everyindoorspaceneedstohave an electric lighting installation.Where there is significant daylight admission, theappearanceofthatspaceanditscontentswillbeaffectedatdifferenttimesandindifferentways,andalightingdesignerneedstogivethoughttohowtheelectriclightinginstallationis to respond to thepresenceofdaylight.This is to takeaccountofbothachievingwhatmaybe perceived to be an appropriate balance of illumination at all times,while at thesametimegainingenergysavingsfromreduceduseofelectriclighting.

Opportunitiesforeitheroftheseobjectivesvaryhugely,andsomeelaborateevaluationsystems have been proposed. Taking a simple approach, fenestration systems can bebroadlycategorisedassidewindows,clerestories,orskylights–orsomecombinationsofthose types. Their impacts upon lighting design may be assessed in terms of thecontributions theymake towards provision of ambient illumination, target illumination,view-outandenergyefficiency(Figure6.8).

Skylightscanprovideveryeffectivelyforambientillumination,aswellasprovidingforHWP illumination.Ashas beendiscussed, ambient illumination, indicatedby theMRSElevel,isconcernedwithhowinter-reflectedlightinfluencestheappearanceofsurroundingroom surfaces, whereas the daylight factor is concerned with enabling effectiveperformance of visual tasks located on desktops or work benches. Skylights that aredesigned so that sloped glazing is orientated in the polar direction (north-facing innorthernhemisphere; south-facing in southernhemisphere)canprovide fairlyconsistent,diffused,ambientilluminationovermuchoftheyear,andthisshouldbetakenintoaccountindevelopinganelectric lightinginstallationtoprovideroomsurface illuminationoutof

normaldaylighthours.Thedaytimeandnighttimeilluminationdistributionsmaybequitedifferent, requiring careful thought about the transitions between these two conditions.Progressivephotoelectricalcontrol is likelytobetheoptionofchoice,andfor largearea,single-storeybuildings,theprospectsfordoingthiseffectivelyandattractivelybydaylightare good (Figure 6.8). By comparison, successful provision of ambient illumination byclerestorywindowsisrestrictedtospaceshavingfairlyhighheight/widthproportions,andmore limited still, are side windows. This is not to imply that they are necessarilyineffective, but rather that their use for providing useful ambient illumination is morerestricted, particularlywhere they occur only in onewall. It is common experience thatwindows in onewall can provide all the illumination required formuch of the time inspacesof domestic scale, but that becomesuncertain in spaceswhere the roomwidth ismorethandoubletheheightofthewindowhead.

Figure6.8Assessmentoflikelyprospectsforvariousrolesforfenestrationinbuildings.

Effectivetargetilluminationfavourselectriclightingasthereliablemeansforsettingupanilluminationhierarchy,butnonetheless,nothingcancomparewiththe impactcreatedwhen sidewindows enable an object, such as a sculpture, to be positioned to ‘catch thelight’ as discussed in Chapter 5. It is, perhaps, the obviously transitory nature of theexperience that adds to its appeal. To design fenestration specifically for the purpose ofproviding target illumination raises the issue of how to integrate that illuminationwithelectric lighting to take overwhen daylight is inadequate.Where the circumstances areseentodemandit,carefulattentiontoachievingeffectivetarget illuminationbydaylight

can be very rewarding, but otherwise it needs to be recognised that the ever-changingnature of daylightmakes it a difficult source to usewhere the appearance of a specifictargetformsanimportantcomponentofthedesigner’soverallconcept.

Thesuitabilityofsidewindowsforprovidingview-outmightseemtobetooobvioustowarrant discussion, but in fact,misdirected thinking on this issue is so common that itreally does need some careful attention.Most surveys of building occupant satisfactionhavebeenconductedonofficeworkers,andagainandagain,theyreportdaylightasbeingahighlyratedoption,andthishasbeentranslatedintostandardsdemandingthatspecifiedminimumdaylightfactorvaluesareprovidedoversomespecifiedpercentageoftheHWP.Thewell-knownfactthat,inopenplanoffices,desksthatarelocatedclosetowindowsareregardedasprimelocations,despitetheirmuchlowerratingsonthermalcomfortindices,iswidelyacceptedasconfirmationofthispreference.However,rethinkingthebasisofthispreference in terms of provision of view is likely to lead to distinctly different designoptions. Side windows designed for view-out combined with reasonably high levels ofthermal comfortwould be quite different fromwindows designed tomaximise daylightadmission. Inparticular, theywouldbe likelyto incorporatesunshadingdevices, fixedoradjustable, specifically designed for the orientation of the window, and which wouldinterceptsunlightwithminimalobstructionofview-out.

Sofar,littleattentionhasbeengiventoenergyefficiency,althoughthistopicisdiscussedin the final chapter. It should not be a matter of surprise that prospects for energyefficiency are shown in Figure 6.8 to be in step with daylight provision of ambientillumination. This is because daylighting systems that have good prospects for ambientillumination are likely to also have good prospects for photoelectric control that willbalance electric light use against daylight availability. The essential difference fromcommon practice is how photosensors are located and commissioned. Sensors need torespondtolevelsofinter-reflectedlightwithinthespace,andsotheyneedtobeshieldedfromdirectlightfromboththedaylightingandtheelectriclightingsystems.Forexample,ina spacewith sidewindows, agood location for a sensor ismountedverticallyon thewall above the window head, and shielded from direct light from the electric lightinginstallationsothatitisexposedtoreflectedlightwithintheroom,butnottodirectlight.Thisisquitedifferentfromconventionalpractice,whichisdirectedbythenotionthatthepurposeforadmittingdaylightintobuildingsistoprovideworkingplaneillumination.Therolethatdaylightcanfulfilbestofallisprovidingambientillumination.

Itmay be noted thatwhere the foregoing procedures have led to an electric lightinginstallation that provides, out of daylight hours, an effective illumination hierarchythrough a combination of target and room surface illumination, the prospect exists foreffective lightingcontrolbydimming just the roomsurface illumination. In thisway,aneffective daytime balance of ambient illumination with a maintained illuminationhierarchymaybeachievedwithworthwhileenergysavings.

Checkingdelivery:measuringthelumens

The on-sitemeasurable illumination quantities that are integral to this perception-basedapproachtolightingdesignare:

meanroomsurfaceexitance,MRSEtarget/ambientilluminanceratio,TAIRvector/scalarratio,VSR,andvectordirection.

Whilehighlightcontrastpotential(HCP)hasalsobeenintroducedasarelevantmetric,itsusefulnessisformakingcomparisonsatthedesignstage,ratherthanasametrictobecheckedon site.Also, at the timeofwriting, there arenogenerally availablemeters forchecking metrics relating to the spectral distributions of illumination, but this may beabout to change. CCD-based spectrometers have recently become reasonably affordable,and this could lead to the availability of portable instruments capable ofmakingon-sitemeasurementsofmanyofthevisibleandnon-visiblefactorsdiscussedinChapter4.

Itisanessentialpartofactingasaprofessionallightingdesignerthattheperformanceofeveryinstallationischeckedagainstthepredictedperformance,regardlessofwhethertheclienthasshowninterestinthedata.Thereneverwillbeaperfectmatch,butknowingthenatureandextentofthedeparturesishowadesignergainsfeelingforexercisingcontroloverperceivedaspectsofillumination.

MeasuringMRSE

WhileMRSEisreasonablystraightforwardtocalculate, it isnotobvioushowtoobtainareliablemeasure of its quantity. Conventional lightmeters have been developed to givemeasurements of lumen density incident on a surface, without regard for the directionfromwhichthelightisincident.MRSEisnotrelatedtoanyparticularsurfaceofincidence,and it discriminates according to the origin of arriving light. It takes account only ofindirect light,anddisregards lightarrivingdirectly from light sources,and this createsadifficultyformeasurement.

ThepurposeofMRSE is toprovideuswithausefulmeasureofambient illumination,andthatmeansthatweneedtomakeameasurementthatrelatesto lightarrivingat theeye, rather than light incident on things that peoplemight choose to look at. Figure 6.9showsasimpleapproach.Chooseapositionanddirectionofviewthat,while it takes inmuchofthespace,avoids lightfromwindows, table lightsandsoforth,andthenholdaluxmeter up to the eye and shield it from any luminaires before taking ameasurement.Depending on the size of the space, repeat this for other positions to obtain an averagevalue.Makingmeasurementsinthiswaycanprovideareasonableindicationoftheextent

towhichapredictedMRSEvaluehasbeenrealised,andthisisvaluableinformationforthedesigner.

ProposalshavebeenmadeformorerigorousproceduresforMRSEmeasurement(Cuttle,2013).Thisapproachinvolvesusinghighdynamicrangeimagingtoproducea

Figure6.9AsimplewayofmakinganapproximatemeasurementofMRSEusingaconventionallightmeter.Exposethe

metertoawideviewofthespaceavoiding,asfaraspossible,windowsandotherlightsources,andshielddirectlight

fromanyoverheadluminaires.

wide-field image of a space defined in terms of luminance. Light sources could then beidentifiedandexcluded,sothattheremainingfieldofviewwouldrepresentthetotalinter-reflectedillumination.Itistobehopedthatsuchasystemwillonedaybecomeavailable,particularlyasitcouldenablenotonlymeasurementofMRSE,butalsoofdiscomfortglare.

MeasuringTAIR

Thenext step in evaluationwill be to checkTAIRvalues.Once again, the designerwillhaveexplainedtheilluminationhierarchyintermsofperceiveddifference,andwillhaveplannedtheluminairelayouttoachievespecifiedTAIRvalues.Checkingbymeasurement

developsconfidenceinrelatingappearancetometrics,aswellasapplicationofmetricsfordeterminingrequiredluminaireperformance.

For a two-dimensional target, TAIR = Etgt/MRSE, which is quite straightforward, butmeasuringEtgtforthree-dimensionaltargetscanposedifficulties.Ifpractical,itisusuallybesttomovethetargettoonesideandtomeasurethecubicilluminationatthespot,usingtheproceduredescribedinthefollowingparagraphs,fromwhichthedirectcomponentofscalarilluminanceisdeterminedbyuseoftheVectorMeasurementspreadsheet,whichisdescribed in the following paragraphs. Otherwise, measurements corresponding to thecubicilluminationaxesmaybetakenovertheobject’ssurface,butitshouldbenotedthatifaparticulardirectionofviewissignificantintheoverallappearanceofthatobjectwithinthespace, thenasinglemeasurementnormaltothatdirectionmightprovideforabetterindicatorofhowwelltheTAIRvaluerelatestotheilluminationhierarchy.

MeasuringVSR

It might seem that the most straightforward way to measure the six cubic illuminancevalues would be to mount a small cube at the measurement point and take successivereadings on the cube’s facets, but in fact, this procedure is cumbersome and tedious,particularly when it comes to taking the measurement on the downward facing facet.Variouscubicilluminationmetersincorporatingsixphotocellshavebeendeveloped,suchas the one shown in Figure6.10, but it isworth taking note of a simple procedure thatmakesuseofjustonephotocellmountedonaphotographictripod(Cuttle,2014).

Envisage the cube tilted, as indicated in Figure 6.11, so that a long corner-to-cornerdiagonal of the cube is vertical, and three facets of the cube face upwards and threedownwards.Thefamiliarx,y,zspatialaxesareunchanged,butnowtheaxesofthecubearedesignatedu,v,w.Figure6.12showsaverticalsectionthroughthetiltedcubeontheyaxis, where BC is one external edge of the cube, AB is a facet diagonal andAC is thevertical longdiagonal.Theratiosof the trianglesidesBC,ABandACare1,√2,and√3,andtheangle .Thisistheanglebywhichtheu,vandwaxesare tilted relative to the horizontal plane, and as shown in Figure 6.11, the u axis isassumedtolieinthesameverticalplaneastheyaxis.

Tomake the cubic illuminationmeasurements, a right-angle bracket is constructed tosupportthephotocellverticallyontheheadofaphotographictripod,asshowninFigure6.13. It helps to have a tripod with a spirit level to ensure verticality and with thehorizontalandverticalmovementsscaledindegrees.Themeasurementprocedureisthenstraightforward. Set the photocell tilt to +35° as shown in Figure 6.14, and rotating thehorizontalmovementofthetripod,readE(u+)at0°,E(v+)at120°andE(w+)at240°.Resetthephotocelltiltto–35°,andreadE(u-)at180°,E(v-)at300°andE(w-)at60°.

Figure6.10Asix-photocellcubicilluminationmeter.Thisinstrumentisself-levellingandisconnectedtoalaptop

computerthatautomaticallyanalysesthedata.PhotographedattheLightingResearchCenter,RensselaerPolytechnic

Institute,Troy,NewYork.

Figure6.11Themeasurementcubeistiltedsothatalongaxisiscoincidentwiththezaxis,andthreefacetsfaceupwards

andthreedownwards.Thefacetsarenormaltotheu,v,andworthogonalaxes.

Figure6.12Averticalsectionthroughthetiltedcubeontheuaxis,whichliesinthesameverticalplaneastheyaxis,

againstwhichitistiltedthroughtheanglea.

Figure6.13Aphotocellheadmountedonaright-anglebracket,ontoaphotographictripod.

Figure6.14Thephotocelltiltedto+35°relativetothehorizontalplane,andreadyformeasuringthethree-dimensional

illuminationdistribution.

Box6.2showstheoutputoftheVectorMeasurementSpreadsheet.Theonlydatatobeenteredby theuserare the sixmeasuredvalues,and from thesea rangeofderiveddatarelatingtothespatialilluminationdistributionisgivenbasedonformulaegiveninChapter5.Whiletheilluminationvectormagnitudeisderiveddirectlyfromthemeasureddata,theoutputdataforvectordirectionareconvertedfromu,v,waxestothemorefamiliarx,y,zaxes,usingthefollowingformulae:

While there are other lightingquantities that are relevant to thedesignprocess, such ascorrelated colour temperature and the highlight contrast, application of these conceptsneednot involvecalculationsandsoconfirmationbymeasurement isnotusualpractice.Nonetheless, when designers are verifying that all is according to expectations, it isnecessarythateveryaspectof interactionbetweenthe lighting installationandthespace

anditscontentsmustcomeunderscrutiny.

Box6.2

CubicIlluminationMeasurement(u,v,w)

140121

Project Box6.2

Illuminancedatainput

E(u+) 109 E(u-) 311E(v+) 365 E(v-) 305E(w+) 342 E(w-) 70

Illuminancecomponents

E(u) -202 ~E(u) 109E(v) 60 ~E(v) 305E(w) 272 ~E(w) 70

Vectorandscalardata

E 344 ~E 161Esr 247 E/Esr 1.39

Horizontalandcylindricaldata

Ewp 236Ecl 268 Ecl/Ewp 1.13

Vectordirection(unitvector)

e(u) –0.587 e(x) –0.436e(v) 0.174 e(y) –0.873e(w) 0.791 e(z) 0.218e(u,v,w) 1 e(x,y,z) 0.999

Vectordirection(altitudeandazimuthangles)

e(x,y) 0.975 alpha 12.7e’(y) –0.895 phi –154.7

Notes

Inputdatashowninredonly.Alltherestaregeneratedautomatically.alpha(vectoraltitude)maybe+iveor-iverehorizontalIfphi(vectorazimuth)=+ivethenanticlockwiserey-axis;elseclockwise

References

Cuttle,C.(1997).Cubicillumination.LightingResearch&Technology,29(1),1–14.

——(2008).LightingbyDesign:Secondedition.Oxford:ArchitecturalPress.

—— (2013). A new direction for general lighting practice. Lighting Research &Technology,45(1):22–39.

——(2014).ApracticalapproachtocubicIlluminationmeasurement.LightingResearch&Technology,46(1):31–34.

Simons, R.H. and A.R. Bean (2001). Lighting Engineering: Applied Calculations.Oxford:ArchitecturalPress.

7DesigningforPerception-BasedLightingConcepts

Chaptersummary

The development of a lighting design proposal involves bringing together the variety ofperception-based lighting concepts into a balance that relates to the design objectivesspecifictothelocation.Itrequirestheabilitytoenvisionaspaceanditscontentsinlight,and seen in thisway, thevolumeof thedesign space ceases tobeavoid, and instead isperceivedasa three-dimensional light fieldcreating interactionswith roomsurfacesandtheobjectswithinthespace.Itisfromthisenvisionedconceptthatthedesignerdevelopsunderstanding of the required characteristics of the light field, leading to the layout ofluminairesandlightsources,togetherwithstrategiesfortheircontrol.Aflowchartlinkingthe lighting concepts to metrics and procedures is introduced. Where the proceduresinvolve calculations, their purpose is seen to be to increase the designer’s level ofconfidencethatthedesignobjectives,statedintermsofperception-basedconcepts,willbeachieved.Thedesignoutcomeisacomprehensivelightingequipmentspecification.

Achievingperception-basedlightingconcepts

Wenowturnourattentiontothetaskofapplyingtherangeofperception-basedlightingconcepts that has been discussed in the foregoing chapters. Each and every project is afresh challenge that calls for understanding of the various roles that the space and itscontents areplanned to serve, backedupby thedesigner’s creative imaginationdirectedtowardsinfluencingpeople’sperceptionsofthespace,itssetting,anditscontentsthroughlighting.

Seen in this context, the lighting concepts provide a framework for ordering thinkingaboutlighting’spotentialforinfluencingpeople’svisualexperiencesoftheirsurroundings.Theserangefromoverallimpressionsofbrightnessordimnessofspacesencounteredinasequenceofenteringandpassingthroughabuilding; theways inwhichthespectrumoflight may arouse or subdue both visual and non-visual responses; through to ordereddistributionsofilluminationthatdifferentiateactivitieswithinspacesandwhichrelatetothe visual significance of objects; and on to the lighting patterns that reveal the form,texture, glossiness, or translucency of individual objects.Within the volume of a space,illumination may, at one extreme, be softly diffused, revealing everything withoutemphasis; and at the other extreme, be selective and sharply directional, differentiatingsurfaces andobjectswith clarity.Aspart of the samevisualisation, lightingmayononehandbeperceivedasbeingwithoutapparentsource,orontheotherhand,sourcesoflightmaybeclearlyexpressedcomponentsofthescene.Thisisthegamutofvariety(oratleasta good part of it!) that a creative designer may bring to bear upon a project, and theperception-basedlightingconceptsprovidemeansforbothorderingcreativethinkingandexercisingcontrol.

Whiledesignisnot tobereducedtoastep-by-stepprocedure, theflowchartshowninFigure7.1presentsarationalorderingofthelightingconceptsandgivesanoverallguidetothissection.

Ambientillumination

Is the first impression tobeof abrightly lit space, or adimly lit space, or something inbetween?This issuehasbeendiscussed in the sections inChapter2 entitled ‘TheMRSEconecpt’and‘Applyingtheambientilluminationconceptindesign’,anditrequirescarefulthought. A person visiting the space will inevitably experience it within a sequence,arriving fromoutdoorsor fromanother indoor space,beforemovingon. Itmaybe theirdestination,ora space that theywill experience inpassing.Perhaps thedesignaim is toarouseattention,oralternatively,toprovideaplaceforrest.Thepossibilitiesarelimitless,asaretherolesthatlightingmayplay,butthroughout,theconnectionbetweentheoverallsenseofbrightnessandthefirstimpressionisstrong.Tables2.1and2.2togetherprovidea

simple introduction to ambient illumination, but it is up to the designer to developsensitivitytotherelationshipsthatmayarise.

While MRSE (mean room surface exitance) may be an unfamiliar concept, it issurprisingly easy and rewarding to come to termswith. The notions of visualising lightwithinthevolumeofaspaceratherthanincidentonsurfaces,andofthinkingintermsoflightattheeye,sothatweachieveouraimsthroughprovidingreflectedlight,anddirectlight is simply a means to achieving that end, lead naturally to effective lightingapplications. TheAmbient Illumination spreadsheet is a useful tool for the first stage oflinkingavisiontoaluminairelayout.

Illuminationcolourappearance

Apart fromthebrightnessordimnessofambient illumination, itscolourappearancecansignificantly affect the appearance of a space. Illumination that is basically ‘white’maynonetheless have a distinct tint, and the acceptability, or even the attractiveness, of thattint,canbestronglyaffectedbycontextandpeople’sexpectations.

Ashasbeenexplained, thegenerally acceptedpractice is to specify the tintof ‘white’illuminationbyCCT(correlatedcolourtemperature)wherelowvalues(CCT<3200K)areassociatedwithyellowish-whitelightand‘warm’colourappearance,andhighvalues

Figure7.1Alightingdesignflowchart.Followthrougheachrowfromconcept,tometric,toprocedure.Thesequenceof

conceptsisproposedasbeinglogical,butmaybeadaptedtosuitcircumstances.Theaimistodevelopproposalsfor

discussion,whichwouldleadtothedesignproposal.Post-installationassessmentandmeasurementshouldalsobe

includedaspartofthedesignprocess.

(CCT>5000K)withbluish-whitelightand‘cool’colourappearance.Atpresent,thisisthechoice that the lighting industry offers, but research findings have beennoted (page 46)which indicate that light source chromaticitiesdeparting from theblack-body locusmayofferpreferredcolourappearancealternatives.

Illuminationhierarchy

Situationsoccurwheretotallydiffusedilluminationthatrevealswithoutemphasiscanbehighlyeffective,butmoreusuallysomeorderingofilluminationdistributioniscalledfor.There may be various reasons for this. The aim may be to distinguish between zoneswithinaspace;itmaybetoincreasethevisibilityofselecteddetail;oritmaybetodrawattentiontoobjectsofvisualsignificance.Theabilitytoenvisionastructureddistributionofilluminationisadefiningskillofalightingdesigner,butitneedstobeunderstoodthatwhiletheenvisionedeffectisadistributionofreflectedflux,itisachievedbyprovidinga

distribution of direct flux onto selected targets thatwill generate that distribution. Thisability to separate in the mind the applied distribution of direct flux and the resultingdistribution of reflected flux is crucial. It is an acquired skill that evolves from carefulobservationofhowappearanceisaffectedbythebalanceofdirectfluxappliedtotargets,andofdiffusely-reflectedambientillumination.

The notion of an illumination hierarchy, by which the lighting designer’s concept ofemphasis forms the basis of a structured illuminationdistribution, is set out in terms ofTAIR (target/ambient illumination ratio). The Illumination Hierarchy spreadsheet is ausefultoolforseeingthroughthisstageoftheprocedure.

Colourrendering

CRI(colourrenderingindex)isthereadilyavailablemetric,anditslimitationshavebeendiscussedatsomelength.CRIservestheneedsofspecifiers,butdesignersneedmore.TheGAI(gamutareaindex)addsanindicationofcolourfulnesstothatoffidelity,buttoooftenthevaluesonthisscaleareunavailable.Reallyuseful information,suchasCMV(colour-mismatch vector) data, is unlikely to be available, so that designers need to developthrough directed observation, as described in the section ‘Source spectrum and humanresponses’ in Chapter 4, the experience to be able to select light sources with colourrenderingpropertiesthatreallysuitparticularapplications.

The‘flow’oflight

The directional properties of a light field that generate shading patterns throughinteractionswiththree-dimensionalobjectsprovideadynamicqualitytotheappearanceofaspace.Thisaspectoflightingisparticularlyassociatedwithspaceslitbysidewindows,andwheredaylightcreatesstrong‘flow’oflighteffects,carefulconsiderationneedstobegiven tohow theelectric lighting is to respond to thevarying shadingpatterns.Distinctshadingpatternsonindividualobjectsareeasilyproducedbyspotlights,butthe‘flow’oflight concept refers to a lighting effect that creates a coherent sense of illuminationdistributionwithinaspace.

The VSR (vector/scalar ratio) relates to the perceived strength of the ‘flow’, and thedirection of ‘flow’may be indicated by the unit vector e, or by the vector altitude andazimuthangles.

The‘sharpness’oflighting

The potential for lighting to generate highlight patterns on glossy-surfaced three-

dimensional objects is indicated by the HCP (highlight contrast potential), which alsorelates to theperceived ‘sharpness’ofshadowpatternsandtheoverallappearanceof the‘crispness’oflighting.

Luminouselements

ThisistheonlyoneoftheconceptslistedinFigure7.1thathasnotbeendiscussedinthetext,butitisinfacttheeasiestofalltheconceptstocometotermswith.Oftenitwouldbetrue to say that, for lightingdesigners, theperfect luminairewouldbe invisible.As it is,designersoftenstrivetoeliminateasfaraspossibleanyvisibleintrusionofluminairesintothescenesthattheycreate.Luminairesarerecessedintoceilings,tuckedaboveshelvesorcornices,orbuilt-inunderfurnitureorhandrails.Theyare,forthemostpart,regardedasnecessarybutunwelcomeintrusionsintothescene.

Thereare,however, timeswhenthe luminairesbecomefeaturesof thedesignconcept.Therecanbeallsortsofreasonsforthis,butarecurringoneisthatthespaceisblandandfeatureless,andwouldbenefitfromthepresenceofself-luminous,eye-catchingobjectsthatadd ‘sparkle’ and interest to the scene. There are nometrics for assessing the perceivedeffectnorarethereproceduralstepsforincorporatingtheseelementsintothedesign,butwhenthedecisionismadethatluminouselementsaretobepartofthescene,itisaswelltokeepinmindthewell-wornadage,“Oneman’ssparkleisanotherman’sglare”.

Thedesignproduct

The spreadsheets that have been used to generate the Boxes shown alongside the textfacilitate the translation from envisioned effects to luminaire performance not only byperforming the calculations, but by providing the designer with almost unlimitedopportunitytoexplorealternativeoptions.Designersareencouragedtousethemasmodelsto develop spreadsheets that serve their own fields of lighting practice. Commerciallyavailable ‘lighting design’ software packages generally fail to address the issues thatconcernacreativedesigner.

Whilemostpeoplethinkofalightingdesigner’soutputbeingtheilluminationthatuserswillexperience,therealitiesoflifeshouldcausethedesignertotakeadifferentattitude.Itis the specification document, listing lamps, luminaires, circuits and controls, thatdetermineswhetherornothisorhervisionofaspaceinlightwillbeachieved,andforthisreason,thespecificationshouldberegardedasthedesignproduct.Ithasbeenstatedabovethat the ability to envision is the essential design skill, but the ability to translate thatvisionintoaspecificationdocumentthatwillnotbecompromisedrunsitaclosesecond.Never losesightofthefactthatwhenthespecificationgoesouttotender, thecontractorwhowillgetthejobwillbetheonethatputsinthelowestprice.

Definingilluminationadequacy

Whilethis‘perception-basedapproach’tolightingdesignisproposedasbeingappropriateforindoorlightingapplicationsrangingfromsimple,everydayactivitiestocomplex,large-scaleprojects, it cannotbedenied that therewillalwaysbe somesituations forwhich itwouldbequitesensibletoprovideuniformilluminationoverthetime-honouredhorizontalworkingplane,wp,whichextendswall-to-wallandmaybecoincidentwiththefloorplane,orelevatedaboveit.

This type of lighting practice is sometimes referred to as ‘lumen dumping’, and theconventions adopted by lumen dumpers for planning their lighting installations includetreatingeveryspaceasarectangularroommeasuringL×W,withthelightinginstallationcomprisingaregulargridofluminaireslocatedontheluminaireplane,lp,whichmaybecoincidentwiththeceilingorbelowit,andwithonlythewallheightHbetweenlpandwpbeingcountedaswallareaw.Allofthesedimensionsareinter-relatedbytheroomindex,whereRI = L.W/H(L+W). (North American practice uses the room cavity ratio, whereRCR=5H(L+W)/L.W=5/RI.)

Clearlythisapproachcontrastswiththatadoptedbytherestofthisbook,soletusnowimagine what would be the implications for lumen dumpers if the designated lightingstandardwere to be based on perceived adequacy of illumination, PAI, being prescribedminimumlevelofambientillumination,specifiedintermsofmeanroomsurfaceexitance,MRSE.

ThedefiningexpressionstatesthatMRSEequalsthefirstreflectedfluxFRFdividedbytheroomabsorptionAα,sothat:

and:

FRF=MRSE×Aα

Roomabsorptionisthesumofproductsofroomsurfaces(workplane,luminaireplane,andwalls)andtheirabsorptancevalues:

Aα=Awp(1−ρwp)+Alp(1−ρlp)+Alw(1−ρw)

Itiscommoninsuchsituationsforroomsurfacefinishestobeundefined,andfor‘typical’surfacereflectancestobeassumed.Providingthatassurancesaregiventhat‘light’finisheswillbeused,thefollowingsurfaces’reflectancesmaybeassumedastypical:

ρwp=0.25;ρlp=0.75;ρw=0.5;

Applyingthesereflectancevaluestotheaboveexpressiongives:

Aα=L.W+H(L+W)

So:

ThefirststepfordeterminingalightinglayoutistouseFormula7.1tocalculatethefirstreflectedflux,afterwhichthenexttaskistodeviseadistributionofdirectfluxfromthelightinginstallationthatwillprovidetherequiredlevelofFRF,andthispresentsthelumendumperwith a novel quandary. There is no stipulated illumination distribution. At oneextreme, s/he coulddirect all of the fluxonto theworkplane, but thatmight create thedreaded“caveeffect”.Attheotherextreme,allofthefluxcouldbedirectedupwardsintothe cavity above the luminaire plane, and while that would be a very efficient way ofprovidingtheFRF,itwoulddistractattentionawayfromtheworkplane.

Theconceptof illuminationhierarchy isallaboutprovidingcontrolleddistributionsofillumination,andforthelumendumper,thesolutionwouldbetonominatetheworkplaneasthetargetandtoworktowardsasuitabletarget/ambientilluminationratio,TAIR.

Target illuminance is thesumofdirectand indirectcomponents, so that ifweassumeluminairesthatdirectall,oratleastalmostall,ofthedownwardfluxontotheworkplane:

FromFormula7.1:

whereUFFR=upperfluxfractionratio

Then:

Inthisway,bothTAIRandUFFRarereadilypredictablefortargetworkplanes.Ifitisdecidedtousefullyrecessedluminaires,oranyothertypeofluminaireforwhichUFFR=

0, then TAIR values can be read from Table 7.1. It can be seen that high values areunavoidable,particularlyforlowRIvalues.

Table7.1Valuesoftarget/ambientilluminanceratio,TAIR,againstroomindexwherethehorizontalworkingplane,

HWP,isthetargetsurfaceandalldirectfluxisincidentontheHWP.Lightsurfacereflectancesareassumed

RI TAIR

1 92 73 6.34 65 5.8

Figure7.2TAIRvaluesforthehorizontalworkingplane,whenitisthetarget.Exceptforatlowvalues,roomindexhas

onlyslighteffect,buttheupperfluxfractionratioisstronglyinfluential.

These highTAIR values can be avoided by use of luminaires that have someupwardlightcomponent.Figure7.2showshowUFFRvaluesrelatetoTAIR,andthismaybeseenasasimpleversionofamorecomprehensivestudyreportedbyLynes(1974).Jay(2002)hascommented that aBZ3 lighting installationwith a 10 per cent upward light componentprovidesasatisfactoryappearance inawiderangeofworkplaceapplications,andFigure7.2showsthistorelatetypicallytoaTAIRvaluearound5exceptatlowRIvalues.TothisIwouldaddmyownobservationthatitneedsaTAIRvalueofatleast3toimpartadistinctdifference of appearance to a target, and for a levelmuch less than 2, the difference is

unlikelytobenoticeable.

Thedifferencebetweenthissituationandcurrentgeneral lightingpractice is thatonlytheamountoflight,asitinfluencesassessmentofilluminationadequacy,isspecified,andthedistributionof that light isundefined.Thismeans that foranyone toplana lightinginstallation, some thought has to be given to the question;What is the purpose of thelighting? Perhaps a grid of luminaires providing uniform work plane illuminance isappropriate,butperhapsnot.MRSEspecificationsmayapplytomanylocationsotherthanworkplaces–infact,theonlyexceptionswouldbelocationswheredistinctlydimlightingmaybe a legitimate design objective.Generally it should be assumed that providing forPAI (perceived adequacy of illumination) doesmatter, and at the same time, that thereneedstobescopeforspecifictargetstobeselectedsothatanilluminationhierarchycanbedrawnupintermsofTAIRvalues.Itisinthiswaythatanilluminationdistributioncanbecreatedthatmeetsthespecificrequirementsofaspacewithoutbeingcompromisedbytheneedtocomplywithalightingstandardthatprescribesuniformity.

Theimportantroleofroomsurfacereflectancevalues

It’stimeforanotherthoughtexperiment.Supposethatyouaredesigningasettinginwhichawhitemarblesculptureistobedisplayed,andyouwanttoachieveastunningeffect.Youwantthesculpturetostandoutfromitsbackgroundsostrikinglythatitappearstoglow.You want the highest possible target luminance contrast. Peter Jay has examined theconditionofmaximumattainablecontrast(Jay,1971)forwhicheverylumenprovidedisincidentonthetarget,andthebackgroundisilluminatedonlybylightreflectedfromthetarget.

Tosimplifythesituation,wewillassumeallsurfacestobediffusingreflectorssowecandefinemaximumattainablecontrastintermsofexitance(M)valuesforatarget,tgt,seenagainstabackground,bg:

In any enclosed space, the total room surface area,Ars, is the sum of the areas of theenclosingsurfacesandanyobjectscontainedwithinthespace.IfwedirectallofthelightfromtheluminairesontoatargetareaAtgt,thentheremainderofthesurfacearea,whichforms the background to the target, isAbg, so thatArs =Atgt +Abg. As the backgroundreceives only indirect illumination, the contrast for this conditionwill be themaximumattainable contrast, Cmax. Target and background illuminances and reflectances are Etgt,Ebg,ρtgtandρbgrespectively.

The target is completely enclosed in a space of exitance Mbg, and so the indirectcomponentof itsaverage illuminancewillbeequal toMbg.Thedirect componentof thetargetilluminanceisthereforeEtgt–Mbg,andthetotalluminousfluxfromtheluminairesisAt(Etgt–Mbg).Weapplytheconservationofenergyprincipletostatethatthisfluxmustequaltherateofabsorptionbyboththetargetandbackgroundareas,sothat:

Atgt(Etgt−Mbg)=AtgtEtgt(1−ρtgt)+AbgEbg(1−ρbg)

So:

AtgtEtgt−AtgtMbg−AtgtEtgt+AtgtMtgt=AbgEbg(1−ρbg)Atgt(Mtgt−Mbg)=AbgEbg(1−ρbg)

DividethroughbyMbg,notingFormula7.4andthatMbg=Ebgrbg:

ThisisJay’sformulaformaximumattainablecontrast(Jay,1971).ItshowsthatCmaxistheproduct of two factors, one being the ratio of the surface areas,Abg/Atgt, and the otherfactor,(1-ρbg)/ρbg,beingdependentonlyonthebackgroundreflectance.Nowthinkbackto

thewhitemarblestatue.Thesetwofactorstellusthattomaximisethecontrast,weneedtoput the statue into a space that is large in relation to the statue, andwith low surfacereflectance.Thereisnothingsurprisingaboutthat,untilwenoticethatthereisnomentionoftargetreflectance.Ifweweretoreplacethewhitemarblestatuewithablackone,alltheexitancevalueswouldbereducedproportionately,butthecontrastwouldbeunchanged.

Let’s lookat this formulaabitmorecarefully.The target reflectancehasdroppedout,and (1-ρ)/ρ term is the background absorptance/reflectance ratio, α/ρ, and as shown inFigure2.9,theinverseofthisratio,ρ/α,describestheinfluenceofreflectanceuponambientillumination.BothoftheseratiosareplottedinFigure7.3,whereitcanbeseenthattheymirroreachother.Thisfigurebreaksdownintothreezones.Wherethevalueofρis lessthan0.3, roomsurface exitancewill be substantially lower thandirect illuminance.Herewehavethepotentialtoachievehightarget/backgroundcontrasts,evenwherethetargetareaisnotmuchsmallerthanthebackgroundarea.Movingtotheothersideofthechart,where ρ is greater than 0.7, room surface exitance exceeds direct illuminance by somemargin,andwhilethiswillgiveanenhancedsenseofoverallbrightness,reasonablyhighcontrastscanbeachievedonlywithtargetsthataremuchsmallerthantheirsurroundings.Inthemid-zone,whereρvaluesareintherange0.3to0.7,roomsurfaceexitancevalueswillbefairlysimilartodirectilluminancevalues.Thisequalbalanceofdirectanddiffuseillumination components gives scope for providing noticeable (but not distinct)illumination differences while avoiding strong contrasts. It is also a prescription forpracticalroomsurfacereflectancevalues,andguidesforgoodlightingpracticeinvariablyrecommend reflectanceswithin this range. Itmay be looked upon as the safe range, inwhich there is some limitedscope foremphasis,butprovidingsufficient light isput intothespace,everythingwillappearadequatelylit.However,thisshouldnotinhibitacreativedesigner.Theimportantthingisforthedesignertohavedeveloped,throughobservationofthe impact that lightingcanhaveon theappearanceof lit spaces, the confidence to stepoutsidetherestrictionsofrecommendedpractice.

Figure7.3Theinfluenceofroomsurfacereflectionproperties.Foreverysurface,ρ=1-α,whereρisreflectanceandαis

absorptance.FromFormula2.1itcanbeseenthatMRSEisproportionaltoρ/α,andfromFormula7.3,maximum

attainablecontrastisproportionaltoα/ρ.Whereoverallroomsurfacereflectance,ρ,iseithermorethan0.7orlessthan

0.3,it’seffectuponappearancewillbepronounced.

Jay’sstudyextendedbeyondatargetobjectsurroundedbyabackground,toexaminethelimitations for contrastwhen the target is part of the space itself. Examplesmight be ademonstrationareainateachingspace,oradancefloorinarestaurant.Itmustnotbelostsight of that the formula is based on the assumption that 100 per cent of the providedluminousfluxisincidentonthetarget,sothatambientilluminationoutsidethetargetareaisdueonlytoreflectedflux.Itis,afterall,aformulaformaximumattainable contrast,and so unlikely to be achieved in practice.However, itmay be noted that as the targetbecomesalargerpartofthetotalsurfacearea,soitbecomesrealistictoassumethatspilllightontothebackgroundismore likelytobesignificant,whichhasthedisadvantageofreducingactualtargetcontrasts,andtheadvantageofreducingtheneedtosupplementthetargetlightingtoprovideforsafemovement.

Finalremarks

The perception-based lighting design approach proposed in this book leaves untouchedsome aspects of lighting that have traditionally been cornerstones of lighting policy. Inparticular,thetopicsoflightingforproductivityinworkplacesandefficientuseofenergyforlightinghavebeenbarelymentioned,andsowewillclosebylookingathowthesetwoaspectsinteractwiththisperception-basedapproach.

Lightingforproductivityinworkplaces

Weliveinanerainwhichifthingsneedtobeseen,theyaredesignedtobeseen.Examplesof this surroundus.Carboncopieswere first replacedbyphotocopies, and thenby laserprintedmaterials,beforepaper-basedmaterialsinturngavewaytoscreen-baseddisplays,originallyCRT screens,which in turn have been replaced by high-definition, full-colourLED displays. At least, that is what has happened where material has to be read by ahumanbeing.Wheretheprocessofreadinghasbeentakenoverbymachines,suchasthebar-codereadersatsupermarketcheckouts,thevisualtaskhasnotsimplybeeneased,buthas actually been eliminated, and similar examples can be found in many industrialworkplaces.

This revolution in the role of vision has not been accompanied by any seriousrevaluationoftheprovisionofillumination.Lightingstandardsandrecommendedpracticedocumentsspecify illuminancevalues forvisual tasks,andforanyonewhocares toreadthe cited literature, these are claimed to be based onmeasured values of the luminancecontrast and angular size of the critical detail at the eye. The reality is that while thespecified illuminancehas climbedduring theprevioushalf century, visual taskdifficultyhaserodedorvanished.Whathasnotchangedisthenotionthatprovidingforilluminationadequacy involves lighting theHWP (horizontalworkingplane) toa specified level, andbecause this is the basis of lighting standards, it applies to all manner of indoorapplications.Everyspacefromawaitingroomtoaprecisionmachineshopisassessedbysomeoneholdinganilluminancemeterataroundwaistheight,andwanderingaroundtoensurethatatnopointdoesthemeasuredvaluedropbelowthespecifiedone.

There are a few exceptions. Some visual tasks cannot be redesigned, and notableexamplesaresurgery,forobviousreasons,andqualitycontrolinspection,wheretheaimistodetectevenveryslightdefectsinmanufacturedproducts.Thecommonfeatureoftheseapplications is that they call for specialised solutions that are quite separate from thegeneral lighting. Consider, for example, that you have undertaken a project to light adentist’spremises.Youthinkthroughtheprogressionofapatientarrivingattheentrance,advancingto thereception,andmovingthroughto thewaitingroombeforebeingcalledinto the surgery.At every stage youhave different ideas about the appearance that you

wanttocreate,andhowyouwilluse lightingtoachieve it.However,oncethepatient istiltedbackinthedentist’schair,andthedentistneedsafewthousandluxonthepatient’sbackmolars, a completelydifferent formof lighting takesover, and theway that that isprovidedisnoneofyourconcern.Aluminairethat incorporatesahighleveloftechnicalexpertiseisbroughtintouse,butitisacomponentofthedentist’sequipmentanddoesnotformpartofthelightinginstallation.

Itmaybesaidthat,generally,inanindoorspacewherethereisanactivitythatinvolvestheneed for visibility, the surfaces associatedwith that activity should be designated astargetsurfacesandincorporatedintotheilluminationhierarchyscheme.Exampleswouldinclude art galleries, retail stores, industrial assembly lines, and the tellers’ counters inbankingpremises.Foractivitiesthatareparticularlyvisuallydemanding,whichincludethealready cited examples of surgery and quality control, specialised lighting solutions thatare designed not merely to deliver lumens, but to enhance the visibility of the criticaldetail, are to be applied.Wherever people are to spend long working periods, whethervisually demanding or not, provision for perceived adequacy of illumination requiresattention.Ifhighlevelsoftargetilluminationaretobeapplied,thenkeepingTAIRdowntomodestvalueswillhavetheeffectofensuringappropriatelyhighlevelsofMRSE.

Efficientuseofenergyforlighting

It goes without saying that energy efficient lighting must make use of high luminousefficacy light sources in optically efficient luminaires. Beyond this, the lightingneeds toprovideforPAI(perceivedadequacyofillumination),nomoreandnoless,atalltimesthatthespaceisoccupied.Thismayinvolveacontrolsystemthatcandimtheelectriclightingto take account of daylight availability, and that will switch it off when the space isunoccupied.TheimportantwayinwhichthisdiffersfromgoodcurrentlightingpracticeisthatitrelatestoPAI,whichmeansthatthelightingsensorisinstalledsothatitrespondstoMRSE, and not to HWP illuminance. The thinking behind this is that the space shouldalways appear adequately lit without ever being lit to excess, and that instead of thedesignerworkingtokeepinsidealightingpowerdensitylimit(W/m2),theaimwouldbeagenuinelylowenergyinstallation,measuredinkWh/m2.yr.

While this scheme seems reasonably straightforward, it could lead to the illuminationhierarchybeing compromised.Overall dimming to allow for changing levels of daylightwould inevitably change the balance of the lighting, particularly in situationswhere thedesignerhasputtogetheraninstallationthatprovidesdifferentTAIRvalues,andinvolvesdifferent types of light sources focussed onto different targets. In such circumstances, itmay be an effective policy to maintain the selective target lighting, and to dim onlylighting that isprovided toboostMRSE,particularly thatwhichwashes lightover roomsurfacesclosetothesourceofdaylight.

So the question arises, would changing from conventional practice of specifyingillumination requirements in terms of minimum HWP illuminance, to basing it uponsatisfyingPAI,leadtolowerenergyconsumption?Thefirstthingtomakeclearisthatthisperception-basedapproachisnotproposedasmeansforreducinglightinglevels.Thebasicrequirement is that a space should appear adequately lit, taking account of the viewer’slikely expectations. Conventional practice can, on occasion, lead to the ‘cave effect’, adismal appearancebrought aboutby themisguidedpursuit ofhigh efficiency.To restatethe illumination standards in MRSE values should have the effect of preventing thisunfortunateoutcome.However,ithastobeunderstoodthattheprescribedlux(orlm/m2)valueswouldneedtobesubstantiallylowerthanthecurrentHWPvalues,notbecauselesslightistobeprovided,butbecauseofthedifferentwayinwhichthemetricevaluatesthelevelofilluminationprovision.

Soiftheaimistocomeupwiththeultimateenergyefficientsolutionthatwillsatisfythe PAI criterion by providing a prescribedMRSE level,whatwould be the outstandingfeaturesofsuchaninstallation?Themostobviousdifferencewouldbetheappearanceofthespaceitself.Everysurfacewithinsuchaspacewouldbewhiteorchromiumplated!Toexperience the space would be like stepping into an integrating sphere. Every lumenemittedwithinthevolumeofthespacewouldbeguaranteedlongevity.Itwouldundergoaprolonged life of multiple reflections before eventually being absorbed by the roomsurfaces.Togetanideaofwhythiswouldbeso,takealookatFigure7.3.Theρ/αwouldbesohighthatitwouldtaketheemissionofonlyafewlumenstobuildupahighlumendensity within the space. Of course high efficacy light sources and high efficiencyluminaireswouldbeapplied,sothatonlyaverylowpowerdensitywouldberequiredtomeetanyreasonableMRSEvalue.

Looknowattheα/ρfunctioninFigure7.3,anditcanbeseenthataspotentialforMRSErocketsupwardswithincreasingroomsurfacereflectance,potentialforcontrastgetseverlower.Weare lookingatanenvironment inwhicheverything isvisible,butnothinghasdistinct visibility. There is no illumination difference, whether a planned illuminationhierarchyoranarbitraryoutcomeofsourceanddistance,andthereisno‘flow’,andthereisno‘sharpness’.

Comparedwiththisoutcome,itcanbeseenthat lightingthatrelatestospace,objects,and particularly to people, comes at a cost. Seen in this way, current notions of goodlightingpracticedo,infact,representoneparticulartypeofenergyefficiencycompromise.Topursueperception-basedlightingconceptsistobringdifferentfactorsintotheequation.Luminaireperformance is still there,but the roomand its contentsare tobe seenas thesecondaryluminaire,whoseroleistodeliverluminousfluxtotheviewer.Theroleoftheprimary luminaires (the lighting hardware) is to energise the secondary luminaire. Thisprocessshouldbeengineeredforeffectivenessandefficiency.

References

Jay,P.A.(1971).Lightingandvisualperception.LightingResearch&Technology,3:133–146.

——(2002).Subjectivecriteriaforlightingdesign.LightingResearchandTechnology,34:87–99.

Lynes,J.A.(1974).Illuminanceratiosasaconstraintonutilance.LightingResearchandTechnology,6:172–174.

Appendix

Abbreviationsusedinthetext

α Absorptance,orvectoraltitudeangleφ Vectorazimuthangleρ ReflectanceA,Aα Area,roomabsorption(m2)CAM ColourappearancemodelCBCP Centrebeamcandlepower(cd)CCT Correlatedcolourtemperature(K)CGA ColourgamutareaCQS ColourqualityscaleCMV ColourmismatchvectorCRI ColourrenderingindexD,D/r Distance(m),distance/radiuscorrectionE,Es(d) Illuminance,directilluminanceonsurfaces(lx)E,E(x) Vectorilluminance,vectorilluminancecomponentonxaxis(lx)e,e(x) Unitvector,unitvectorcomponentonxaxis~E,~E(x) Meansymmetricilluminance,symmetricilluminanceonxaxis(lx)FRF Firstreflectedflux(lm)HCP HighlightcontrastpotentialHWP HorizontalworkingplaneMS Exitancefromsurfaces(lm/m2)MRSE Meanroomsurfaceexitance(lm/m2)PAI Perceivedadequacyofillumination(MRSE)RI RoomindexS/P Scotopic/photopicratioTAIR Target/ambientilluminanceratioTCS TestcoloursampleVSR Vector/scalarratio

Index

Adelson,EdwardH.6

Ambientillumination6,11,103,120

Perceivedbrightnessordimnessof18

Attributes(ofobjects)6,28

Bezold-Brückehueshift41

‘Black-body’45

‘Caveeffect’131

Checkershadowillusion3

Circadianresponse44,60

Colour:

appearancemodels(CAMs)55

‘ClassA’60

gamutarea57,60,122

mismatchvector57

qualityscale54

renderingindex51,122

Correlatedcolourtemperature45,48,60,120

D/rcorrection97

Daylightfactor106

Energyefficiency108,131

Exitance19

Fenestrationsystems106

‘Firstbounce’lumens17,93

Flux:

Firstreflected17,22,33,92

Inter-reflected20

‘Flow’oflight50.66,75,122

Flowchart:

Illuminationhierarchy32

Lightingdesign121

Gershun,A.A.77

Gretag-Macbeth‘ColorChecker’61

Highlightcontrastpotential87

Horizontalworkingplane12,131

Hunt,R.W.G.55

Illuminance:

Indirect19

Ratios28

Recommendedlevels12

Scalar81

Illumination:

Adequacy124

Colourappearanceof45,120

Colourrenderingof50

Cubic99,112

Daylight106

Hierarchy28,103,122

Perceivedadequacyof30,33

Perceiveddifferenceof18,29

Solid76,81

Vector79

InternationalCommissiononIllumination(CIE)39

Intrinsicallyphotosensitiveretinalganglioncells44

Jay,Peter127

‘Kruithofeffect’49,60

Lightingpatterns:

Highlightpattern5,70,84

Shadingpattern5,67

Shadowpattern5,70,89

Threeobjectlightingpatterns66

Lightingstandards12

‘Lumendumping’124

Luminousefficiencyofradiantflux40

Luminoussensitivityfunction(V(λ))39,60

Othersensitivityfunctions41–45,60

Lynes,J.A.28,81

McCandless,Stanley50,60

Maximumattainablecontrast127

Meanroomsurfaceexitance16,17,33,92,120,131

Calculation103

Measurement109

Melanopsin44

Melatonin44

Mesopiccondition40

Nayatani,Y.55

Perceivedadequacyofillumination30,33,124,131

Perceived‘tint’46

Photopiccondition40

Phototropism27

Productivity,lightingfor130

Rea,Mark41

ReciprocalmegaKelvinscale46

Related(andunrelated)colours4,27

Roomabsorption17,92

Roomsurfacereflectancevalues127–129

Scalarilluminance81

Scotopiccondition40

Scotopic/photopicratio42–43,60

‘Sharpness’ofillumination67,84

Spreadsheets:

Ambientillumination23

Cubicillumination102

Cubicilluminationmeasurement117

Illuminationhierarchy35–36,122

Symmetricsolid80

Target/ambientilluminanceratio30,33,122

Calculation103

Measurement111

Testcolourmethod51

Thoughtexperiment:

Howbrightlylit?12

‘Sharpness’ofillumination84

Uniformityfactor12

Umbra,penumbra89

‘Visualclarity’59,61

Vectordirection:

Altitudeangle82

Azimuthangle82

Unitvector83

Vector/scalarratio67,82,122

Measurement111

Vectorsolid79

View-out108

Waldram,J.M.28

Worthy,J.A.84