the impact of light including non-image forming effects on ... · prof. j.-l. scartezzini, dr m....

POUR L'OBTENTION DU GRADE DE DOCTEUR ÈS SCIENCES

acceptée sur proposition du jury:

Prof. Ph. Thalmann, président du juryProf. J.-L. Scartezzini, Dr M. Münch, directeurs de thèse

Dr M. Knoop, rapporteur Prof. L. Ortelli, rapporteur Dr B. Paule, rapporteur

The Impact of Light Including Non-Image Forming Effects on Visual Comfort

THÈSE NO 6007 (2013)

ÉCOLE POLYTECHNIQUE FÉDÉRALE DE LAUSANNE

PRÉSENTÉE LE 9 DÉCEMBRE 2013

À LA FACULTÉ DE L'ENVIRONNEMENT NATUREL, ARCHITECTURAL ET CONSTRUITLABORATOIRE D'ÉNERGIE SOLAIRE ET PHYSIQUE DU BÂTIMENT

PROGRAMME DOCTORAL EN ENVIRONNEMENT

Suisse2013

PAR

Apiparn BORISUIT

1

Tothememoryofmygrandmother.

Abstract

2

i

Abstract

Visualcomfortatworkplaceshasoftenbeenconsideredintermsofdiscomfortglare,luminancedistributionandtaskvisibility.Besidesvisualeffects, thelightingenvironmenthasalsoimpactonhumanphysiologyandbehaviour.Theseeffectsof lightaretransmittedviaanovelclassofphotoreceptorsinthemammalianretina,whichwasdiscoveredonlyadecadeago.Sincethen,ithas become evident that light also plays an important role in regulating Non‐Image Forming(NIF)functionssuchascircadianrhythms,alertness,well‐beingandmood.Inlightingdesignitisaccordinglynecessarytotakeintoaccountnotonlyluminousintensity,butalsolight'sspectralcomposition, since the novel class of photoreceptors ismoremaximally sensitive to differentluminouswavelengthsthantheclassicalphotoreceptors(e.g.rodsandcones).

Themain focus of this doctoral thesis is on visual comfort assessment at workplaces. It washypothesized that the impact of light on visual comfort comprises not only luminancedistribution and/or luminous intensity, but also other qualitative aspects of the lightingenvironment.Officelightinginfluencesbuildingoccupantsintermsofvisualtaskperformance,alertness,healthandwell‐being.TheaimofthisthesiswastoassesstheimpactofofficelightingonvisualcomfortincludingNIFeffects.

Firstly,inordertomonitortheluminancedistributionwithinascene,anewphotometricdevicebased on a high dynamic range logarithmic visual sensor (IcyCAMTM) was set up. Aftercalibrationsandvalidations,thephotometricdevicewasusedtoassessluminancedistributionofofficespacesinaveryefficientway.Secondly,twoexperimentalstudieswereperformedwithhumansubjects,aimingtotest theacuteeffectsof lightonvisualcomfortvariables,subjectivealertness,moodandwell‐being.Lastly,thenoveldevicewasalsousedduringoneofthestudiestomonitortheimpactsofluminousdistributionovertimeandundervariouslightingconditions.

ThenovelphotometricdeviceenablestoassessluminousdistributionalsoincircadianmetricswithrespecttoNIFeffectsoflight.Theresultsfromthetwostudiesshowedtheeffectsofofficelighting includingdifferent sky conditionsand time‐of‐day changeson visual comfort andNIFfunctions.Inter‐individualdifferences,asassessedinextremechronotypes,alsohadaninfluenceonvisualcomfort.Interestingly,luminancedistributionwasnotonlyfoundtoimpactonvisualcomfortbutalsoonsubjectivealertness,moodandwell‐being.Toconclude,theresultsobtainedwiththenewdeviceprovideamorecomprehensivescientificframeworkandpracticalbasisforindoorlightingdesignatworkplaces.

Keywords: Lighting conditions, visual comfort,work environment, non‐image forming effects,photometricmeasurement,circadianrhythms

Abstract

ii

iii

RésuméLeconfortvisueldanslecadreprofessionnelasouventétéabordéentermesd'éblouissement,de distribution de luminance et de visibilité adaptée à la tâche effectuée. Au‐delà des effetsvisuels, l'environnement lumineux a aussi un impact sur la physiologie et les comportementshumains.Ceseffetssontprincipalementrégisparunenouvelleclassedephotorécepteurs,situéssurlarétinedesmammifères,découverteilyaunedizained'années.Depuislors,l'importancedu rôle de la lumière sur le rythme circadien, la vigilance, le bien‐être et l'humeur a étédémontrée.Ceseffets, indépendantsde la formationd'images(‘Non‐imageFormingEffects’or‘NIFeffects’),s'ajoutentauxeffetsvisuelsprécédemmentétudiés.Ilestdoncnécessairedetenircomptenonseulementdel'intensitélumineuse,maiségalementdelacompositionspectraledela lumière lors de projets d'éclairage, puisque cette nouvelle classe de photorécepteurs estsensible à des longueurs d'onde différentes des photorécepteurs classiques, que sont lesbâtonnetsetlescônes.

Lethèmeprincipaldecettethèseestl’étudeduconfortvisueldansl'environnementdetravail.Lepostulatdebaseestquel’impactdelalumièresurleconfortvisueldépendnonseulementdeladistributionet/oudel’intensitélumineuse,maisaussidelaqualitédesconditionsd'éclairage.Enoutre,ilestdésormaisévidentquel'éclairagedebureauxaffectelesperformancesvisuelles,lavigilance,lasantéetlebien‐êtredesoccupants.Lebutdecettethèseestd'évaluerl'influencedel'éclairagedebureauxsurleconfortvisuel,enincluantlesfonctionsNIF.Dansunepremièreétape,etafindepouvoirmesurerlarépartitiondesluminancesauseind’unescènevisuelle,unluminancemètre digital basé sur un senseur lumineux à grande dynamique et à réponselogarithmique (IcyCAMTM) a été mis sur pied. Deux études expérimentales ont ensuite étéréaliséesdanslebutdetesterleseffetslesplussignificatifsdelalumièresurdifférentsaspectssubjectifsduconfortvisuel,ledegrédevigilance,l'humeuretlebien‐êtreaucoursdelajournée.Finalement, lenouveaucapteurphotométriqueaétémisenœuvreaucoursdesesétudes: leseffets de la distribution lumineuse ont ainsi été mesurés au fil du temps sous plusieursconditionsd'éclairage.

L'utilisationdunouveaucapteurapermisainsid'atteindrelebutdecettethèseavecsuccès.Enplusdesmesuresphotométriques,celui‐cidonneaccèsàladistributiondeluminancedansuneunitédemesurecircadienne,quitientcomptedeseffetsNIFdelalumière.Lesrésultatsdesdeuxétudesontmis en évidence l'effet des conditionsd'éclairage, de la couverturenuageuse et del'heure sur le confort visuel et les fonctions NIF. Les différences entre individus, tels que leschronotypes extrêmes, ont aussi une influence sur le confort visuel. Par ailleurs, l'effet de ladistribution lumineuse a étédémontré,non seulement sur ladimension subjectivedu confortvisuel,maiségalementsurlavigilance,l'humeuretlebien‐être.Finalement,lesrésultatsdecettethèse fournissentunebase scientifique et pratiqueplus complète, en vuede la conception enéclairagedanslecadredel’environnementdetravail.

Mot‐clés:éclairage,confortvisuel,environnementdetravail,impactsnon‐visuels,mesurephotométrique,rythmecircadien

Abstract

iv

v

Acknowledgements

Throughouttheyearsofworkingonmythesis,manypeoplesupportedmewithexcellentadvice,helpandtime.Iwouldliketotakethisopportunitytoexpressmygratitudetoallofthem.

IamindebtedtoProf.Jean‐LouisScartezzini,thethesisdirectorforgivingmetheopportunitytoworkatLESO‐PB,EPFLandforprovidingmewithexcellentguidanceuntilthecompletionofthisthesis;

Furthermore,IwouldliketosincerelyacknowledgeDrMirjamMünch,mythesisco‐director,formuchvaluableadvice, supportand timeand for givingme theopportunity todomy thesisatLESO‐PB,EPFL;

IalsothankDrMartineKnoop,Prof.PhilippeThalmann,DrBernardPaule,andProf.LucaOrtelli,thepresidentofthejuryandthesiscommitteemembersfortheirtimeandcomments.

I amverygrateful to theVeluxFoundation for the financial support for this thesisandwouldalsoliketothanktheSwissFederalCommissionforScholarshipsforForeignStudents(FCS)thatgavemetheopportunitytocometostudyinSwitzerland.

Inaddition,IwouldliketothankEPFLandthePhDMobilityAwardsforgrantingmefundingforashort‐termvisitatLawrenceBerkelyNationalLaboratory(LBNL).

ManyheartfeltthanksgotoDrSigolènePangaudforherfruitfuladviceandsupportintheCLLS‐IcyCAM project and to the group of Sensory Information Processing at Centre Suissed’ElectroniqueetdeMicrotechnique(CSEM)forthecollaborationwithCLLS–IcyCAM.

Furthermore,IwouldliketoexpressmygratitudetowardsDrAnothaiThanachareonkitforalltherecommendationsandsupportbothinlifeandwork,atLESO‐PB,EPFLandatBerkeleyLabAdvancedWindowsTestbedfacility,LBNL.IwouldalsoliketothankEleanorLeeforhostingmeduring my visit at LBNL. Moreover, I gained considerable benefit in the field of variousdaylighting systems including valuable comments from Dr Steve Selkowitz, Dr Greg WardLarson,DrAndrewMcNeilandDrLuisFernandesfromthesamegroupatLBNL.

I would like to thank the Radiance community for the helpful guidance for using Radiance,Evalglare and the other simulation functions. Particular thanks to Lars O. Grobe, Dr DavidGeisler‐Moroder,DrJanWienoldandDrGregWardLarsonfortheirfruitfuladvices.

ManywarmthankstoDrJérômeKämpfandLaurentDeschampsfortheexcellentcollaborationin the CLLS – IcyCAMproject. Iwould also like to thankmy colleagues ‐ LenkaMaierova forexcellentteamworkduringtheChroliproject,DrFriedrichLinhartforeffectivecollaborationin

Acknowledgements

vi

theLiperproject,ChantalBasurtoforhelpingwiththeCLLS‐IcyCAM,andPierreLoeschforthetechnicalsupportinalltheprojects.

ItismypleasuretothankallmycolleaguesatLESO‐PBforprovidingmewithanunforgettableexperience, warmth and help. Particular thanks go to Dr Maria Cristina Munari Probst andChristian Roecker for providing excellent advice. I am also very grateful to the secretaries atLESO‐PB, Suzanne L’Eplattenier and Barbara Smith as well as the IT group for facilitatingprocedures at the laboratory and assisting me in many ways. Thanks to Diane Perez fortranslating the summary of this thesis fromEnglish to French and Iwould also like to thankSylvia Coccolo for generous support and food. I would like to thank all my colleagues whocontributedtomaketheLESOamorethanapleasantplacetowork,mypresentfellows:NikosZarkadis, Govinda Upadhyay, Antonio Paone, Olivia Bouvard, André Kostro, StefanMertin, DrNahidMohajeri,DrVahidNik,DrNicolasMorel,and the former fellows:DrFrédéricHaldi,DrUrs Wilke, Dr David Daum, Dr Paola Tosolini, Dr Adil Rasheed, Dr Virginie LeCaër, NicolasJolissaint,RaquelP.Gagliano,MariaPapadopoulou,AndreaCuéllarandMarjaEdelman.

I thankmy English editors, Barbara Smith andWipapanNgampramuan for proofreading andeditingthethesis,andalsoNapatRujeerapaiboonforhelpingmereviewsthethesis.

Lastbutnot least,Iwouldliketothankmyfamily,myparents,mysisterandbrotherandalsomy friends for their generous and wholehearted support. I dedicate this thesis to mygrandmother.Herunconditionalsupportwasaconstantsourceofmymotivationandstrength.

Lausanne,29November2013

vii

Contents

Abstract...................................................................................................................................................................................i

Acknowledgements...........................................................................................................................................................v

Contents..............................................................................................................................................................................vii

ListofFigures.....................................................................................................................................................................xi

ListofTables....................................................................................................................................................................xiii

ListofEquations...............................................................................................................................................................xv

Glossary.............................................................................................................................................................................xvii

Chapter1 Introduction.............................................................................................................................................19

1.1 Theroleofofficelighting....................................................................................................................19

1.2 Shortsummaryandopenquestions..............................................................................................25

Chapter2 Problemstatement................................................................................................................................27

2.1 Principlesofvisualcomfort...............................................................................................................27

2.2 Stateoftheartofresearch.................................................................................................................28

2.2.1 Impactoflightingonvisualcomfort....................................................................................................28

2.2.2 Lightingpreferencesforofficeworkers............................................................................................33

2.2.3 Brightnessperception................................................................................................................................36

2.2.4 Visualperformance.....................................................................................................................................37

2.3 Relevantquestionsandhypotheses...............................................................................................40

2.4 Structureofthesis..................................................................................................................................43

Chapter3 Photometricmeasurements..............................................................................................................45

3.1 Physicalparameters..............................................................................................................................45

3.2 Glarerisksassessment.........................................................................................................................49

3.3 HighDynamicRange(HDR)imagingtechniques.....................................................................51

3.3.1 Background.....................................................................................................................................................51

3.3.2 CalibrationofHDRImages.......................................................................................................................51

3.3.3 Practicalapplicationsofluminancemaps.........................................................................................52

3.3.4 LimitationsofHDRimagingtechniques............................................................................................53

Contents

viii

3.4 TheCamera‐LikeLightSensor(CLLS)...........................................................................................54

3.4.1 Principlesofthedevice.............................................................................................................................54

3.4.2 Calibrationprocedure................................................................................................................................55

3.4.3 ExperimentalvalidationofCLLS...........................................................................................................60

3.4.4 Outlook.............................................................................................................................................................66

3.5 ApplicationofCLLSforassessmentofnon‐imageformingeffectsoflight....................66

3.5.1 Assessmentofnon‐imageformingeffectsoflight........................................................................66

3.5.2 Calibrationprocedure................................................................................................................................67

3.5.3 Circadianweightedradiancecalibration...........................................................................................68

3.5.4 ExperimentalvalidationoftheCLLSwiththeC(λ)filter............................................................69

3.5.5 Outlook.............................................................................................................................................................72

3.6 Conclusion.................................................................................................................................................72

Chapter4 Non‐ImageForming(NIF)biologicaleffectsoflight...............................................................73

4.1 PrinciplesofNIFeffectsoflight.......................................................................................................73

4.1.1 Circadianrhythms.......................................................................................................................................74

4.1.2 Behaviouraleffectsinofficeworkers..................................................................................................75

4.1.3 Inter‐individualdifferences.....................................................................................................................77

4.2 LuminouspropertiesmodulatingNIFeffectsinhumans......................................................77

4.2.1 Illuminance.....................................................................................................................................................78

4.2.2 Spectralcompositionoflight..................................................................................................................78

4.2.3 Timeofday.....................................................................................................................................................80

4.2.4 IssuesofNIFeffectsinarchitectureandbuildingphysics.........................................................80

4.3 Conclusion.................................................................................................................................................81

Chapter5 Experimentalstudiesinofficelightingconditions...................................................................83

5.1 Impact of office day‐ and electric lighting conditions on visual comfort,alertnessandmood(ExperimentI)................................................................................................83

5.1.1 Participantsandmethods........................................................................................................................84

5.1.2 Results...............................................................................................................................................................86

5.1.3 Discussion........................................................................................................................................................93

5.2 Impact of different lighting conditions on visual comfort, discomfort glare,alertnessmoodandwell‐beinginextremechronotypes(ExperimentII).....................94

5.2.1 Subjectsandmethods................................................................................................................................94

5.2.2 Results...............................................................................................................................................................98

5.2.3 Discussion.....................................................................................................................................................111

5.3 Conclusion..............................................................................................................................................113

Contents

ix

Chapter6 Discomfortglareandsubjectiveassessments.........................................................................115

6.1 Impactofdiscomfortglareonvisualcomfort,alertnessandmoodinrealisticofficeday‐andelectriclightingconditions(ExperimentI)...............................................116

6.1.1 Experimentalmethods...........................................................................................................................116

6.1.2 Results............................................................................................................................................................117

6.1.3 Discussion.....................................................................................................................................................119

6.2 Impactofdiscomfortglareonvisualcomfort,subjective,alertnessmoodandwell‐being for various lighting conditions in extreme chronotypes(ExperimentII)....................................................................................................................................120

6.2.1 Subjectsandmethods.............................................................................................................................120

6.2.2 Results............................................................................................................................................................121

6.2.3 Discussion.....................................................................................................................................................126

6.3 Conclusion..............................................................................................................................................127

Chapter7 Conclusion..............................................................................................................................................129

7.1. Statementsandresearchquestions.............................................................................................129

7.1.1 ValuationofaCamera‐LikeLightSensor.......................................................................................129

7.1.2 Impact of office lighting on visual comfort includingNon‐Image Forming (NIF)effects.............................................................................................................................................................130

7.1.3 Impactofluminousdistributiononvisualandnon‐visualfunctions................................131

7.2. Recommendationsforofficelighting..........................................................................................132

7.3. Futureoutlook......................................................................................................................................133

References.......................................................................................................................................................................135

AppendixA......................................................................................................................................................................148

I. QuestionnairesforExperimentI..................................................................................................149

a) Visualcomfortvariables........................................................................................................................149

b) Non‐ImageForming(NIF)functions................................................................................................150

II. QuestionnairesforExperimentII.................................................................................................151

a) Visualcomfortvariables........................................................................................................................151

b) Non‐ImageForming(NIF)functions................................................................................................153

CurriculumVitae..........................................................................................................................................................154

xi

ListofFigures

Figure1.1:Conceptualmodelofthephysicalenvironmentandjobsatisfaction.......20

Figure1.2:Examplesofofficespaceswithdaylighting,issuedfromwindows............21

Figure1.3:Varioussimulatedeffectsoflightcreatedbydifferentlightsources.........22

Figure1.4:Relationshipsbetweenlighting,visualandnon‐visualeffects.....................23

Figure1.5:Linksofvisualandnon‐visualeffectsoflightingconditions.........................24

Figure1.6:Themodelofvisualcomfort,satisfactionandperformance.........................24

Figure2.1:MaximalluminanceratiosrecommendedbyIESNA.........................................29

Figure2.2:Relativefractionofobserversratingtheirtasklighting..................................31

Figure2.3:Summaryofpreferredilluminancesfromdifferentstudies..........................33

Figure2.4:Kruithof’slawastheassociationbetweenCCTandilluminance.................35

Figure2.5:Schematicviewofscotopic,mesopicandphotopicregions..........................38

Figure2.6:Conceptualmodelofvisualandnon‐visualsystems........................................42

Figure3.1:Relativesensitivityfunctions......................................................................................46

Figure3.2:Experimentalset‐upforthecalibrationofHDRimagingtechnique..........52

Figure3.3:Markedareasofthereferencetaskareasandtheirsurroundings.............52

Figure3.4:GlarerisksassessmentusingRADIANCEandEVALGLARE..........................53

Figure3.5:TheoriginalmodelofIcyCAMdevelopedbyCSEM...........................................55

Figure3.6:ImagetakenbyCLLSwiththreedifferentviewoptions.................................55

Figure3.7:Experimentalset‐upforthespectralsensitivitycalibration.........................56

Figure3.8:RelativespectralsensitivityassessmentsoftheCLLS.....................................57

Figure3.9:Experimentalsetupforthephotometriccalibration.......................................57

Figure3.10:Correlationbetweenthepixelgrayscalevaluesandluminances.............58

Figure3.11:VignettingeffectsoftheCLLSequippedwithafisheyelens.......................59

Figure3.12:Measurementset‐upforthevignettingcorrections.......................................59

Figure3.13:CharacterizationofthevignettingeffectsoftheCLLSfisheyelens..........60

Figure3.14:Locationsof15differentroomelementsfortheCLLSvalidation............61

ListofFigures

xii

Figure3.15:LuminancemappingextractedfromtheCLLSandaCCDcamera............62

Figure3.16:RelativedifferencesbetweentheCLLSandaCCDcamera..........................65

Figure3.17:Relativespectralsensitivityassessmentforcircadianmetric...................68

Figure3.18:CorrelationsbetweenpixelgreyscalevaluesandassociatedLec..............69

Figure3.19:LocationsofdifferentroomelementsforthevalidationoftheCLLS......69

Figure3.20:LecvaluederivedfromCLLSandspectroradiometer...................................70

Figure3.21:RoomAandRoomC....................................................................................................71

Figure3.22:Luminance,Lecandratio(Lec/L)forRoomAandRoomC.........................71

Figure4.1:SchematicoverviewofvisualandNIFpathwaysinthehumanbrain.......74

Figure4.2:Spectralsensitivitycurvesformelatoninsuppression....................................79

Figure4.3:Phaseresponsecurvetobrightlightwithrespecttothetimeofday......80

Figure5.1:Twolightingconditionsduringthestudy.............................................................84

Figure5.2:Averagedphotometricvariablesinthecourseoftheafternoon................87

Figure5.3:Averagedsubjectiveassessmentsofthevisualcomfortvariable..............88

Figure5.4:Averagedsubjectiveglareassessments.................................................................89

Figure5.5:Changesinphysicalwell‐beingandsubjectivealertness...............................91

Figure5.6:Availableday‐andelectriclightingsystems........................................................97

Figure5.7:TimecoursesofaveragedphotometricvariablesforBLandSSL............100

Figure5.8:Timecoursesofaveragephotometricvariablesundervariedskies.......102

Figure5.9:Averagedvisualcomfortvariablesforalllightingconditions...................103

Figure5.10:Timecoursesofvisualcomfortvariablesforbothchronotypes............104

Figure5.11:Effectsofweatheronsubjectiveglareovertime..........................................105

Figure5.12: TimecoursesunderSSLforsubjectivealertnessandmood...................107

Figure6.1:SampleHDRimagewithpotentialglaringsources........................................117

Figure6.2:Dynamicsofaveragedglareindicesinthecourseoftheafternoon........118

Figure6.3:HDRimagesotakeninthemorning,afternoonandevening......................121

Figure6.4:Averagedglareindices...............................................................................................122

Figure6.5:Averageglareindicesunderdifferentskyconditions...................................124

xiii

ListofTables

Table3.1:Radiometricandphotometricvariables............................................................45

Table3.2:Criteriaofglaresensations......................................................................................50

Table3.3:R2ofluminancemappingbytheCLLSandluminancemeter...................62

Table3.4:Glareindicesobtainedfromimages....................................................................63

Table3.5:GlareindicesassessedbytheCLLSandaCCDcamera................................64

Table5.1:Summaryofthetwolightingconditionsduringtheafternoons..............85

Table5.2:Summaryofthetwolightingconditionsduringtheafternoons..............87

Table5.3:AveragedphotometricvariablesduringDLandEL......................................90

Table 5.4: Correlation coefficients between visual comfort variables, non‐visualfunctionsandlightproperties.....................................................................................................92

Table5.5:Averagedhabitualwake‐andbedtimes,studybeginandendtimes.....95

Table5.6:Criteriaforthethreedifferentlightingconditions........................................96

Table5.7:Averagedphotometricvariables...........................................................................98

Table5.8:Averagedphotometricvariablesunderdifferentskyconditions..........99

Table5.9: Correlation coefficients of photometric variables with visual comfortvariables,andnon‐visualfunctions.......................................................................................109

Table5.10:Correlationcoefficientsbetweenandwithinvisualcomfortvariablesandnon‐visualfunctions............................................................................................................110

Table 6.1: Correlation coefficients between glare indices and subjectiveassessments.....................................................................................................................................119

Table6.2:Averagedglareindicesandbackgroundluminanceunderdifferentskyconditions.........................................................................................................................................123

Table 6.3: Correlation coefficients between glare indices (DGP, DGI, UGR) andvisualcomfortvariablesandnon‐visualfunctions..........................................................125

xv

ListofEquations

Equation2.1........................................................................................................................................29

Equation2.2........................................................................................................................................29

Equation3.1........................................................................................................................................49

Equation3.2........................................................................................................................................50

Equation3.3........................................................................................................................................50

Equation3.4........................................................................................................................................56

Equation3.5........................................................................................................................................56

Equation3.6........................................................................................................................................58

Equation3.7........................................................................................................................................60

Equation3.8........................................................................................................................................60

Equation3.9........................................................................................................................................67

Equation3.10.....................................................................................................................................67

Equation3.11.....................................................................................................................................68

Equation3.12.....................................................................................................................................68

Equation3.13.....................................................................................................................................68

xvii

Glossary

Generalterms

NIF Non‐ImageForming

ipRGC intrinsicallyphotosensitiveRetinalGanglionCells

HDR HighDynamicRange

CLLS Camera‐LikeLightSensor

f1' CIEdeviationoftherelativespectralresponsefromtheV(λ)functionPhotometricvariables

Ev Verticalilluminance(lx)CCT CorrelatedColourTemperature(K)CRI ColourRenderingIndex

Eec Circadianweightedirradiance(W/m2)

Lec Circadianweightedradiance(W/sr·m2)LightingconditionsDL DaylightingconditionEL ElectriclightingconditionDIM DimlightingconditionsBL BrightlightingconditionsSSL Self‐selectedconditionsGlareindicesDGI DaylightGlareIndexDGP DaylightGlareProbabilityUGR UnifiedGlareIndexSubjects ET extremeEveningTypesMT extremeMorningTypes

19

Chapter1 Introduction

1.1 Theroleofofficelighting

Thequalityoftheindoorenvironmentwithinofficebuildingsinfluenceswhite‐collarworkersintheirbehaviourandwithregardtoorganizationaloutcomes.Anoptimalindoorenvironmentcanimprove comfort, productivity, health and well‐being [1‐3]. It was reported that workers’satisfaction with regard to their work environment induced greater job satisfaction andorganization commitment as well as less absentees [4‐7]. Environmental satisfaction derivesfromseveralfactors;lightisveryimportantinthisprospect:itrepresentsthebulkofthisthesis.

Intheframeworkofalargeinterdisciplinaryprojectdedicatedtoworkers’satisfactioninopen‐planofficeconditions[5,8,9],datawerecollectedfrom779workstationswithin9buildingsinCanada and US cities during two years (2000‐2002). Physical variables at individualworkstations,suchaslighting,acoustic,thermalandairmovementconditionsweremonitoredfor thatpurpose.Theparticipantswererequestedtoexpress theirsatisfactionregardingtheirworkenvironment,aswellastheiroveralljobsatisfactionbymeansofasurvey.Thestatisticalanalysis of the survey indicated the existence of a clear link between environmental and jobsatisfaction(Figure1.1).Theoverallenvironmentalappraisalwasaffectedbytheworkers’ownsatisfactionwiththephysicalvariablesregardingventilation,privacy,acousticsandlighting[5].In a subset analysis, predictor variables related to satisfaction with lighting conditions wereanalysed on the basis of the monitored data. Five variables that specifically related tosatisfactionwithlightingwere:i)preferredilluminance,ii)glaresensationsfromVideoDisplayTerminal(VDT) iii), lightinguniformity, iv) lightdirectionality(ratioofhorizontalandverticalplanes),aswellasv)presenceofawindow.Themostimportantvariablewasthelastone:thefact of having a window next to the work space, or the fact of benefitting from daylight,significantlyimprovedsatisfactionofofficeworkerswithregardtolighting[8].

Introduction

20

Figure 1.1: Conceptual model showing the relationships between the physical environment and jobsatisfaction, basedon thedata fromquestionnaires (modified from [5, 9]). Thepredictor variables forsatisfactionwithlightingareindicatedontheleftside(orangerectangles)(adaptedfrom[8])

Interestingly, according to the study, lighting satisfaction does not only imply sufficientbrightnessforthesakeoftaskvisibility.Therelatedphysicalfactorsalsoincludevisualcomfortand the presence of windows. Figure 1.2 illustrates office environments benefitting fromdaylightaswellasfromthepresenceofwindows.

Introduction

21

a)

b)

Figure1.2: Examplesof office spaceswithdaylighting fromwindows; a) large openoffice at LufthansaheadquartersinFrankfurtairport,Germany[10];photographybyH.G.Esch;permissionby“ingenhovenarchitects”;b)smallofficespaceatLESOSolarExperimentalbuilding(EPFL,Lausanne,Switzerland).

Differentcoloursof lightsources impact theatmosphereofaworkingspace;day‐andelectriclighting can create different atmospheres of light. Figure 1.3 describes these different effectssimulatedbyKarcher[10].Anofficewithdifferentlightingsystemsisillustratedbysimulationsincluding: daylight from the window (Figure 1.3a), daylight and incandescent lamps (Figure1.3b), daylight and fluorescent lamps with Correlated Colour Temperature (CCT) at 4000 K(Figure1.3c), incandescent lampsonly(Figure1.3d)andfluorescent lampwithCCTat4000Konly (Figure 1.3e). Our perceptual system is affected by different lighting conditions, such aslight sources, colour temperature and luminous intensity; this influences our mood andbehaviour.Theeffectsof lighthavebeenfurtherexploredinordertounderstandhowlightingimpactsourdailylifeandwhattheotherconsequencesare.

Introduction

22

a)

b) c)

d) e)

Figure 1.3: Various simulated effects of light created by different light sources: a) daylight from thewindow; b) daylight and incandescent lamps; c) daylight and fluorescent lampswithCCT at 4000K; d)incandescentlampsonly;e)fluorescentlampsonly.TheseimagesweresimulatedbyKarcherforElectricGobo,Berlin,Germany[10].ThesehavebeengivenpermissiontobereproducedinthisworkbyERCOGmbH,www.erco.com.

A decade ago, a new class of photoreceptors in the retinal ganglion cells (the so‐calledintrinsically photosensitive Retinal Ganglion Cells; ipRGC) were discovered; they are mainlyresponsible for the non‐visual light perceptionwhich regulates circadian functions, the pupil

Introduction

23

lightreflexandhormonalsecretions[11‐15].Sincethen,ithasbeendemonstratedthatbesidesvisualeffectsoflight(perceivedbyrodsandcones),therearealsonon‐visualbiologicaleffects:thesocalled“Non‐ImageForming(NIF)”effectsoflight,whichimpactmanycircadianfunctionsinhumans, aswell as alertness andwell‐being (seeChapter 4). Emotion is impactedbyboth,visualandNIFeffectsoflightasillustratedinFigure1.4[16].Visualeffectsoflightaffectvisualperformance;NIFeffectsoflightimpacthealthandwell‐being.

Figure1.4:Conceptualrelationshipsbetweenlighting,visualandnon‐visualeffectsof light,asdescribedbyVanBommel[16]

ThegroupsofBoyceandVeitchperformedtwoexperimentswith188participantstoassesstheimpactofdifferent(mainlyartificial)lightingconditionsonparticipants’behaviourwithseveralquestionnaires [17]. According to the results, Veitch et al suggested different links betweenlightingeffects(theso‐called“linkedmechanismsmap”),asillustratedinFigure1.5[18].Theirmodelincludedtwopaths:anappraisalpath(lightperception)andavisualpath.Viathevisualpath, lighting conditions fostered the visibility leading to better task performance; thismightreflect the “visual effects” of light. The appraisal path describing the influence of lightingconditionswasrelatedtohealthandwell‐beingintermsof lightingpreferenceandmood.Theauthors pointed out a positive correlation between lighting appraisal and visual comfort: ahigher appraisal of lighting conditions was linked with a better visual comfort. Initially, theauthorshypothesizedthatvisualcomfortwasdirectlyaffectedby lightingconditions,andalsorelatedtovisualcapabilitiesandtaskperformance.However, theycouldnotdemonstrate thatthese linkswerestatistically significant: theyexplained thisby the fact that lightingquality inthestudywasratedhigh,whilevisualdiscomfortratingswereratedlow[18].

Introduction

24

Figure1.5:Linksofvisualandnon‐visualeffectsoflightingconditions.Solidlinesshowtheresultswhichweresupportedbymediatedregressionanalysis(heavyblue linesshowtheappraisalpathanddouble‐headedarrowsshowthevisionpath);dottedlinesshowresultswhichweresupportedbythestatisticalanalysis(MANOVAandANOVA);solidlinesshowweakerlinkssupportedbymediatedregressionresults(accordingto[18]).

Visualcomforthasbeengenerallyconsideredasapredictorof lightingquality,whichincludesvisualcomfort,satisfactionandperformance,theso‐called“CSP”index.Thisindexwasbasedona survey of officeworkers’ opinion regarding lighting quality [19, 20]. The authors suggestedthat the appraisal of a lighting environment depended on these three elements (Figure 1.6).Visualcomforthasgenerallybeenconsideredtobefullyindependentfromvisualperformance:lightingconditionsfavourableforvisualperformancemightbeuncomfortable.Visualcomfortisa human‐responsive factor linked to occupants’ expectationwhich can change over time [21,22]:thismayimplythatvisualcomfortcouldbelinkedbothtovisualandtoNIFeffectsoflight.

Figure1.6:Thethreedimensionsofvisualcomfort,satisfactionandperformance,definingthe‘’CSP’’indexaccordingto[20]

Introduction

25

1.2 Shortsummaryandopenquestions

Satisfaction with lighting is related to job satisfaction, which contributes to organizationaloutcomes.Lightingconditionsalsohaveaninfluenceonoccupants’individualbehavioursuchasmood, well‐being and task performances. Light induces both visual and non‐visual biologicaleffects;thelinkbetweenthesetwoopposedpathshassofarremainedunclear.

Internationalstandardsandlightingrecommendationsforworkspaceshaveexistedforseveraldecades. However, most defined criterions aimed solely at meeting visual requirements forrelated visual tasks. There are doubts whether or not the current lighting norms are alsoadequate for other human factors such as health and behaviour, leading to better workperformance and organizational outcomes. To take into account the NIF effects of light, it isnecessary to optimize lighting conditions in order to satisfy both visual and non‐visualrequirements.Thismightleadtoboth,betterworkperformanceandbetterhealth.

Visual comfort is themain topic of this doctoral thesis: it is assumed to be the link betweenvisual andNIF effectsof light.As such, the impactof light onvisual comfortwas investigatedwithrespecttobotheffects.Itwasexpectedthatamorecomprehensivescientificbasisforthedefinitionof”healthylighting”atworkplacescouldbecreated,tobeusedbypractitionersinarathershortfuture.

27

Chapter2 Problemstatement

2.1 Principlesofvisualcomfort

The definition of visual comfort has been extensively debated for more than half a centurywithoutauniversallyacceptedversion. Ithasgenerallybeendefinedasa lightingsituation intheabsenceofvisualdiscomfort [21,22].However, thisdefinition isstill insufficient todefinevisualcomfortforseveralreasons:

• Alightingsituationthatprovidesbettervisualcomfortisnotnecessaryrelatedtohigherluminousintensity,whilethelatterisassociatedwithhighertaskvisibility[21,22].

• Peoplecantolerateglaresensations(asacauseofvisualdiscomfort)fromadaylightingconditionbetterthanfromanartificiallightingcondition[23].

• Visualcomfortcanchangeovertimewhilethecausesofvisualdiscomfortdonotdiffer(astestedunderartificiallightingconditions)[17,24].

Generally,theterm‘comfort’indicatestheperceptionofwell‐beingandaesthetics[25]andtheterm‘visual’relatesto'seeing'orthesenseofsight.Inthiswork,‘comfort’couldthusbedefinedas[26]:

• Theabsenceofdiscomfort• Positivefeelingsofwell‐being• Apleasantstateofphysiological,psychologicalandphysicalharmonyofahumanbeing

withher/hisenvironment([27]referencedin[28]).

Accordingtotheabovestatements, lightingconditionsshouldavoidcausingvisualdiscomfort.Whatarethecausesofvisualdiscomfort?

Since the visual system perceives 'information' from a visual environment, the aspects thatinterrupt the abilities of 'information' perception can be defined as potential causes of visualdiscomfort,suchasvisualtaskdifficulty,under‐orover‐stimulationbylightexposureinascene,distractionwithin the visual field andperceptual confusiondue to the luminancedistribution[21]. After excluding the causes of visual discomfort, a lighting conditionwith optimal visualcomfortmayleadtopositivefeelingsintermsoflightingpreferences,aesthetics,brightness,orwell‐being. Moreover, the ‘comfort’ sensation is known to be affected by physiological andpsychological interactions of humans with their environment [28]. Visual comfort shouldtherefore also comprise lighting conditions optimized for human psychological andphysiological well‐being whichmeans not only visual functions but also Non‐Image Forming

Problemstatement

28

(NIF)effectsoflight.Withrespecttothelatter,thelightingconditionsshouldmaintainoptimalsleep‐wake homeostasis in humans. Healthy lighting conditions mean not only luminousconditionsduringdaytime;theyconcernalsooptimallight(ifnecessary)duringdarkperiodsforthesakeofgoodhealth.

This doctoral thesis ismainly focused on the impact of light on visual comfort as opposed tovisual performance. One reason is that it is the most effective method to determine lightingqualityduetothefactthatthechangesofvisualcomfortaremoresensitivethanthoserelatedtovisualperformanceperse[29]. Althoughoptimalvisualcomfortenhancesvisualperformance,uncomfortable lighting conditions do not always reduce visual performance ([30, 31] asreferencedin[22]).

At present, visual comfort cannot be objectively quantified due to the fact that no empiricalmodelofvisualcomfortisgenerallyaccepted.Somenumericalmethodscanhoweverbeusedtodetermine luminous distribution within spaces. Moreover, the luminous distribution can beconsidered as a key factor for visual discomfort depending on glare and luminance ratios.Furthermore,lightingqualityalsoneedstobeassessed;thiscanbedonewithhumansubjectsbyapplying subjective assessments. Until now, reliable human and physical factors referring tovisualcomfortarenotwellcircumscribed:itisimportanttounderstandhowlightingconditionscanhavean impactonvisualcomfort.Mostof thepublishedvisualcomfortassessmentscalesfocuson lightingpreference and lightperception,manyof themare finally resulting in visualperformanceassessments[20,32‐37].'Visual'performancemeanstaskperformancewhichonlyrelatestothevisualsystem.However,'task'performancecanbeinfluencedbyboth,visualandnon‐visual aspects [38]. Photometric variables, such as the luminous intensity, luminousdistribution or colour appearance of light might also influence visual comfort. In order tocombine these effects with subjective ratings, simultaneous lighting monitoring and visualcomfort assessments are needed. A comparison between the physical properties of light andsubjectiveassessmentscanfosterabettercomprehensionofthefundamentalsofvisualcomfort.

The next section describes the impact of office lighting on office workers. Visual discomfort,lightingpreference, brightnessperception and visual performancewill be discussedbased onresults from the literature. Several findings from physical light properties, such as luminousdistribution, luminous intensity, colour appearance of light and light sources (daylight vs.artificiallight)willalsobecovered.

2.2 Stateoftheartofresearch

2.2.1 Impactoflightingonvisualcomfort

Luminousdistribution

Luminous distribution is a key factor underlying the causes of visual discomfort withinworkspaces.Theratiobetween theverticalplane illuminanceand theworkplane illuminance

Problemstatement

29

must be checked particularly in office rooms equipped with Video Display Terminals (VDT).According to the recommendations of the Illuminating Engineering Society of North America(IESNA) [22], luminance ratios between the task and their immediate surroundingsmust notexceed1:3(or3:1)andshouldnotbelargerthan1:10(or10:1)formoredistantsurroundings.Similarly, the ratiobetweenaVDTand its immediateandremotesurroundingsshouldnotbelargerthan1:3and10:1,respectively(seeFigure2.1).Inaddition,theluminanceratiooutsidethefieldofviewmustbelowerthan1:40[39].

Figure2.1:MaximalluminanceratiosrecommendedbyIESNAforVDTworkstations(after[22]and[39])

However, La Toison [40] states that “objective contrast, such as the ratio of luminances (R;Equation2.1),doesnotallowlightingsituationstobecharacterizedwherethesubjectiveaspectisdominant.Apparentcontrast(C;Equation2.2),asadvancedbyBodmann[41]fordiscomfortglare, is completely different, since there is a difference in brightness”. This implies that theapparentcontrast,whichisgenerallyusedinformulaeofdiscomfortglare(orglareindices),ismorerelevantfortheassessmentofvisualdiscomfortthananobjectivecontrast(oraluminanceratio).

R=L2/L1 (2.1)

C=ΔL/L1 (2.2)

whereL1,L2aresurfaceluminanceofagivenarea(resp.task/background)[cd/m2].

Discomfort glare can be induced by poor luminance distribution, generally caused by anextremelyhigh luminance source, located in the fieldofview.Twomainglare situationshavebeen identifiedanddescribedas:1)excessive luminancecontrasts in the fieldofview,and2)(over‐) saturation of the visual system. Excessive contrasts are usually caused by very brightsurfaceswhichareperceivedinamuchdarkerenvironment,suchasapocketlightspotonthefloorinacellar.Saturationaffectsthevisualsystem,whentheretinaisstimulatedbyatoolargeflux of light when for example a pocket light beam is oriented toward the eyes. Daylightingwithinbuildingscaninduceeithereachoftheminanindividualmanner,orthetwosituations

Problemstatement

30

together. Glare reduces visual performance by impeding certain visual tasks and is thereforecalled‘disabilityglare’.Besides,glarethatleadstoasensationofannoyance,eyestrainorpainwithoutaffectingvisualperformanceiscalled‘discomfortglare’[42].

Both daylighting and artificial lighting conditions can cause discomfort glare, althoughoccupantsmaytoleratedaylightingandartificial lightingconditions invariousdegrees.Hence,therewasadevelopmentofglareindicestoassessthedegreesofdiscomfortglareunderagivenlighting situation. Glare indices, as well as the way to calculate them using photometricvariables,willbediscussedinmoredetailsinChapter3.

Luminousintensity

Ingeneral,ithasbeeninvokedthatlightingconditionsleadingtohightaskilluminanceimprovevisual performance; nevertheless, this is not always the case for visual comfort. A studyconductedbyMuckandBodmanninvestigatedvisualcomfortandvisualperformanceachievedfor different task illuminance ([31]as referenced in [21, 22]). In this study, a group of youngvolunteersperformedavisualtaskbyidentifyingaspecifictwo‐digitnumberamongsthundredothernumbers.Theilluminanceinthisstudyvariedbetween50and2000lx;higherworkplaneilluminanceledtobettervisualperformanceandvisualcomfortinthisilluminancerange.Workplane illuminance exceeding 2000 lx led to a lower visual comfort and a higher visualperformance at the same time. This implies that optimal visual comfort conditions areapparently associated with higher illuminance, but only to a certain level: when a certainilluminancethresholdisexceeded,lightingconditionsinducelowervisualcomfort.

Balder and colleaguesalso investigatedvisual comfort fordifferent luminance values ([11] asreferencedin[2]).Thevolunteersinthisstudywereaskedtoqualifythelightingontheirdesksusingthreecategories,namely,i)“toodark”,ii)“good”oriii)“toobright”.Figure2.2showstheresultsforeachcategoryoflighting.Itcanbeseenthatthelargestfractionof“good”appraisalsappearsonlywhentheluminanceatthedeskiscloseto130cd/m2.Thisstudyexperimentallyconfirmed that “good lighting” is not necessary the one showing the largest photometricalvariables(e.g.luminanceorilluminance).Toobtainoptimalvisualcomfortconditions,alightingenvironmentshouldbalanceallthesevariables.

Problemstatement

31

Figure2.2:Relativefractionofobserversratingtheirtasklightingas“toodark”,“good”or“toobright”fordifferentluminance(cd/m2)(after[30]referencedin[22]).

Colourappearanceoflight

Thespectralcompositionoflightsourcesdeterminesthecolourappearanceofthelightflux;italsoinfluencesindoorlightperception.Generally,thecolourappearanceofalightsourcecanbecharacterisedbythewayoftheCorrelatedColourTemperature(CCT)andtheColourRenderingIndex(CRI)[43].Thelightspectrumorthespectralpowerdistributionofalightsourceisusedas a metric in some studies; it indicates the colour characteristics of a light source for eachwavelength over the visible part of electromagnetic radiation. More details about the colourappearanceoflightaregiveninChapter3.

Sofar,therehasbeenlittleevidencetosupporttheimpactoflightcolourappearanceonvisualcomfort; most studies only provide evidence of the impact of colour appearance on lightingpreferenceorbrightnessperception(seeSections2.2.2and2.2.3).

Astudyinvestigatedtheimpactofstreetlightingonvisualcomfortbetweenthreedifferentlightsourceswith two typesofmetal halide lamps (CCT=2800K,CRI=83 andCCT=4200K, CRI=90)and a high pressure sodium lamp (CCT=2000K, CRI=25). In this study, the observers rated ahigher visual comfort from the twometal halide lamps in comparisonwith thehighpressuresodium lamp. Interestingly, no differences were found in comfort ratings between the lightgeneratedbythetwometalhalidelampsat2800Kand4200K.IthastobenotedthattheCRIofthe sodium lampwas very low (CRI=25)whichmight have induced an uncomfortable visualambiance. The result of CCT is in agreement with another indoor lighting study whichinvestigated the lightappraisalsof subjects fromtwodifferentCCTs (sameCRI=80)at500 lx,750 lxand1000 lx fromaworkplane illuminance.Thesubjects ratedahighervisualcomfortandspaciousnesswith4000Kthanwith2700Kforallilluminancevalues.

Problemstatement

32

However,thisiscontradictedbyastudyofcolouredlightsources,havingaircraftpassengersasparticipants.ThestudyinvolvedtheuseofmultichipLEDlightsourceswhichweresettofourdifferent colours. The first two types generated a yellow colourwith centroidwavelengths at603.2 and 607.1 nm, while the other two types generated a blue colour with centroidwavelengthsat496.8and504.6nm.Unfortunately,therewerenoCCTorCRIdataprovidedinthe publication. The results show that the observers did not express different visual comfortsensations for the four different lighting conditions, even though they rated differentperceptions of brightness and colour temperature sensations. They expressed the feelingof a‘darker’and‘warmer’atmosphereundertheyellowlight,whilehavingthefeelingofa‘brighter’and‘cooler’atmosphereunderthebluelight.Furtherresearchisstillneededtounderstandthebasicsandimpactoflightcolourappearanceonvisualcomfort[44].

Daylightvs.artificiallight

Thereisevidencethatofficeoccupantstolerateglaresensationsfromwindowsbetterthanfromartificial light sources [23,45]; onemain reasonmightbe thedifferent sizesof glare sources.This is the reason why daylight glare indexes were developed in a different way than thosededicated to artificial lighting. Many prior studies dedicated to office lighting involvingdaylightingfoundastrongbiasinfavourofviewsthroughwindowswithinoccupants’surveys.The results show that the further the occupantswere seated from awindow, the lower theirsatisfactionwas[8,46‐48].

The view throughwindows can also lead to glare sensations and visual discomfort. [49, 50].TuaycharoenandTregenzacarriedoutaninvestigationundercontrolledlaboratoryconditions.Inthisstudy,volunteershadtorate ‘interesting’viewimagesprojectedonasmallscreen.Theresultsshowthatalargerinterestintheprojectedimagewasassociatedwithalargertoleranceregarding discomfort glare [35]. The Unified Glare Rating (UGR), a glare index set‐up forartificial lighting condition [51], was a poor predictor for discomfort glare in this case. Theauthors also conducteda study in a20‐floorbuilding inSheffield (UK) inorder to investigatediscomfortglarewithinidenticalofficeroomsatdifferentfloorsandwithdifferentorientations.Thisstudycompareddifferentwindowviewsbetweenanopaquewall(noview),urbanscenes(low‐scoring view), and the view on a landscape scene (high‐scoring view). It appeared thatdiscomfortglareratingsdecreasedwhentheinterestfortheviewincreased.Hence,twofactorsforanewdaylightglareindexwereproposed:1)theinterestfortheviewand2)thevariationsof luminance depending on the changing weather conditions. The authors developed anempiricalequationforanovelglareindexbasedontheDaylightGlareIndex(DGI)whichismoreappropriatefordaylightingstudies[49].

Ontheotherhand,anegativeimpactofdaylightinthepresenceofwindowswasalsofoundbyRoche et al.; these authors reported that dissatisfactionwith daylight was also expressed byoccupants when the average daylight factor was over 5%, with complaints regarding thereflectionfromthesunandglare[48].AnotherstudybyAriesshowsthatoccupants,whowerelocatedclosertothewindowsexpressedmorethermalandglarecomplaints[46].

Problemstatement

33

2.2.2 Lightingpreferencesforofficeworkers

Luminousintensity

Althoughwell‐established international standards for illuminance exist [22, 52], only little isknown about inter‐individual and subjective light preferences of building occupants. Thepreferred illuminanceonahorizontalworkplanewasreported tobeeither larger [53,54]orlower[36,55,56]than500lxaccordingtointernationalrecommendations[22,57]).

Fotios andCheal [58] reviewed the rangesof preferred illuminance fromdifferent studies, asillustrated inFigure2.3.These authors examinedpreferred illuminancesusing an adjustmenttask,whereparticipantscouldsettheirtaskilluminanceaccordingtotheirpreferencewithintheavailableilluminancerange(so‐calledthe“stimulusrange”).Figure2.3illustratesthepreferredilluminancesandstimulusrangesofeachstudy.Scholzetal.[59]aswellasBoyceetal.[60]alsoinvestigatedmultiplestimulusranges.Scholzetal.[59]determinedtheoptimaltaskilluminanceset for 50 anesthetists which was used in order to visualise larynxes during laryngoscopies.Three different types of discharge lamps were used in these studies, namely 1) vacuum, 2)xenon and 3) halogen lamps. Each light source provided different ranges of illuminance, e.g.‘vacuum’=5‐327lx,‘xenon’=11‐2583lx,‘halogen’=10‐2557lx.Boyceetal.[60]observedthepreferredilluminanceinawindowlessofficewhichcomprisedtwostimulusranges(low:7‐680lx, high: 12‐1240 lx). They found that the preferred illuminances for the same room weredifferent(low:398lx,high:518lx)duetothestimulusrange.VeitchandNewsham[36],Mooreetal.,Boyceetal.[61]aswellasBegemannetal.[62]alsoconducteddifferentstudiesinofficerooms,while an investigation of Junlèn et al. [54]was carried out in awindowless industrialworkspace.TheaveragepreferredilluminancesandstimulusrangesobservedinthesestudiesareillustratedonFigure2.3.

Figure 2.3: Summary of preferred illuminances from different studies (after [58]). The dashed lineindicatesthe500lxonahorizontalworkplanwhichistheminimumrequirementatmostworkingplaces.

Problemstatement

34

According to their investigations, Fotios and Cheal argued that the average preferredilluminance is often close to the middle of the available illuminance ranges, the so‐called“stimulusrangebias”.Interestingly,theyalsoconductedanotherstudy,wherethesubjectswereaskedtoevaluatetheirpreferredilluminancefromthreedifferentavailableilluminanceranges,namely, 1) low: 48‐1037 lx, 2) middle: 83‐1950 lx, 3) high: 165‐2550 lx) [58]. The resultssupportedtheirhypothesisthattheaveragepreferredilluminanceapproachedthecenterofthestimulusrange(forallthreeilluminanceranges).Logadóttiretal.[63]usedasimilaradjustmenttask and confirmed the results of Fotios et al. Furthermore, they noted that the initialilluminanceusedatthebeginningofthestudy,theso‐called“anchor”,affectedalsothepreferredilluminance: higher anchors settings (higher initial illuminance at the starting point) led tohigherpreferredilluminance[63].

Colorappearanceoflight

ItisgenerallyarguedthatofficeoccupantspreferhighCCTlightsourcesinthepresenceofhighilluminance,andlowCCTlightsourcesinthepresenceoflowilluminance.ThesefindingswerefirstreportedbyKruithof[64]andwerecoinedas‘Kruithof’slaw’.Figure2.4showstherangeofpreferred illuminance levels versus the CCT of different light sources. However, the originalstudyhasnodatareferringtotheCRIofthelightsources.

Many subsequent studies were not in agreement with the Kruithof’s law. According toinvestigationsbyDavisandGinthner[65],arelationshipbetweenCCTand lightingpreferencewasnotfound.Itseemsthatlightpreferenceismoreassociatedwithtaskilluminance[65]thanwithCCT.BoyceandCuttle [66] reported that theCCTof lampsvirtuallyhadnoeffecton theobservers’ impression regarding lighting; higher illuminance led to amore positive appraisalregarding the room appearancewhichwas qualified asmore pleasant andmore comfortable[66].

Problemstatement

35

Figure2.4:Kruithof’s lawastheassociationbetweenCCTandilluminanceofagivenlightsource:thex‐axisshowstheCCT(K),andthey‐axisshowsilluminance(lx).Thewhitearearepresentstheilluminationconsideredasbeing“pleasant”.Thelowerbluesectionrepresentstheilluminationconsideredas“cold”,whileintheupperorangesectiontheilluminationisconsidered“unnatural”(after[64]).

It was found that colour rendering properties, rather than CCT, were the critical judgmentvariableforlightingconditionsquality[67].CorrelationsbetweenCRIandlightingpreferenceswerealsofound:higherCRIsareassociatedwithgreaterlightpreferencescores[68,69].FotiosandCheal[69]evaluatedtheimpactoflampspectrumonlightingpreferencesbyusingside‐by‐sidebooths.Thesubjectshadtodeterminethepreferredcolourappearanceoftheirownhand(naturalnessoftheskincolour),ofacolourchart(preferredcolourappearanceonthechart),aswell as of an illuminated space from two identical models. The experiment employed fivedifferentlamptypes:twodifferentmetalhalidelamps(MH2,CPO),acompactfluorescentlamp(CFL2), ahigh‐pressure sodium lamp (HPS) and twoLED light sourceswithdifferent colours.Thelampswereequippedwithatranslucentdiffuserinordertoavoiddifferencesinthespatialdistributionoflight;14pointsontheboothsweremaintainedatconstantluminancesunderthedifferent light sources. Two booths were illuminated for comparative purposes using twodifferent light sources of equal illuminance and brightness. The results showed that the CRIprovidedthestrongestcorrelationwithlightingpreference(R2=0.85),whencomparedtotheGamutAreaIndex(GAI),CCT,scotopicandphototicratio(S/P)andtheCIEmesopicsystem.


Access to daylight generally improves users’ satisfactionwith lighting; it is preferred to pureelectriclightingconditions[8,48,70,71].Itiswidelyknownthatdaylightismoredesirableforpsychological reasons, environmental appearance and pleasure [71‐73]. Daylight is ratedsuperior to other light sources: it provides higher illuminance and allows a better colourdiscrimination and office appearance [71, 73, 74]. Some studies mention that a workplacelocatednearthewindowissignificantlymoreattractiveandhigherranked[8,36,46‐48,75,76]

Problemstatement

36

thanonelocatedfurtherintheroom,duetotheopportunityofaviewthroughwindows,aswellas the ability to identify the weather and to “open thewindow” to increase ventilation [75].However,windowsalsoshowednegativeimpactsforthosepeoplesittingnearby,suchasglare,thermaldiscomfortandlackofprivacy[46,48,76].

Begemannetal.conductedastudywith96occupantsintheirusualofficelightingenvironmentduringdaytimeworkinghours[53].Theyfoundthatunderdaylightingconditions,thepreferredworkplane illuminances were much higher than the values recommended by current indoorlighting standards (500 lux at desk), which significantly correlated with the biologicalstimulation (as assessed by subjective sleep quality). The authors finally concluded that‘biological’ lighting needs were very different from visual requirements. Another interestingfindingofthisstudywasthatlightingpreferencesincombinationwithtaskvisualperformancevaried strongly among individuals. Since then, several studies investigated inter‐individualpreferencesandinfluencesoflightonbiologicalfunctions[77‐79].

Otherphysicallightingproperties

Dynamic lighting for office spaces has recently been used in order to mimic the decreasingnatural light at dusk and to achieve energy savings for electric lighting. Begemann [53]suggested that people preferred a lighting environment which follows the daylight cycle tosteadystatelighting[53].KortandSmolders[79]conductedastudyaboutthedynamiclightingeffectbysettingahigherworkplane illuminanceandCCT in themorningandafter lunchtimeandgraduallymodifyingbothlightingfeaturesafter9AMand1PM(700lxto500lx,CCT:4700Kto3000K).Participantsweremoresatisfiedwiththedynamiclightingconditionsthanwiththesteadylightingconditionsof500lxand3000K.However,theydidnotfindanysignificanteffectsof lighting conditions on vitality, alertness, headache, eyestrains, sleep quality or subjectiveperformance. Dynamic lighting conditions were also employed in another study for energysavingpurposes.AkashiandBoyce[80]foundthatreducingtheambientoffice illuminancebyaboutone‐thirddidnotcausealong‐termmodificationofoccupantsatisfactionwiththelightingenvironment.However,suchareductionintheambient illuminancecanstillproducenegativeshort‐term reactions, which will fade as time passes due to the increasing perception ofbrightness[80].

2.2.3 Brightnessperception

Luminousintensity

Brightness is a subjective assessment indicating ‘howmuch light’ is perceived by the humanvisual system; photometric measurements can lead to an objective assessment of thecorrespondingphysicalvariables[81].Brightnessperceptionisafunctionofluminance;hence,it inevitably relates to the light intensity and the reflectance of the targets’ surfaces.Understandingthefactorsresponsibleforabrightnessjudgmentcancontributetothedesignofmoreeffectiveandefficientspacelighting.AsstatedbyFlynnetal.,“…correctdesignattentiontoseveral ‘non‐quantitative’ lighting factors (might)compensate insomedegree forreduction in

Problemstatement

37

overallquantityoflight”...([82]referencedin[83]).Moreover,brightnessperceptionalsohasanimportantimpactonthefeelingofsafetywithinaspace.


Brightnessdoesnotdependonlyonilluminance;itisalsoafunctionofthecolourappearanceoflight.IthasbeendebatedformorethanadecadewhetherSpectralPowerDistribution(SPD)oflightsourceshasan impactonperceivedbrightness.SomestudiesdemonstratedthatSPDdidinfluencebrightnessofroominteriorsorspatialbrightness[80,84‐86],whileitwasnotthecasefor other studies [65, 87]. Prior investigations proved that brightness perception was notdependingonphotopicsensitivity.Variousindicatorslinkingthecolourappearanceoflightwithbrightness were suggested, such as CCT [80, 88], the colour gamut of light sources [86] orscotopic/photopicratios[84].ItwasreportedthatCCT,CRI,gamutareaorchromaticityalone,canbegoodpredictorsofbrightness.Morerecently,amesopicluminousefficiencyfunctionwasintroducedtoquantifythevisualperceptionunderlowilluminance.Thiswasdonebybridgingthe scotopic and photopic luminous efficiency functions into a unified system of photometry[89]. Rea et al. suggested amodel of brightness judgments based upon the shortwavelengthsensitive cone (S‐λ) fundamental [90]. The colour of objectswas also found to be relevant tobrightnessperception;however, itmightnotbeinfluencedsimplybytheirownluminancebutbytheircolourcontrastsagainstthebackground[91].Fromthisstudywasconcludedthatanewmetricforspatialbrightnessisstillrequired.


Luminancedistribution ina roomaffects thebrightnessperceptionof anobserver.One studyused two types of office rooms with relatively uniform luminance distribution to assessperceivedbrightness.Theofficeroomswereequippedwithrecessedfluorescentlightingwith1)acrylic lenses and 2) parabolic louvers. The offices with uniform lighting conditions wereperceivedatthesamebrightnessastheofficeroomswithnon‐uniformluminancedistribution,although the latter provided 5‐10% less task illuminance. The results showed that the officeroomswith non‐uniform lighting conditionswere perceived as brighter than the roomswithuniformlighting.Thenon‐uniformroomswere,inthiscase,characterizedbyalargerluminancecontrastbetweenthebrighteranddarkerwallportionsleadingtoabrighterappearance.

2.2.4 Visualperformance

Lighting conditions can impact human performance by three pathways: i) through the visualsystem, ii) through circadian effects, such as alertness (see Section 4), and iii) through theperceptual system, such as mood and motivation [1, 76]. Light is the key factor for visualperformanceviathevisualsystem,whiletherearemorefactorsconcernedwiththeothertwopaths (the circadian and perception systems). Visual performance is affected by illuminance,luminancedistributionandthespectralpowerdistributionoflight.

Problemstatement

38

Luminousintensity

Therangeoflightfluxreachingtheeyesduringadayislarge.Infact,sunlightgeneratesaone‐million‐timeshigherluminancethanmoonlight,i.e.100’000lxvs.0.1‐1lx.Naturally,thehumaneye has the flexibility to adapt to these various illuminance levels. Figure 2.5 illustratesschematically the scotopic,mesopic and photopic regions defined by Hurden et al. PhotopicvisionislinkedtotheV(λ)function(withapeakat555nm)reproducingthevisualsensitivityofbright light conditions that stimulate the cone receptors. At very low illuminance, the eye’sspectralsensitivityisdescribedbytheV’(λ)function(withapeakat505nm)determinedbytherodreceptors in thescotopic region.Themesopic regioncoversa luminousrange in‐betweenthephotopicandscotopicregions.Inthisregion,thehumaneyeisabletoperformregulartasks,wherebothrodsandconesareactive.Morerecently,anewmesopicphotometry(CIEmesopicsystem) has been introduced. It describes the spectral luminous sensitivity Vmes(λ) in themesopic region as the linear combinationof theV(λ) andV’(λ) luminous sensitivity functionsbasedontheratioofscotopicandphotopicvision(S/P)[73].Theborderlinebetweenmesopicandphotopicvisionisdifficulttodefine,manyfactors,suchassizeandpositionintheviewfield,have also an influence on the latter [78]. Figure 2.5 illustrates the fact that certain tasks canrestrictedlybeperformedinthephotopicregionduetocolourdiscriminationandvisibility.

Figure2.5:Schematicviewofscotopic,mesopicandphotopicregionsmodelledbyHurdenetal.[92]


Rea[93] investigatedvisualperformanceunderaconstant illuminance(278 lx)withdifferentluminancecontrasts.Hefoundthattaskcontrasthadasignificantimpactontaskperformance,asrelated tospeedandaccuracy [93].Lateron,healsouseddifferent illuminancerangesandfound that visual performance increased at higher illuminance [94]. Finally, a Relative VisualPerformance(RVP)modelwasdevelopedinordertopredictthevisualperformance,associatedwith task illuminance, task size and luminance contrast [95, 96]. The model was tested andvalidated[97,98];however,itcanonlybeappliedtotaskswithimagesonthefoveaandimplies

Problemstatement

39

thattheobserversknowwheretolookat,suchasinthecaseofreading.Themodelcanbeusedtoassessvisualperformancewithoutveilingreflections,disabilityglareandimpactofthelightspectrum.


Colour might influence visual performance during a chromatic task, but there is still littleevidence to support whether colour appearance of light significantly affects achromatic taskperformance.Besides,manystudies failed to identifyanyeffectof the light spectrum [85,99‐101].

However, Berman et al. demonstrated that the light spectrum influences visual performanceunder low luminance contrasts for achromatic tasks [102]. They studied the influence ofdifferent light spectra on the accuracy of task performance by using Landolt ring tests [102,103].SubjectshadtoassigntheorientationoftheopeningofaC‐shapedringinoneofthefourcardinal directions, under differing background luminance. It emerged that young and oldersubjects had a better task performance under green‐blue light sources (scotopically enrichedilluminance;scotopic luminance=228cd/m2), thanunderred‐pink lightsources(scotopicallydeficient illuminance; scotopic luminance = 13 cd/m2) at the same photopic luminance (53cd/m2; Figure 2) [102, 103]. Given this phenomenon, increased scotopic luminance with anassociatedsmallerpupilsize,can leadto improvedvisualacuity [104].Thesameauthorsalsoshowedthatpupilsize,brightnessperceptionandvisualperformancearelinkedtotheratioofscotopicandphotopiclumen(S/P)[84,104,105].Berman’stheoryimpliesthatalamptypewithalargerproportionofshort‐wavelength(higherscotopicluminance)wouldleadtobettervisualperformance. This was contradicted later on by other studies carried out under photopicconditions. Veitch andMcColl [99] found no correlation between visual performance and thetypeoflightsource(studyingfluorescentlampswithdifferentlightspectrums)for200luxtaskilluminance. Likewise, Boyce et al. found no impact of two different light sources (3000K vs.6500K;344luxand500lux)onvisualperformance,althoughthespectralpowerdistributionoflight influenced the pupil size differently [85]. It is still questioned if the light spectrumoperatingthroughpupilsizecouldhaveanimpactonvisualperformance.Pupilsizemightbeasmall determinant of visual performance,which canbe ignored formost lighting applications[106].Onlyalittleimpactonperformanceofsupra‐thresholdtaskswasidentifiedforpracticallightingconditions(photopicregion)[21].

Nevertheless, after the discovery of the novel class of photoreceptors, the intrinsicallyphotosensitive Retinal Ganglion Cells (ipRGC), it has been demonstrated that the ipRGCs areresponsibleforsomedynamicsofthepupilreflex[12,107].Itwasfoundthatashortwavelengthluminousradiationat460nm(‘blue’light)reducedthepupilsize[12,107]andledtoadelayedpost‐illuminationpupilresponse[108].


There isnoempiricalevidencethatdaylight,perse, issuperiortootherkindsof lightsourcesregarding thework performance via the visual system. The great advantage of daylight is itsquantity,thecontinuousspectrumandtheluminousdistribution.Somebenefitsofdaylightfor

Problemstatement

40

visualperformancewereidentifiedwhenthetaskinvolvedfinecolourdiscrimination;thiswashardlyfoundincaseofanachromatictask[109].

Itappears that the findingsofdifferingvisualperformancestudiesbetweenofficeroomswithwindows and windowless offices have so far been contradictory. According to a surveyperformedinanofficebuildinginSeattle(USA)[71],subjectiveworkperformancesweresimilarforofficeoccupantsworkingwithorwithoutwindows.The results showthatmostoccupantspreferredworkingunderdaylighting conditions [71].However, Figueiro et al. [110] observedproductivity of workers in a software development company in New York (USA). It washypothesisedthattheoccupantsinthewindowlessofficemightnotreceiveenoughlightfluxtoentraintheircircadiansystem(seedetailsinSection4.1),seekingoutforexposuretodaylightand consequently spending less time in theiroffices (or seeking formore social interactions).Theresultsshowedthatworkersinwindowlessofficesspent lesstimeoncomputertasksandmore time talking on the phone than office workers in glazed offices [110]. The authorsconcludedthattheoccupantsspentlesstimeontasksbyspendingmoretimetalkingtotheirco‐workers; thismight indirectly indicate that the officeworkers inwindowless officewere lessproductive than those benefitting fromwindows [110]. Other studies to be conducted in thelaboratory and field studies to test the reliability and consistency of these results are stillrequiredtoserveasafoundationforarchitecturalpractice.

2.3 Relevantquestionsandhypotheses

Visual comfort at the workspace has been investigated in the past with respect to glare,luminancedistribution,luminousintensityandinfluencesonvisualperformance.Beyondthesevisual effects, the lighting environment also impacts humanphysiology, behaviour andmood,mainly via a recently discovered novel class of photoreceptors inmammals, the ipRGC. Sincethen, it has become evident for indoor lighting to consider not only the quantity (e.gilluminance) but also the quality of light (e.g. the spectral composition). Nevertheless, theimpactsofa lightingenvironmentonvisualcomfortandtheotheraspectsneedstobe furtherinvestigatedintermsofbothdaylightingandartificiallightingconditions.

Thisdoctoralthesisaimedtoeffectivelyassesstheimpactof lightonvisualcomfort, includingbothvisualandnon‐imageformingeffects.Thetwomainrelevantquestionsforthenon‐imageformingeffectswere:

i. Howcanthedistributionofenvironmental light, its intensityandspectralcompositionbeassessedmoreeffectivelyinsideofbuildings?

ii. What is its impact on human visual comfort including visual andNon‐Image Forming(NIF)effects?

Thefollowinghypotheseswerederived:

Problemstatement

41

An appropriate device for monitoring luminance distribution must be efficient andaccurate in order to investigate the luminance distribution simultaneously underrealisticofficelightingconditions,especiallyunderdynamicdaylightingconditions.

Forthepurposeofvisualcomfortassessmentsincludingnon‐imageformingeffects,thedevicemust be able tomonitor those distributions according to both visual and non‐image formingeffectsoflight.

The firstquestionwasansweredbymeansofanovelphotometricdeviceusedtomonitor theluminancedistributionwithinofficerooms.ItwillbepresentedinChapter3.

Toanswer the secondquestion, studieswithhumansubjectswereperformedunderdifferentlightingconditionsinvolvingbothdaylightingandartificial lighting.Duringtheseexperimentalstudies, the view throughwindowswas excluded in order to limit the impact of the outdoorenvironment.Daylight,perse, aswell as itsphotometricpropertieswas examinedwithvisualcomfort,includingtheNIFeffects.TheresultsarepresentedinChapter5and6.Thehypothesiswithregard to thestudiesof impactof lightonvisualcomfort, includingbothvisualandnon‐visualsystems,wasthefollowing:

Inter‐individual differencesbetween subjects are expected, for example fromdifferentchronotypesorgender.Likewise,thechangeovertimeisalsoassumedtohaveaneffecton the results of the studies due to the varying endogenous biological rhythmsof thesubjects.

Inordertosupportthemainhypothesesofthisthesis,aconceptualmodelofvisualcomfortandlightimpactswassetupandillustratedinFigure2.6(followingFigure1.3byVeitchetal.[18],presentedinChapter1).Lighthasanimpactonnon‐imageformingeffects,whichincludesacutelighteffects,circadianrhythmsorthephysicalandbehaviouralchangesduringa24‐hourcycle,whichareexplainedinChapter4.Itisknownthatlightsynchronisestheinternalbodyclocktothe external 24‐hour light‐dark cycle on earth [111]. Hence, apart from visual comfort andperception,lightalsoimpactsourphysiologyviathecircadiansystemwhichisalsoincludedinthis new conceptualmodel of visual comfort. Figure 2.6 shows the light paths aswell as theimpactsof eachphysicalpropertyof light, suchas luminousdistribution, intensityandcolourappearance.ThemainvariablesonwhichthisthesisisfocusedarepresentedinthewhiteboxesinFigure2.6.

Problemstatement

42

Figure 2.6: Conceptualmodel of visual and non‐visual systems impacted by the different physical lightproperties of light: impact of luminance distribution (red), luminous intensity (blue) and colourappearanceoflight(green),constitutingthemainhypothesesofthedoctoralthesis

Thedevelopedphotometricdevice(issued fromquestion i)wasusedduringtheexperimentalstudies in order to examine the relation of luminancedistributionwith visual comfort,whichmight also impact other aspects of human eyes beyond vision. The hypotheses regarding theimpactoflightonvisualcomfortincludingNIFeffectswereformulatedasfollows:

SincethelightingenvironmentalsoinfluencestheNIFeffects,theremightbesomelinksbetween thevisualsystemand thenon‐visualsystem,orvisualcomfortmightdirectlyrelate to both systems. It is believed that light is first perceived, and then a givenbrightness is judged later on. In addition, light also directly affects subjective lightingpreference and mood; light enhances well‐being through the perception path. Thecircadiansystemisaffectedbylightandalsocontributestotheimprovementofhealthandwell‐being.

Photometric variables are likely to greatly influence visual comfort and perception.Luminancedistribution(showninredinFig.2.6)isexpectedtoinfluencevisualcomfort,brightness,lightpreferenceandmood.Luminousintensity(blue)andcolourappearanceoflight(green)mightimpactallvariables.

LuminousdistributionsmightalsoinfluencetheNIFeffectsregardingvisualcomfort.

Problemstatement

43

2.4 Structureofthesis

Thisthesisreportisstructuredinthefollowingway:

Chapter 1 gives an overview of the impacts of office lighting conditions on occupants. Itintroducessomeopenquestionsregardinghealthylightingcriteria.

Chapter 2 reports the fundamentals of visual comfort. It presents the relevant questions andhypothesesofthethesis.

Chapter 3 describes luminancemapping techniques including the recent high dynamic rangeimaging technique. It also introduces the novel camera‐like luminance sensor, which wasdevelopedandusedintheframeworkofthisthesis.

Chapter 4 reports the fundamentals of non‐image forming effects. It also explains how lightimpactsthenon‐visualbiologicalfunctions.

Chapter5reportstheinvestigationoflighteffectsonvisualandnon‐visualfunctionsaccordingtotwoexperimentalstudies.

Chapter6presents the visual comfort and luminancedistribution assessedby thewayof thenovelcamera‐likeluminancesensor.

Chapter7discussestheimpactsoflightonvisualcomfortincludingnon‐imageformingeffects.Theconclusionintendstoprovideknowledgeandgivesomesuggestionsforfurtherresearch.

45

Chapter3 Photometricmeasurements

3.1 Physicalparameters

In order to understand the influence of light on human performance and behaviour, physicalpropertiesofthelightsourceshavetobedetermined,‐andmeasurementshavetobeperformedbyusingappropriatemetrics.

A light flux transports radiant energy and can be characterized using two different metricsreflectingitsradiometricandphotometricproperties.Radiometricunitsdefinethepropertiesoflight in terms of quantities of absolute emitted electromagnetic radiation (from the entirespectrum). Photometric units definemeasures in the visible range of light, i.e. from 380‐780nanometres(nm).Theyareweightedbythesensitivityofthehumanrodsandcones,sinceoureye is not equally sensitive to allwavelengths of electromagnetic radiation. Both radiometricandphotometricmetricsincludefourbasicvariables,illustratedinTable3.1.

Quantity FunctionRadiometry(Unit) Photometry(Unit)

Radiantpower(W) Luminousflux(lm) Totaloutputfluxfromalightsource(pertimeunit)

Radiantintensity(W/sr) Luminousintensity(lm/sr=cd) Fluxemittedbyalightsourceinadirectionperunitsolidangle

Irradiance(W/m2) Illuminance(lm/m2 =lx) Amountoflightfallingontoasurfaceatacertaindistancefromthelightsource

Radiance(W/m2sr) Luminance(cd/m2=nit) Amountoflightemittedbyanelementoftheareaofalightsourceinagivendirectionperunitsolidangle

Table3.1:Radiometricandphotometricvariables,assummarizedfromreference[112]

Photometry isusuallybasedonphotopic visiondescribedby thehumaneye sensitivity curvedesignated by the V(λ) function; it is used in photometry as an average action spectrum forphotopic responses following the CIE luminous efficiency function [113]. At very lowilluminance,theCIEscotopicluminousefficiencyV’(λ)functiondefinestheactionspectrumforthescotopicresponse[113].Figure3.1presentstheV(λ)andV’(λ)functionsacrossthevisibleelectromagnetic spectrumof light (from380 to 780nm).Moreover, in betweenphotopic and

Photometricmeasurements

46

scotopicregions,anewCIEmesopiccurvehasbeenrecentlyintroducedtodescribethespectralluminous efficiency for the mesopic region [89], as mentioned in Section 2.2.4. In order tocharacterize the response of the human eyes in the mesopic region, the ratio of luminousquantityaccordingtothescotopicandphotopicvisions(S/Pratio)isused.MesopicAdaptationLuminance(MALorLMES)canbedeterminedusingtheS/Pratiofordifferentluminancevaluesintherangeof0.01‐3cd/m2[89].LargeS/Pratiosindicateahigherluminousefficacyofthelightsourcewithrespecttothemesopicsensitivityinhumans.Sincethemesopicregionisdynamicand depends on previous light and dark adaptation [114], the S/P ratio determines the LMESvaluesfordifferentluminances[89].

Figure 3.1: Relative sensitivity functions: a) circadian sensitivity function C(λ)[115] (dashed line); b)scotopic luminous efficiency function V’(λ) [89] (dotted line); c) photopic luminous efficiency functionV(λ)[89](solidline).

Besides the scotopic and photopic luminous efficiency curves, a circadian sensitivity functionhas been recently introduced. It is based on the relative sensitivity of a new class ofphotoreceptors located in the retinal ganglion cells, the so‐called intrinsically photosensitiveRetinalGanglionCells(ipRGC)[116,117],showingamaximalsensitivityaround480nminthebluerangeofvisiblelight.TheipRGCsareknowntoconveymanysocalledNon‐ImageForming(NIF)functionssuchascircadianrhythms,pupillightreflexes,andhormonalsecretion[11‐15].Priorresearchshowedlargerinducednocturnalmelatoninsuppressioninhumanswithnarrow‐bandwidth blue light (446‐477 nm) [118, 119], indicating that melatonin suppression is anindirect sensitivity marker of ipRGCs. A few research groups suggested their own circadiansensitivity curves [115, 120, 121] based on melatonin suppression data, first published byBrainard et al. [118] and by Thapan et al.[119]. A relative circadian sensitivity function C(λ),illustratedinFigure3.1,wasproposedbyGall[115],whoalsosuggestedthreenewmetricsinordertoquantifythecircadianeffectsof lightsources: i) thecircadianweighted irradianceEec(unit in W/m2); ii) the circadian action factor acv (no unit) and iii) the circadian weightedradiance Lec (unit in W/m2sr) [122, 123]. Another research group, led by M. Rea (LightingResearchCenter,RensselaerPolytechnicInstitute,NY,USA),alsosuggestedaC(λ)curvebasedonhumanmelatoninsuppressiondata[120,124].Accordingtotheirmodel,theterms‘CircadianLight’ (CLA) and ‘Circadian Stimulus’ (CS) were introduced as new circadian metrics [125].


47

Briefly,theCLAisalsoaspectrallyweightedirradiance(unitinW/m2)andtheCStransformstheCLAintoarelativeunitfrom0to1, inordertodescribetheeffectivephoticstimulusofalightsourceforthecircadiansystem[125].AthirdapproachforacircadiansensitivityfunctioncamefromKozakovetal.[126].Accordingtotheseauthors,themaximalcircadiansensitivityiscloseto450nm.Moreover,the“melanopicspectralefficiencyfunction”orVz(λ)(thenamemelanopicrefers to melanopsin, the ipRGC photopigment) with a peak sensitivity around 480 nm wasrecentlyintroducedbytheLucasgroup(UniversityofManchester,UK).Basedonthisfunction,a“melanopicilluminance”(in‘m‐lux’)canbederivedinordertomeasuretheluminousintensityfortheNIFeffectsinanimalsandhumans[127,128].Besidescircadianefficiencyfunctions,thecorrelation factors of radiometric/photometric variables with circadian metrics were alsorecentlysuggested.BelliaandBisegna[121]proposedtwoconstantvalues,alpha(α)andbeta(β), for different light sources as conversion factors for radiometric (α) and photometric (β)functions. Their suggested β value depends on the spectral power distribution of the lightsource. This approach has been proposed to lighting designers and practitioners in order toevaluatethecircadianeffectoflightsourceswith‘easytoapply’parameters.However,todate,none of the above described circadian sensitivity functions have been incorporated intointernational lighting regulations; there is currentlywork inprogressby several internationalinstitutions[129].

Eventhoughthereisnotyetaformalcircadianmetricavailabletoquantifythebiologicaleffectsof light in humans, there are already devices for measuring light fluxes in circadian metricsbased on existing sensitivity curves. Besides spectroradiometers with a circadian metricfunction,a fewnoveldeviceshavebeenintroducedas“personalexposuredevices” inordertoassess long‐term light exposure for both visual and circadian spectral sensitivity. The“Daysimeter”(developedbytheLRC,RensselaerPolytechnicInstitute,NY,USA)[130]containstwophotosensors:aphotopicdetectorandashortwave(blue)lightsensor;thelatterisderivedfrom the C(λ) curve suggested by Rea et al. [124], asmentioned above. The device convertsincoming light fluxes into illuminance, CLA and CS. The “Daysimeter” comprises anaccelerometer which monitors activity as well as an optical sensor array wired to a circuitboard;itisavailableintwoversions[131]:i)the“Daysimeter‐S”,asmallhead‐mounteddevice,andii)the“Daysimeter‐D”,withamicrocliptobepinnedonthecollarofashirt.Bothdeviceswerevalidatedsimultaneouslytorecordlightexposureandactivityrhythmsfromday‐shiftandrotating‐shiftnurses[132].

Theseconddevice,called“LuxBlick”,wasdevelopedbyHubaleketalatETHZurich,Switzerland[133].The“LuxBlick”includestwolightsensorstorecordilluminanceandbluelightseparatelyaswell as acvmentionedabove.Blue lightandacv are circadianmetricsderived from theC(λ)curvebyGall[122].Thedevicecanbefixedonspectacleframesinordertomonitordailylightinflux during an observation period. “LuxBlick” was used as a tool to investigate daily lightexposures and their relationship with subjective sleep quality and mood in office workersduring seven consecutive days. The results revealed the positive impact of light, includingluminousexposure,durationandspectrumparameters,duringdaytimeonsleepqualityonthefollowing night [134]. The measurements taken with office workers pointed to the fact thatexposuretolightonworkingdayswasratherconstant,whereasitvariedonthedaysoff[134].

The last device, a wrist worn activity watch “Actiwatch Spectrum”, is commercially availablefromPhilipsHealthcare (Eindhoven, NL) [135]. It can simultaneously record illuminance and


48

spectral irradiance for red, green and blue (RGB) components via separate spectral channels.Theequipmentwasusedtoinvestigatedailylightexposuresinhumansacrossseasonsaswellas the relative red, green, and blue wavelengths content of these seasonal changes inphotoperiods[136].TheaccuracyforthephotometricmonitoringandRGBsensorsystemofthedevicewere recently assessed [137].The same study suggested touse a given fractionof theblueandgreenchannelsoutputsforthemonitoringofthe“bluelight”circadianactionspectrumtobeusedasacircadianpersonalexposuredevice[137].Anotherstudycomparedtheuseofthe“Actiwatch Spectrum” for circadian studies with the “Daysimeter” [131]. It resulted that thephotometricperformanceofthe“Daysimeter”wasbetterthanthatofthe“ActiwatchSpectrum”.TheCIEstandarderrorof theV() function (designatedby f1’) [138]of the “Daysimeter”wasequal to 25%; the corresponding value of the “Actiwatch Spectrum” showed a substantialdifference with a f1’ value of 83%. The authors argued that systematic errors in photopicsensitivityofthespectralsensorsexisted.Theyassumedthat“ActiwatchSpectrum”potentiallyoverestimates the light flux and thus led to measurement errors, especially for the bluespectrum,whichisessentialtoassessthelightimpactonthehumancircadiansystem[131].

Besides absolute and photometric measurements, the perception of colours in workenvironments also influences lighting quality [79, 87, 139]. Themost common techniques toassess the colour appearance are the Correlated Colour Temperature (CCT) and the ColourRendering Index (CRI), used in various lighting applications to compare the colorimetricpropertiesofdifferentlightsources.TheCCTisaspecificationofthecolourappearanceofthelight fluxemittedbya lamp,relating itscolourtothecolourofablackbodyheatedtoagiventemperature,measuredinKelvin(K)[22].AhigherCCT(over5000K)relatesto‘coldercolours’andcontainsalargerfractionofshortvisiblewavelengths(i.e.bluecomponents),whilealowerCCT(2700‐3000K) is related to ‘warmercolours’ (i.e. those that containmoreredandyellowwavelengths)[22].TheCRIisaquantitativeassessmentoftheabilityofagivenlightsourcetorender colours similar to their colour appearance under a reference light source (i.e. usuallydaylightoranincandescentlamp).HigherCRIvaluesindicateabettercolourrenderingforthelightsource,whicharecloserto“true”colours.

Lastly,thecharacterizationoftheluminancedistributionintheviewfieldisalsoimportant,asitisakeyforvisualdiscomfortassessment.Previously,luminancedistributionassessmentswereperformedusingaluminancemeter,bymonitoringpointtopointtheluminanceofsurfaces.Asimilarprocedure is used for glare risk assessments: afterdeterminingpoint topoint surfaceluminances,glareindicescanbecalculatedusingagivenformulae(moredetailsinSection3.2).Recently,advancedmethodsforassessmentof luminancedistributionusingtheHighDynamicRange(HDR)imagingtechniquewereemployed(seeSection3.3).AnHDRimageisacquiredbyusingseveralLowDynamicRange(LDR)images,takenwithdifferentapertureexposureswithaCharge‐Coupled Device (CCD) camera. Using this technique, a ‘luminance map’ of lightingenvironmentscanbeobtained[140‐143].ThebenefitofthistechniqueistheabilityofanyCCDcamera to capture detailed spatial information relating to luminance distribution in a givenspace.AnotherapplicationofHDRimagingtechniques isdeterminationofglareriskswiththehelpofavailablesoftware,asdescribedindetailinSection3.3.3.

In the course of this thesis, an innovative device has been introduced in order to assessluminance distributionmore rapidly andmore efficiently. By using this device, it is nomorenecessary to capture several images: a single HDR image can be captured by the way of a


49

snapshot.Thisisusefultoassessdynamic(day‐)lightdistributionsaswellasthepropertiesofanylightsourceincircadianmetricsaccordingtoGall’scircadiansensitivityfunction[122].Thecalibrationandthevalidationof thisnovel luminancemeteraredescribed inSection3.4. TheapplicationwithcircadianmetricsispresentedinSection3.5.

3.2 Glarerisksassessment

Asmentionedbefore (seeChapter 2, Section2.2.1), discomfortglare is anunwanted effect ofexcessiveluminancecontrastintheviewfield.Nowadays,noneofthedaylightglareratingsareunanimouslyrecognizedasaninternationalstandard.ThemostcitedoneistheDaylightGlareIndex (DGI)which is amodified version of the glare index recommended by the IlluminatingEngineeringSociety(IES):theIESglareIndex(IESGI),suggestedbyHopkinsonforlargeglaringsources [144]. The DGI glare formula, expressed by Equation 3.1, was created for large arealightingsources,suchaswindows;itsusewasrecommendedfordaylightingconditions[45].

(3.1)

where:

n[‐]isthenumberofglaringsources;Lsi [cd/m2] istheglaringsources luminance(i),suchastheluminanceofthevisibleskypatch,theobstructionsandthegroundperceivedthroughthewindows;Ωi[sr]isthesolidangularsubtendedbytheglaringsources(i);Lb[cd/m2]isthebackgroundluminance,suchasthoseoftheroom;Lw[cd/m2]istheaveragewindowluminanceweightedaccordingtotherelativeareasoftheskyandtheground;ωsi [sr] is the solid angle of the luminous parts of each light source (i) as viewed from theobserver’seye.

SeveralauthorscriticizedthereliabilityofDGI,andotherformulasfordaylightdiscomfortglareassessments were suggested [145, 146]. Osterhaus et al stated that the direct verticalilluminanceat theobserver’seye(vertical illuminanceat theeye)showed largestcorrelationswiththesubjectiveglareratingexpressedbyobservers,comparedtootherglareindices,suchastheCIEGlareIndex(CGI),theUnifiedGlareRating(UGR)andtheDGI[147].Recently,anotherglare index, the Daylighting Glare Probability (DGP) has been suggested by Wienold andChristoffersen [148], based on a glare risks assessment within an office room showingdiscomfort glare inducedbywindows.TheDGP combined the vertical eye illuminancewith amodifiedglareindexgivenbyEquation3.2.Itwasusedinrecentlightingstudiesforevaluationsofdiscomfortglareinducedbywindows[149‐152].

wsib

isi

i LL

LDGI

5.0

8.06.1

110

07.048.0log10


50

(3.2)

where: Ev[lx] istheverticalilluminanceattheeye; 5.87 10 ;Lsi[lx] istheluminanceofsource(i); 9.18 10 ;ωsi[sr]isthesolidanglesustendedbysource(i); 0.16;Pi[‐]istheGuthpositionindexofsource(i); 1.87.

Anotherglareindex,theUnifiedGlareRating(UGR)recommendedbyCIE,isintendedtobeusedforartificiallightingconditionsandgivenbyEquation3.3[153],

(3.3)

where:

Lb[cd/m2]isthebackgroundluminance;Li[cd/m2]istheluminanceofthebrightestpartofeachlightsource(i) inthedirectionoftheobserver’seye;ωi[sr]isthesolidanglesustainedbythebrightestpartofeachlightsource(i)attheobserver’seye;Pi[‐]istheGuthpositionindexofeachlightsource(displacementfromthelineofsight).

TheUGRformulaiswidelyusedforvisualcomfortanalysisofofficeroomswithartificiallightingconditions.AccordingtoCIErecommendation,UGRvaluesmustnotovercomeanupperlimitof19 in office rooms equippedwithwork stations and Visual Display Terminal (VDT) (writing,typing, reading,dataprocessing) [153].Table3.2givesacomparisonof the threeglare ratingscales,togetherwiththeassociatedglaresensations.

GlareSensation DGP DGI UGR

Imperceptible <0.30 <18 <13

Perceptible 0.30‐0.35 18‐24 13‐22

Disturbing 0.35‐0.45 24‐31 22‐28

Intolerable >0.45 >28 >28

Table3.2:CriteriaofglaresensationsassessedbytheDGP,DGIandtheUGRscales(after[149])

Glareindices,suchastheDGI,DGPandUGRcanbeassessedusingluminancemappingsbasedon the HDR technique employed in photography (see Section 3.3) or HDR image renderingsgeneratedbycomputersimulations.Thesoftware‘RADIANCE’[154]providesspecificcodesfor

31

24

,2

,1021 1log c

PE

LcEcDGP

i icv

isisv

12

225.0log8

i i

ii

b P

L

LUGR


51

thatpurposesuchas: ‘Findglare’, ‘Glarendx’ [10]and ‘Evalglare’ [148].These functionscanbeused forHDRdigital image renderings. Thus, bymeans ofHDR imaging techniques, potentialglare sources are first identifiedon the images, and in a second step the corresponding glareindexcalculationscanbeexecuted.TheglareassessmentbasedonHDRtechniquesisdescribedfurtherinthenextsection.

3.3 HighDynamicRange(HDR)imagingtechniques

3.3.1 Background

Thehumanvisualsystemcancopewithlightintensitieswhichvaryover10ordersofmagnitude[155]. It is capable of simultaneously adapting to contrasts over a range of five orders ofmagnitude within a scene [156]. An ordinary digital 8‐bit image is characterized by a muchlower luminousdynamic range.ALowDynamicRange (LDR) image cannotprovide sufficientdetailsoftherealscene,especiallyinshadedoroverexposedregions.Thus,thepixelvaluesofanLDR image do not correspond to the luminance values of the real scene. Conversely, HDRimagingtechniquescancapturethefulldynamicrangeofrealscenes.AsingleHDRimagecanbegeneratedby fusing a sequenceof LDR images underdifferent exposure intervals, takenby aCharge‐Coupled Device (CCD) camera. As a result, each pixel data can bemerged by using aresponsecurvederivedfromthissetofimages.Analgorithmthatrecoverstheresponsecurveofthe CCD camera allows obtaining an accurate radiance value for each pixel [157]. Luminancemappings can be produced using any available open source software, such as hdrgen [158],RADIANCE [159], Photosphere [160], or by accessing theWebHDR HTML [161] (which useshdrgenasaback‐end).

3.3.2 CalibrationofHDRImages

In order to empirically calibrate an HDR device, luminance values assessed by HDR imagingtechniquesmustbecomparedwiththosemonitoredpoint‐to‐pointbyaluminancemeter;thisisgenerallydonetodeterminetheaccuracyandreliabilityofthistechnique[140‐142,151].Figure3.2showstheexperimentalsetupusedforthatpurpose,basedonaCCDcamera(NikonCoolpix5400, Tokyo, Japan) which is equipped with an FC‐E9 fisheye lens (Nikon, Tokyo, Japan), acalibrated luminance meter (Minolta LS110, Tokyo, Japan), tripods and targets (Macbeth®colour chart, X‐Rite, Michigan, USA), as previously reported [151]. Relative errors can bedetermined by comparing the device output with a conventional point‐to‐point luminancemeter. Most studies showed that this technique provided very reasonable 10‐20% relativeaccuracyforawiderangeofluminance(0.5–12,870cd/m2)[140,142].


52

Figure 3.2: Experimental set‐up comprising a Nikon Coolpix 5400 CCD camera equippedwith a FCE‐9fisheye lens, a calibratedMinolta LS 110 luminancemeter and tripods. The luminance calibrationwasmadeatdefinedtargetpointswithintheroomusingaMacbethColourChart[151].

3.3.3 Practicalapplicationsofluminancemaps

Luminanceratiocalculation

Theluminanceofagivensurfacecanbeextractedusingoneoftheabovementionedpiecesofsoftware (see Section 3.3.1). In the next step, the luminance of different task areas andsurroundingscanbeobtainedandluminanceratioscalculated.Toassessthetaskluminanceorthesurroundingsofaspecific locationinamoreprecisemanner,certainareascanbemaskedusingimageprocessingsoftware,suchasAdobePhotoshop®oranopensourceasGIMP.Using“pcomb”and“pvalue”commandsinRADIANCE[154,159],eachmaskedareacanbeprocessedin order to determine its average luminance value [151]. Figure 3.3 illustrates the masks ofdifferentareasasobtainedbytheimage‐processingsoftwareAdobePhotoshop®[151].

Figure3.3:Markedareasof thereference taskareasand their surroundings inanoffice roombyusingmaskingfunctionsfortheassessmentofluminanceratios[151].


53

Glareindicescalculation

Glare indices in the occupants’ field of view can be calculated throughHDR images using the‘Evalglare’inRADIANCEroutines[148,159];moreover,‘findglare’inthesoftwareRADIANCEisalso applicable [154]. Several parameters of the glare formula, such as the luminancebackground,glaresourceluminanceandsolidanglescanalsobedeterminedusingHDRimagingtechnique; luminance values can be extracted for each pixel from the luminance mapping.Potential glare sourcesare identifiedbybeingdetectedon thebasisof a luminance thresholdvalue[160,162].Thisway,DGI,UGR,DGPaswellasotherglareindicescanbecalculated.

Figure3.4showspixelsofpotentialglaresources,ascircledby theRADIANCEsoftwareusingthe‘findglare’and‘glarendx’commands.Applying‘Evalglare’,irregularshapesofpotentialglaresourcescanalsobedetectedandcolouredasshowninFigure3.4b).Moreover,thetaskareacanalsobemanuallydefined;contrastsbetweentaskareasandpotentialglaresourcescanthenbecalculatedmoreprecisely. Inthecourseof thethesis, theglare indicesweredeterminedusing‘Evalglare’,whichwillbepresentedinSection3.4.3andinChapter6.

Figure3.4:Glareriskassessmentusing:a)RADIANCEwith‘findglare’and‘glarendx’commands[159]andb)‘Evalglare’[148,162]

3.3.4 LimitationsofHDRimagingtechniques

GlareriskassessmentsandluminancemappingcanbeaccuratelyperformedbythewayofHDRimagingtechniques.However,anon‐negligibletimelapisrequiredforimagecapturingandpostprocessing. Inorder toachieve luminancemaps, anumberofLDRphotographsareneeded inordertogeneratethefulldynamicrangeofluminanceinthescene:theLDRimagesmustcovertherangeofminimalandmaximalluminancevaluesoftheconsideredscene.Alargenumberofpictures leads toabettersignal‐to‐noiseratio.Theprocedure is timeconsumingsinceseveralimageshavetobecollectedbythewayofaconventionalCCDcameratoproduceasingleHDRpicture. Another disadvantage is the difficulty to copewith dynamic daylight conditions: theluminancedistributionofadaylitscenemightchangewhile takingthose images.Dealingwithhighlydynamicandcontrasteddaylightingenvironmentsalwaysresult inrapidchangesofthedifferent glare indices due to the variable luminance of the sun and the sky vault. It wouldtherefore be better to capture images more rapidly, especially under dynamic daylight


54

conditions.Toachievethisgoal,anewlydevelopedCamera‐LikeLightSensor(CLLS)wasusedin the course of this thesis (see Section 3.4). It aims to produce faster HDR images and tosimultaneouslyassessglarerisksundervarying(day‐)lightingconditions.

3.4 TheCamera‐LikeLightSensor(CLLS)

3.4.1 Principlesofthedevice

TheCamera‐LikeLightSensor(CLLS)wasdevelopedbytheCentreSuissed’ElectroniqueetdeMicrotechnique (CSEM; Neuchâtel, Switzerland). The original model, “IcyCAM” (Figure 3.5a),was introduced for the purpose of security surveillance, optical character recognition andindustrialcontrol[163].Theessentialelementsofthisdevicearetheintra‐scenedynamicrangeof the optical front‐end and a data representation that is independent from the illuminationlevel.Itcontainsa‘system‐on‐chip’(SoC)combininganopticalfront‐endandaprocessoronthesamedevice.TheSoCenablesasinglechiptoperformimageacquisition,analysisanddecision‐making [164]. Each pixel achieves a 132 dB intra‐scene dynamic range due to a logarithmiccompression [163]. The logarithmic encoding allows one single HDR image to be real‐timerecordedin780sto2sexposuredurations[165].Accordingto its intra‐scenedynamicrangefunction,theCLLSprovidesarecordingofimagesinrealtime.HDRimageswithmuchgreaterspeed performance than the ‘Classical HDR’ technique of a conventional CCD camera can berecordedaccordingly.TheCLLScancapturean image (ora video)andpresent it in: i) classic(luminance)mode(Figure3.6a),ii)contrastmode(Figure3.6b)andiii)directionmode(Figure3.6c). The contrastmode represents the contrastmagnitude of an image; the directionmodeshowstheorientationcontrastofeightdirections.

TheaiminthisthesiswastousetheCLLSdevicefordynamicluminancemappingandglarerisksassessments.TheCLLSwasfirstcalibratedforthatpurposeforphotometricmeasurements;thecorresponding calibration procedure according to the photopic spectral luminous efficiencyfunctionV(λ)isdescribedinSection3.4.2.Inordertoextendthefieldofviewtoapproximately154°X115°whichisclosetothehumanvisualfield(~120°x120°forthepanorama[26,155]),theCLLSwasequippedwithaFujinonfish‐eyelensFE185C057HA[166](asshowninFigure3.5b);thevignettingeffectswerealsocorrected(seeSection3.4.3).

The followingsectiondescribes thecalibrationand thevalidationof theCLLS forphotometricpurposes. The results were presented at the SPIE 2012 conference and reported in the SPIEconferenceproceedings[167].


55

a) b)

Figure3.5:a)TheoriginalIcyCAMdevelopedbyCSEM;b)theIcyCAMequippedwiththefisheyelens

a) b) c)

Figure3.6: ImagetakenbyCLLSwiththreedifferentviewoptions;a) thedefault imagewiththeoption“luminance”,b)imagewiththeoption“contrast”andc)imagewiththeoption“orientation”.

3.4.2 Calibrationprocedure

SpectralSensitivitycalibration

In order to match the spectral sensitivity of the CLLS to the human eye (according to thephotopic sensitivity function), theCLLSspectral responsewasdeterminedand thencorrectedby three customized filters at the optimal thickness. The calibrationwas set up as shown inFigure 3.7, based on measurements performed previously in our laboratory [168]. Narrowbandwidthmonochromatic lightbeams (generatedbya1000WattXenon lampanda gratingmonochromator) were used as a reference light source. The CLLS and a calibratedspectroradiometer(Specbos1201,JETI,Jena,Germany)monitoredtheemittedradianceonthesametarget inan integratingsphere.Thereference lightsource(monochromatic lightsource)wassetata555nmwavelength(usingthemonochromator),correspondingtothemaximumofthephotopicsensitivitycurve,fordifferentluminousintensities.


56

Figure 3.7: Experimental set‐up for the spectral sensitivity calibration. The luminance meter, thespectroradiometerandtheCLLSwereplaced in frontofanUlbrichtsphere; thesametargetwasaimedusing the “spot laser” of the spectroradiometer. The irradiances were measured using thespectroradiometer;simultaneously,imageswerecapturedbytheCLLS.

TheCCLSdigitalcameraprovidedonevalueperpixelonagrayscale(inarbitraryunitsfrom0to1024 digits), while the spectroradiometer measured the corresponding radiance (inμW/m2.sr)[168].Equation3.4wasusedtodescribetherelationbetweengrayscalevaluesandmonitoredradiances(R²=0.9994).

y 1.36 ∙ 10 ∙ 5.71 ∙ 10 ∙ 1.29 ∙ 10 ∙ 9.51 ∙ 10 ∙ 6.76 ∙ 10 ∙

1.83 ∙ 10

(3.4)

Inanext step, the luminous intensitywaskeptconstant; the samemeasureswere then takenwiththeCLLSandthespectroradiometer,thewavelengthofthelightbeambeingmodifiedusing5nmsteps from380 to780nm.Therelative rawspectral sensitivityof theCLLS is shown inFigure3.8a,togetherwiththestandardV()photopicluminosityfunctionsuggestedin1931bytheCIE[169].Threecolouredglasseswereselectedasfilters:Kopp3307,Kopp3384andKopp7071(KoppGlassInc.,PA,USA).TheiroptimalthicknesswasdeterminedinordertomatchthespectralresponseoftheCLLSwiththephotopicsensitivityfunctionV().Afterimplementationof the filters on the CLLS, the previous procedurewas repeated to obtain a sensitivity curvewhichcorrespondstotheV()(Figure3.8b). Inordertoassesstheerrorbetweentherelativespectral sensitivity of the CLLS and the V() curve, the CIE standard error f1’ [138] wasdeterminedusingEquation3.5:

0.93584 | | d

(3.5)


57

Thecorrespondingstandarderror isequal to8.3%.Basedontherequirements fromstandardorganizations (DIN 5032 [170]), the maximally tolerated error for commercial devicescorresponds to f1’ = 9% ([170], cited in [133]); the error obtained from CLLS, after spectralcalibrationlieswithinanacceptablerange.

Figure3.8:RelativespectralsensitivityassessmentsoftheCLLS[167];a)relativerawspectralsensitivityofCLLS(filledcircles)shownwiththeCIE1931photopicsensitivityfunction[169](V();opentriangles);b)correctedspectralsensitivityresponseoftheCLLSequippedwithfilters(filledcircles)shownwiththeCIE1931photopicsensitivityfunction[169](V();opentriangles),withinthevisiblerange.

Photometriccalibration

Toperformaphotometriccalibration,atotalof138measurements(from0.01cd/m2to15,560cd/m2)were carried‐out by using the CLLS simultaneouslywith a calibratedMinolta LS‐110luminance meter (Konica Minolta, Tokyo, Japan), as well as a calibrated spectroradiometer(Specbos1201,JETI,Jena,Germany).Polychromaticwhitelightfroma1000WXenonlampanda 1200 W metal halide spotlight were used as reference light source for the photometriccalibration.Theexperimentalset‐upisshowninFigure3.9.

Figure 3.9: Experimental set‐up for the photometric calibration. The luminance meter, the spectro ‐radiometer and the CLLS are aiming at the same target. The latter was located using the “spot laser”functionof thespectroradiometer.Luminancesweresimultaneouslymeasuredbythe luminancemeter,thespectroradiometerandbythewayofCLLSimages.

Wavelength (nm)

300 400 500 600 700 800

Rel

ativ

e Sp

ectr

al S

ensi

tivity

(-)

0.0

0.2

0.4

0.6

0.8

1.0

1.2Raw Spectral Sensitivity of IcyCAM without Filter CIE 1931Photopic Sensitivity Function (V-lambda)

Wavelength (nm)

300 400 500 600 700 800R

elat

ive

Spe

ctra

l Sen

sitiv

ity (

-)0.0

0.2

0.4

0.6

0.8

1.0

1.2 Corrected Sensitivity Response of IcyCAM CIE 1931Photopic Sensitivity Function (V-lambda)


58

Thereferenceluminancevalueswereobtainedbythecalibratedluminancemeter.Thecameraprovidedtheassociatedpixelsonagreyscale(inarbitraryunitsfrom0to1024digits)andtheluminance meter the corresponding luminance (cd/m2). The correlation between the pixelgrayscale digits and the luminance values is illustrated in Figure 3.10. The luminance wasassociatedtothegreyscale[168]byanexponentialfittingfunction,asgivenbyEquation3.6.Thecorrespondingcoefficientofdeterminationishigh:R2=0.97.

y=e0.017x

(3.6)

Theexponentialfunctionwasimplementedinthe“ImageCAM”softwareandwasdevelopedbyLESO‐PB,EPFL(Lausanne,Switzerland),inordertosupporttheluminancemappingfunctionoftheCLLS[167].

Figure 3.10: Correlation between the pixel grayscale values and the associated luminances. The blackcrossesindicatethemeasurements;thesolidlineindicatesthefittedexponentialfunction[167].

Vignettingcorrection

A Fujinon fisheye lens (index “FE185C057HA”) [166] was installed on the CLLS in order tomimicthefieldofviewofahumaneye(≈120°x120°)[1,155].Fisheyelensesnormallyexhibitnoticeablelightfalloffs(so‐calledvignettingeffects)forimagepixelslocatedfarfromtheopticalaxis.Figure3.11illustratesthevignettingeffectoftheCLLSequippedwiththeFujinonfisheyelens.

Grayscale

0 100 200 300 400 500 600 700

Luminance(cd/m

2 )

0

2000

4000

6000

8000

10000

12000

14000

16000

18000

CLLSvsLuminanceMeter

Exponential;R2=0.97


59

Figure3.11:VignettingeffectsoftheCLLSequippedwithafisheyelens.

Vignettingeffectshadaccordinglytobecorrectedtocompensatetheluminancelossatthelens’sedges.Toassessthepotentialvignettingeffects,ameasurementwasset‐upasshowninFigure3.12a)‐c). The camerawas fixed to a rotating tripodandplacedonan axis orthogonal to thelight source. It was then rotated by using 5° steps both in the horizontal and the verticaldirections; snapshots were taken at each position, while maintaining the luminous intensityconstant,asshowninFigure3.12d)[167].

Figure3.12:Measurement set‐up for the vignetting correction; a‐c) CLLS located in front of a constantlightsource,d)thedevicewasrotatedin5°anglestepsinboth,horizontalandverticaldirectionsinordertotakepicturesatdifferentangles.

a)

d)

b)

c)


60

Theluminancesofthelightsourceatdifferentangleswereextractedfromtheluminancemapsgeneratedby the“ImageCAM”software.Themonitored luminancesof the lightsourceplottedforeachangleareillustratedinFigure3.13;theywerethenfitted(R2=0.97)withapolynomialfunctionaccordingtoEquation3.7.Aninverseequationwasthencomputed(Equation3.8)andfinallyimplementedinthe“ImageCAM”softwareinordertocorrectthevignettingeffects.

y 2.16 ∙ 10 ∙ 1.47 ∙ 10 ∙ 5.52 ∙ 10 ∙ 2.72 ∙ 10 ∙ 2.66 ∙ 10 ∙

1.000473

(3.7)

y 2.27 ∙ 10 ∙ 1.60 ∙ 10 ∙ 5.91 ∙ 10 ∙ 2.71 ∙ 10 ∙ 2.88 ∙ 10 ∙

0. .9995

(3.8)

Figure3.13:Characterizationof thevignettingeffectsof theCLLS fisheye lens.Thecrosses indicate themeasurements for different angles; the solid line shows the fitted polynomial function (according toEquation3.7);0=centre.

3.4.3 ExperimentalvalidationofCLLS

Luminancemapping

Aftertheoverallcalibrationprocedure,luminancemappingofrealsceneswasperformedwiththeCLLS.TheluminancemapswerevalidatedbyusingacalibratedMinoltaLS‐110luminancemeter (Konica Minolta, Tokyo, Japan). The validated luminance maps were compared withluminancemappingsobtainedwithaNikonCoolpix5400CCDcamera,equippedwithaFCE‐9

AnglefromCentre(Degrees)

‐80 ‐60 ‐40 ‐20 0 20 40 60 80

RelativetoCenter(‐)

0.85

0.90

0.95

1.00

1.05

1.10

MeasuredDatafromCLLS

PolynomialFunction;R2=0.97


61

fisheye lens.Thevalidationwascarriedout inanoffice roomat theLESOSolarExperimentalbuilding on the EPFL campus (Ecole Polytechnique Fédérale de Lausanne, Switzerland). Astandardizedcolourchart(Macbethcolourchecker,X‐Rite,Michigan,USA)aswellasdifferentroomelementswereusedastargetsforluminancemapping[151].Asetofimageswastaken:i)under dynamic daylighting conditions and ii) under electric lighting conditions (2 x 36 Wfluorescent tubes; 3000K). Fifteen surfaces of different room elements were used beside thestandardizedcolourchart;thelocationsoftheroomelementsareshowninFigure3.14.

Figure3.14:Locationsof15differentroomelementsfortheCLLSexperimentalvalidation:1=PCscreen(white),2=PCscreen(black),3=desktopnearkeyboard,4=keyboard,5=desktoprightside,6=lowerwall,7=middlewall2,8=upperwall,9=wallnearceiling,10=wallleftside,11=shelf‐top,12=shelf‐left,13=shelf‐middle,14=desktopoftheleftsidedesk,15=desktopleftsideandthecolourchartonthedesk.

The luminance values of the targets were extracted in two different ways: i) using theimplemented “ImageCAM” software for the CLLS images and ii) using the software“Photosphere” [160] for images obtained by theway of HDR techniques (CCD Camera NikonCoolpix).Bothdatasetswerethencomparedwiththe luminancemonitoredbythe luminancemeter. The luminance data of the colour chart and the room elements obtained under pureelectriclightingconditions(EL)anddaylightingconditions(DL),areshowninFigure3.15.Thecoefficientsofdetermination(R2)betweentheluminancemappingsachievedwiththeCLLSandthe‘ClassicalHDR’imagingtechniqueassociatedtotheluminancemeterareshowninTable3.3(fordifferentlightingconditions).


62

Figure 3.15: Luminancemapping derived from standardized colour charts (Macbeth Colour Checker)anddifferentofficeroomelements.ThedatapointswereextractedfromHDRimagestakenwiththeCLLS(orange bars) and a Nikon Coolpix CCD camera (HDR imaging technique; blue bars). The referenceluminance(fromtheluminancemeter;greybars)arealsoshown:a)assessmentofthecolourchartunderelectric lighting conditions; b) assessment of the room elements under electric lighting conditions; c)assessmentofthecolourchartunderdaylightingconditions;d)assessmentoftheroomelementsunderdaylightingconditions.

Luminancemappingfrom

luminancemeterincomparisonwith

Coefficientdeterminant(R2)ColourChart RoomElements

EL CLLS 0.9838 0.9994‘ClassicalHDR’ 0.9872 0.9990

DL CLLS 0.9957 0.9803‘ClassicalHDR’ 0.9953 0.9946

Table3.3:Coefficientofdetermination(R2)ofluminancemappingwiththeCLLSandthe‘ClassicalHDR’techniquecomparedwiththeluminancemeterfordifferentlightingconditions.

DaylightingCondition(DL)

c)ColourChartunderDL

Color1

Color2

Color3

Color4

Color5

Color6

Color7

Color8

Color9

Color10

Color11

Color12

Color13

Color14

Color15

Color16

Color17

Color18

Color19

Color20

Color21

Color22

Color23

Color24

Luminance(cd/m

2 )

0

100

200

300

d)RoomElementsunderDL

Room1

Room2

Room3

Room4

Room5

Room6

Room7

Room8

Room9

Room10

Room11

Room12

Room13

Room14

Room15

Room16

Luminance(cd/m

2 )

0

200

400

600

800

1000

ElectricalLightingCondition(EL)

a)ColourChartunderEL

Luminance(cd/m

2 )

0

5

10

15

20

25

30

CLLSLuminanceMeterClassicalHDR

b)RoomElementsunderEL

Luminance(cd/m

2 )

0

50

100

150

200

250


63

Glarerisksassessments

AsetofHDRimageswassimultaneouslygeneratedusingboth,theCLLSandtheNikonCoolpixinaLESOdaylitofficeroomduringthreedifferentdays.TheEvalglareroutine[148]wasusedtocompute glare indices for both images. The Daylight Glare Probability (DGP), Daylight GlareIndex(DGI)aswellastheUnifiedGlareRating(UGR)weredetermined; theyarepresented inTable3.4.Thevaluesobtainedwiththe ‘ClassicalHDR’technique(CCDcameraNikonCoolpix)wereslightlyhigherthanthoseoftheCLLS:therelativedifferencesacrossthreedaysareequalto2.9%fortheDGP,12.0%fortheDGIand15.4%fortheUGR.

GlareIndices

Day1 Day2 Day3

CLLS‘ClassicalHDR’ CLLS

‘ClassicalHDR’ CLLS

‘ClassicalHDR’

DGP 0.22 0.22 0.23 0.23 0.22 0.24DGI 12.21 12.74 8.31 10.79 13.3 14.91UGR 14.38 16.40 11.83 15.21 18.34 21.04

Table3.4:Glare indicesobtained fromimagesat identicalofficeroompositionsanddifferentdayswiththeCLLS and theNikonCoolpixCCD camera (‘ClassicalHDR’).DGP=DaylightGlareProbability;DGI =DaylightGlareIndex;UGR=UnifiedGlareRating[167].

Glarerisksassessmentfordifferentlightingconditions

InordertoconfirmtheresultsobtainedfromHDRimageswiththeCLLSandtheNikonCoolpixfor different (day‐) lighting conditions, the overall procedure was repeated in a LESO officeroom. A set of LDR images was taken with the conventional CCD camera (duration of HDRprocedure: ~60s) while a sequence of real‐time digital images was simultaneously recordedwiththeCLLSundersteady‐stateelectric lightingconditions(for10s)anddaylightconditions(for60s).Threeglareindices(DGP,DGI,UGR)andbackgroundluminanceswerethencomputedusingEvalglare.Thevaluesobtainedbyusingthethreeglareindices,aswellasthebackgroundluminancearepresentedinTable3.5a)forelectricallighting,andTable3.5b)fordaylighting.

Forelectriclightingconditions,theCLLSvaluesobtainedafter10saswellasthoseissuedfromaconventional ‘HDR imaging’, were compared with data obtained with the CLLS at 0s. Theobserveddifferencesaresmallerthan3%forbothtechniques,except fortheDGI(5.5%lowerthanwiththeCLLS)andthebackgroundilluminance(11.8%greaterthanwiththeCLLS).

Table3.5billustratestheglareindices,asmeasuredwithCLLS(threedifferentrecordingtimes)and the Nikon Coolpix for varying daylighting conditions. For the latter, no difference wasobservedfortheDGPacrossthedifferentrecordingtimesorbetweenthetwotechniques.ThevariationsfortheDGI,UGRandthebackgroundilluminationwereintherangeof0%to36.2%(mean ± SD: 11.3 ± 13.2%). The largest difference for background illumination (36.2%)wasfound for imagesderived fromthe ‘ClassicalHDR’ technique [167].TherelativedifferencesoftheglareindicesandthebackgroundluminancesareillustratedinFigure3.16[167].


64

LightingConditions

GlareIndices(Daylight)

DGP DGI UGR BackgroundLuminance

a)ElectricLightingCondition

CLLS<1s

0.18 9.21 11.74 48.0

CLLS~10s

0.18 9.17 11.69 48.04

‘ClassicalHDR’From14LDR

imagesduring~60s

0.18 8.7 11.34 54.4

b)DynamicDaylightingCondition

CLLS<1s

0.23 11.61 15.53 137.2

CLLS~10s

0.23 12.55 16.70 101.62

CLLS~20s

0.23 10.89 15.08 154.96

CLLS~30s

0.23 11.36 15.29 200.61

CLLS~40s

0.23 11.60 15.60 191.82

CLLS~50s

0.23 11.58 15.62 201.58

CLLS~60s

0.23 11.68 15.75 201.65

‘ClassicalHDR’From14LDR

imagesduring~60s

0.23 10.79 15.21 215.14

Table 3.5: Glare indices (DGP, DGI and UGR) and background luminance were assessed at the samepositionfora)steady‐stateelectrical lightingconditionsandb)dynamicdaylightingconditionswiththeCLLS(between<1sand60s),andtheNikonCoolpix(‘ClassicalHDR’;during~60s).


65

Figure3.16:RelativedifferencesbetweentheCLLS(from10sto60s)andtheNikonCoolpix(after60s)fordaylightconditionsregardingglareindicesandbackgroundluminanceat~0s(withCLLS).Blackcircles=DGP;whitecircles=DGI;blacktriangles=UGRandwhitetriangles=backgroundluminance.

Discussion

TheCLLSwassuccessfullycalibratedandtestedinrealscenes.Luminancemapsweregeneratedwith the CLLS and comparedwith those obtained by conventionalHDR techniques. Differentglare riskassessmentsundersteady‐stateandvaryingdaylightingconditionsshowed that thenewsensorhastheadvantagetocaptureHDRimagesmorerapidly,whichisofgreatbenefitindynamic(day‐)lightingsituations.

The calibration procedure and vignetting corrections resulted in high correlations with thereferencedevices:theCLLScanthusbereliablyusedtomonitorrealscenes.Thecoefficientofdetermination(R2)betweenluminancemapsgeneratedbytheCLLSandtheconventionalCCDcameracomparedtoapoint‐to‐pointluminancemeterwaslargerthan0.98.Thisconfirmsthatbothdevices(CLLSandconventionalHDRimaging)producedsimilarresults,evenwhentestedfordifferentlightingconditions.

Under steady‐state electric lighting conditions, the glare indices’ differences between bothapproacheswererathersmall(lessthan6%),exceptforthebackgroundluminance,wheretheconventional CCD camera assessed higher values. Under varying daylighting conditions, thedifferencesofglareindicesbetweenbothtechniqueswerelarger,comparedtoelectricallightingconditions; the background luminance varied by a maximum of 36%. The main parameterstaken into account for the glare indices calculation, were the background luminance, theluminance of potential glare sources, and the solid angles [45, 153]. The glare indices’differencesaremostlikelyduetodifferentbackgroundluminance.Whenthelattervarieswithinshort time periods, the glare indices are also modified. Distortions of the two fisheye lensesmight also indirectly contribute to the differences of glare indices, since they can impact thesolidangles.ThedifferencesofDGPbetweenbothapproachesweresmaller than thoseof theDGIandUGR.Thiscanbeduetotheverticalilluminance,whichisanadditionalDGPparameter[148].Thevertical illuminances for theDGPcalculationwere similarwhatever techniquewas

Icycam10s 20s 30s 40s 50s 60s 'ClassicalHDR'

RelativeDifferences

0.0

0.1

0.2

0.3

0.4DGPDGIUGRBackgroundluminance


66

used:whentheyweremanuallymeasuredduringtheimagecapturing,theDGPrevealedsmallerdifferencesthantheothertwoglareindices[167].

3.4.4 Outlook

Theresultsof thecalibrationandvalidationof theCLLSwerereportedrecently inconferenceproceedings [167]. The main advantage of the CLLS luminance mapping is its high‐speedperformance. Besides the benefits of time savings, a more accurate assessment is achievedcompared to ‘Classical HDR’ imaging techniques, particularly under dynamic daylightingconditions.TheCLLScamerahasthepotentialtobecomeanewandusefulapplicationtoolforarchitects, building designers and lighting experts. Dynamic CLLS luminancemapping is thusadvantageous, in particular for the assessments under varying lighting conditions such asdaylighting.Thehighspeedof theCLLSallowsalsoconsideringusing it in the futureasa skyscannertoassessskyluminancedistributionsevenwithverydynamicdaylightconditions.Sincethe CLLS is equipped with photopic filters, based on human cone sensitivity, it can also beemployed to assess lighting situations in the visible range of wavelengths. An additionalpotential applicationofCLLSmightbe tousea secondsensitivity curvewhich isblue‐shifted,alsoknownasC(λ),tomonitorbiological,non‐imageformingeffectsoflight;thisisconsideredinthenextsection.

3.5 ApplicationofCLLSforassessmentofnon‐imageformingeffectsoflight

3.5.1 Assessmentofnon‐imageformingeffectsoflight

TheC(λ)curveorthecircadianefficiencyfunctionshavebeenintroducedinSection3.1.AsthecircadiansensitivityfunctiondiffersfromtheV(λ)curve,lightmonitoringregardingNon‐ImageForming(NIF)effectsdiffersfromphotopicmeasurements.

Inthissection,theCLLSisadaptedtotheC(λ)spectralsensitivityfunctioninordertoevaluatelight distributions with respect to non‐visual functions. Using the C(λ) function, luminousdistributionmaps of a circadianweighted radiance (Lec) can be created. The CCLS, equippedwithcustomized filters,enables toadapt thecamera’sspectralsensitivity to theC(λ) function.The distributions of Lec using circadian luminance mapping under electric and daylightingconditionsweredeterminedatdifferenttimesofday.


67

3.5.2 Calibrationprocedure

Spectralsensitivitycalibration

As described in Section 3.4, the CLLS was calibrated based on the photometric sensitivityfunction V(λ) and corrected for vignetting effects; in this section, the spectral calibration isdescribed with respect to the circadian sensitivity function C(λ). The experimental set‐up issimilar to the one used for V(λ) (Figure 3.7). Narrow‐bandwidthmonochromatic light beamswerealsousedasareferencelightsource.TheCLLScaptureddigitalimagesofthe470nmlightbeams for different intensities. Simultaneously, the emitted radiance was measured with acalibratedspectroradiometer(Specbos1201,JETI, Jena,Germany).TheCLLSprovidedasinglevalueperpixelonagreyscale(inarbitraryunitsfrom0to1024digits).Thegreyscalevaluesandthe emitted radiance of the spectroradiometer were correlated using a polynomial functiongivenbyEquation3.9.TheresultsshowedahighcoefficientofdeterminationwithR²=0.9998.

y 1.25 ∙ 10 ∙ 5.06 ∙ 10 ∙ 1.15 ∙ 10 ∙ 8.62 ∙ 10 ∙ 7.28 ∙ 10 ∙ 3.16 ∙ 10

(3.9)

TheCLLSandthespectroradiometermonitoredsimultaneouslytheconstantemittedradianceofalightsourceatdifferentwavelengthsintherangeof380to780nm,separatedby5nmsteps.The raw spectral sensitivity of the CLLS, normalized at 470 nm, and the circadian sensitivitycurve,asreportedfromtheliterature[115],areshowninFigure3.17a.Thissensitivitycurveisbasedontheactionspectrumforlight‐inducedmelatoninsuppression,suggestedbyBrainardetal.[118]andThapanetal.[119],asmentionedinSection3.1(withapeaksensitivitywithin460‐480nm).

ThreeappropriatefilterswereidentifiedandcombinedinordertomatchtheC(λ)curve:Scott‐S8612,ScottGG420,andKopp5030filters(SchottAG.Mainz,Germany;KoppGlassInc.,PA,USA)wereused for thatpurpose. Theoptimal thicknessof the filterswascomputed tocorrect thespectral responseand implementedonto theCLLS.Todetermine theCLLSspectral sensitivityfunction, the sameprocedure alreadyused for theV(λ) filterwas applied in this case (Figure3.17b). The corrected sensitivity function is shown in Figure 3.17b andwas compared to thecircadiansensitivityfunctionsuggestedbyGall[115].InordertoassesstheerrorbetweentheCLLSrelativespectralsensitivityandtheC(λ)curve,theCIEstandarderrorf1’suggestedfortheV()function(Equation3.5[138])wasappliedtotheC()functionasshownbyEquation3.10.Astandarderrorof10.4%wasobtainedinthiscaseusingthemodifiedformula.

dCsf rel )()(93584.0'0

1

(3.10)


68

Figure3.17:RelativespectralsensitivityassessmentoftheCLLS:a)relativerawspectralsensitivityoftheCLLS (normalized data at 470 nm; black circles) and circadian sensitivity function (C() [115]; whitecircles), b) corrected spectral sensitivity response of theCLLS equippedwith filters (black circles) andcircadiansensitivityfunctionC()(whitecircles[115]).

3.5.3 Circadianweightedradiancecalibration

Toperformthecircadianweightedradiance(Lec)calibration,atotalof83measurements(from0.04 cd/m2 to 23,871cd/m2) were made by using simultaneously the CLLS and a calibratedspectroradiometer(Specbos1201,JETI,Jena,Germany)asthereferencesensor.Polychromaticwhitelightfroma1000WXenonlampandfroma1200Wmetalhalidespotlightwereusedasreference light sources. The spectroradiometermonitored the Lec values (W/sr∙m2),while thecameraprovidedtheassociatedpixelsonagreyscale(inarbitraryunitsfrom0to1024digits).ThecorrelationbetweenthegreyscalevaluesandthemeasuredLecisshownonFigure3.18.Thebest fitbetween thepixelgreyscalevaluesand theirassociatedLecwasdeterminedusing twodifferentfunctions:

For greyscale values lower than 425 digits: the exponential function expressed byEquation3.11wasemployed(R2=0.97);

For greyscale values higher than 425 digits: the polynomial function expressed byEquation3.12wasused(R2=0.98).

Both functions were finally implemented in the “ImageCAM” software in order generate Lecmaps.

0.00013 .

(3.11)

y 6.22 ∙ 10 ∙ 7.86 ∙ 10 ∙ 4.13 ∙ 10 ∙ 1.04 ∙ 10 ∙ 1.24 ∙ 10 ∙ – 5.56 ∙

10 ∙ 5.30 ∙ 10

(3.12)

Wavelength (nm)

300 400 500 600 700 800

Rel

ativ

e S

pect

ral S

ensi

tivi

ty (

-)

0.0

0.2

0.4

0.6

0.8

1.0

1.2Raw spectral sensitivity of CLLS without filterCircadian sensitivity curve [4]

Wavelength (nm)

300 400 500 600 700 800

Rel

ativ

e S

pect

ral S

ensi

tivi

ty (

-)

0.0

0.2

0.4

0.6

0.8

1.0

1.2Corrected sensitivity response of CLLS with filtersCircadian sensitivity curve [4]


69

Figure3.18:Correlationsbetweenpixelgreyscalevalues(arbitraryunitsfrom0to1024)andassociatedLec(W/sr∙m2).TheblackcrossesindicatemeasurementstakenwiththeCLLSandthespectroradiometer;thesolidlineindicatesthefittedexponentialfunctionforthegreyscalevalueslowerthan425(R2=0.97);the dashed line indicates the fitted polynomial function for the greyscale values higher than 425(R2=0.98).Thevertical dashedanddotted linedepicts theborder for the two regression functionsat agreyscalevalueof425digits.

3.5.4 ExperimentalvalidationoftheCLLSwiththeC(λ)filter

Circadianweightedradiancemapping

TheCLLSwastestedinanofficeroomoftheLESOSolarExperimentalbuildinglocatedontheEPFLcampus(EcolePolytechniqueFédéraledeLausanne,Switzerland).Twentydifferentroomelements were used as targets for measurements (Figure 3.19). A set of pictures was takenunder electric lighting conditions (2 x 36 W fluorescent tubes; 3000K) as well as underdaylighting conditions (clear sky). The Lec values of the room elements were simultaneouslyassessedbythespectroradiometerandbytheCLLS.Thetwodatasetswerethencompared,asshown inFigure3.20.Thecoefficientsofdetermination (R2)betweenLecmeasuredwithCLLSandthespectroradiometer.

Figure3.19:Locationsofdifferentroomelementsforthe validation of the CLLS. 1=desktop, 2=PC screen,3=desktop leftside,4=desktoprightside,5=seconddeskleftside,6=chair,7=telephone,8=window,9upperwindow, 10= backwall 1, 11=white board,12=backwall2,13= closet,14=wall left side,15=door,16=bottomofthedoor,17=poster,18=behindPC screen, 19=document on PC screen, 20=blackwallpaperonPCscreen.

0

5

10

15

20

25

0 100 200 300 400 500 600 700

Circadian W

eighted Radiance (W/sr*m

2)

Grayscale(0‐1024digits)

Greyscalevs.Lec

ExponentialFunctionR²=0.97

PolynomialFunctionR²=0.98


70

Figure3.20:Lecvalueofdifferentofficeroomelements(seeFigure3.19).ThedatapointswereobtainedfromCLLS(blackbars)andthereferenceluminancemeasurements(spectroradiometer;greybars).Left:assessmentunderelectriclightingconditions;right:assessmentsunderdaylightingconditions.

Comparison of circadian weighted radiance mapping for different daylightingsystemsandtimesofday

TheCLLSwasalsoemployedintwotestroomsoftheAdvancedWindowsTestbedfacilityattheLawrence Berkeley National Laboratory (LBNL; CA, USA), during a short‐term visit and acollaborationbetweenLESO‐PB/EPFLandLBNL.RoomA is equippedwith standard venetianblinds (Figure3.21a);RoomC is equippedwithLight Louvers™ (Figure 3.21b).Both complexfenestration systemswere located in theupperpartof thewindows,whereas the lowerpartswerecoveredwithopaquescreens.AsetofimageswastakenwiththeCLLSat9AM,12PMand3PM. Luminance and Lec were extracted for the same reference points in both rooms (walls,windows,taskareaandceiling).TheratioofLecandluminance(Lec/L)wasthendeterminedtoassess the circadian efficiency of the light distribution in the room; a higher ratio indicates alargercircadianefficiency.Atotalof84measurementsweretakenunderclearskyconditions;extracted luminance, Lec and ratio on log‐transformed drawings were analysed with 2‐wayrepeatedanalysesofvariance(rANOVA)withthefactors‘time’and‘room’.

ElectricLight

RoomElement

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

L ec(w/srm2 )

0.00

0.02

0.04

0.06

0.08

0.10

0.12

0.14

0.16

0.18

CLLS;R2=0.96

Spectroradiometer

Daylight

RoomElement

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Lec(w/srm2 )

0.0

0.5

1.0

1.5

2.0

2.5

CLLS;R2=0.91

Spectroradiometer


71

Figure3.21:a)RoomAequippedwithstandardvenetianblinds;b)RoomCequippedwithLightLouver™

Luminance and Lec in Room A (venetian blinds) were higher than those in Room C(LightLouverTM) for all three timesofday. For rooms, luminanceandLecwere larger at12PMthanat3PMandlowerat9AM(Figure3.22a‐b, ‘room’x ‘time’;p<0.05).TheratioofLec/LwasoveralllargerinRoomAthanthatinRoomC(p<0.05;maineffectof ‘room’Figure3.22c);theratioforbothroomswashigherat3PMthanat9AM(p<0.05;maineffectof‘time’;Figure3.22c).Firstcomparisonsbetweentwodifferentlocationsintheroom(nearbythewindowanddeeperintheroom),didnotrevealanydifferenceincircadianefficiencyoflightdistribution(p>0.20).

Figure 3.22: a) Luminance, b)Lec and c) ratio (Lec /L) for Room A (black circles) and Room C (whitesquares)at9AM,12PM,and3PM(mean±SEM);overall, largerluminance,LecandratioinRoomAthanRoomC;largerluminanceandLecat12PM;largerratioat3PMthan9PM(*=p<0.05).

Discussion

TheCLLSwassuccessfullyusedinrealscenes;itrevealedthatthecorrelationofLecissuedfromthe CLLS and the spectroradiometer was larger for steady‐state electric lighting than fordaylightingconditionsunderclearsky.

The room equippedwith standard venetian blinds provided larger (photopic) luminance andcircadianweightedluminance(Lec)thantheroomequippedwithLightLouverTMthroughouttheday;largervalueswereobservedatnoon.ThemostlikelyreasonforthedynamicsofluminanceandLecisthevaryingincidentangleofsunlight.Interestingly,circadianefficiencywashighestat3PM;onereasonforthismightbethatthedifferentsourcesofdaylight(sunandskyvault)didnotonlyprovidedifferent light fluxes,butalsoinducedchangesinthespectralcompositionofdaylight.

5b) Lec

Time

9AM 12PM 3PM

W/s

r*m

2

0.0

0.2

0.4

0.6

0.8

1.0

1.2 Room ARoom C

5a) Luminance

Time

9AM 12PM 3PM

cd/m

2

0

200

400

600

800Room ARoom C

* **

*

*

*

5c) Ratio (Lec/L)

Time

9AM 12PM 3PM

[-]

0.0008

0.0010

0.0012

0.0014

0.0016

Room ARoom C

*

*


72

3.5.5 Outlook

The calibration and the validation of the CLLS were previously published in conferenceproceedings[171].FurtherstudieswiththeCLLSwithrespecttocircadianweightedluminanceat different room locations and different times of day are required. One important questionremains:Whatdoesahigherorlowercircadianefficiencymean?Canitbeusedasaproxyforother biological functions? Future experiments should also test other variables, for examplecircadianefficiencyoflightforhumanalertness,moodandperformance.

3.6 Conclusion

Photometric variables, such as luminous intensity and colour appearance of light, have beengenerally used for characterization of office lighting. However, office lighting criteria forcircadian metrics are not established yet. Lighting conditions must be further investigated,especiallywithrespecttoNIFeffects.Appropriatemonitoringandeffectivedevicesincircadianmetricsmustbefurtherdevelopedandbecomeinternationallyaccepted.

An efficient monitoring of luminance distribution, which has an important impact on visualcomfort, has yet to be developed. Since the HDR imaging technique has been introduced,luminancedistributioncanbeassessed.However,theassessmentofvisualdiscomfortbyHDRcannot be performed instantaneously: a set of photographs must be acquired to generate asingle HDR image, and capturing several LDR images takes considerable time. The varyingluminance of glare sources or background might induce different results, particularly underdynamicdaylightconditions.

In the course of this thesis, a Camera‐Like Light Sensor (CLLS) has been introduced andcalibratedforbothphotopicandcircadiansensitivityfunctions.Thenoveldevicehasalsobeensuccessfully validated, showing that the CLLS considerably facilitates luminance mappingcompared to the HDR technique with respect to speed and accuracy. Besides the use forluminancemapping,acircadianweightedradiancemaphasbeenintroducedforthefirsttime.With the CLLS, the circadian weighted radiance (Lec) can be easily monitored. A firstexperimental assessment of circadian weighted luminance maps for Complex FenestrationSystems (CFS) was also carried out in this thesis; it reveals different circadian weightedluminousdistributions.ThisparticularuseoftheCLLSisthusinnovativefortheassessmentsoflight properties with respect to NIF functions within architectural settings. The CLLS has inadditionthepotentialtobecomeanovelandveryusefulapplicationtoolforarchitects,buildingdesignersandlightingexperts.

73

Chapter4 Non‐ImageForming(NIF)biologicaleffectsoflight

4.1 PrinciplesofNIFeffectsoflight

More thanadecadeago,anovelphotoreceptorclass, theso‐called intrinsicallyphotosensitiveRetinalGanglionCells (ipRGC),wasdiscovered in themammalianretina [116,117,172,173].Sincethen,ithasbecomeevidentthatocularlightinginfluencesnotonlythevisualsystem,butalsomanynon‐visualeffectsonawiderangeofbiological functions,suchasphysiology,moodandbehaviour.Theseeffectsarecalled“Non‐ImageForming”(NIF)functions.

Light via the ‘visual path’ is conveyed by rods and cones. Incoming light information(electromagnetic rays) is transformed into electrochemical signals and sent via polysynapticpathwaystothevisualcortex.LightviaNIFpathwaysisconveyedbyintrinsicallyphotosensitiveRetinalGanglionCells(ipRGC)viatheretinohypothalamictracttotheSuprachiasmaticNucleus(SCN)andotherbrainregions,suchasthepinealglandwhichsynthesizesmelatonin:Figure4.1schematicallyillustratesthepathwaysofvisualandNIFeffectsinthebrain.TheIpRGCsaremostsensitiveinthebluepartoflightbetween460‐480nm[118,119,174],whencomparedtocones(555 nm) and rods (505 nm), as mentioned in Section 3.1. Besides wavelength‐dependentdifferences,otherphysicalpropertiesofthevisibleelectromagneticspectrumsuchasluminousintensity,timing,andexposuredurationtolightalsomodulatethemagnitudeofNIFeffects.Inordertooptimizeindoorlightingconditions,thefundamentalsoftheluminouspropertiesthatimpact NIF effects must be recognized. Environmental light exerts a signal to regulate NIFeffects, such as circadian, circannual (seasonal) rhythms [175], the pupil light reflex andhormonalsecretions[11‐15].Manyoftheseluminouseffectsalsoinfluenceofficeworkersandwillbeexplainedinthefollowingsection.

Non‐ImageForming(NIF)biologicaleffectsoflight

74

Figure4.1:Schematicoverviewofvisual(red)andNIF(blue)pathwaysinthehumanbrain(adaptedfrom[176]).

4.1.1 Circadianrhythms

ThecircadiansystemwithitsmasterclockintheSCNgovernsourdailyrhythmssuchasbodytemperature,sleepandwakefulness,hormonalsecretion,andmostotherphysiologicalfunctions[9, 15‐18]. Endogenous concentration changes of two hormones, melatonin and cortisol areoften used as circadian ’phase markers’ in humans. Melatonin, also referred to as a “darkhormone”, is secretedby thepineal glandduringnighttime. It relates to several physiologicalprocessesincludingcircadianentrainment,retinalphysiologyandseasonalreproduction[177].It reflects the environmental photoperiod via its secretion profile; high levels are secretedduring the night, and very low levels or nomelatonin is secreted during daytime [178‐180].Sincemelatoninisimportantforcircadianrhythms,light‐inducedmelatoninsuppressioncanbeused to determine themagnitude of these acute light effects [11, 15, 181, 182] Cortisol, alsoknown as “stress hormone” is maximally secreted around habitual wake time and falls tominimumclosetohabitualbedtime.Morninglightexposure,beforethedailycortisolpeaks,canadvancethecircadianphaseofcortisolrhythms,whilsteveninglighthasnoorlittlesignificanteffectoncortisolsecretion[25].

Byourdailylightanddarkcycle,thebiologicalclockintheSCNissynchronisedtotheexternal24‐hour cycle. The average period length of human circadian rhythms is approximately 24.2hours [111]. The circadian phase can be shifted by bright light pulses, depending on theirluminous intensity, timing and duration [182‐185]. Circadian response curves and circadianactionspectraforhumanswerecreated[118,119].Disruptionofthissynchronycauseschronicsleepdisordersandmanyothersymptoms,alsodescribed incombinationwith jet‐lagorshiftworkdisorders[186,187].Lackofsufficientlightstimuliduringdaytimeandinappropriatelightstimuli during nighttime influence circadian disruption, which in turn often leads to theemergenceofsleepdisordersandothermedicalorpsychiatricdiseases[186].


75

4.1.2 Behaviouraleffectsinofficeworkers

PreviousstudiesreportedseveralNIFeffectsof lightonhumanphysiologyandbehaviour;thefollowingconsequencesaresomeexamplesofdaily(light‐dark)environmentaleffectsonofficeworkersandtheirworkperformances.

Alertness

Under controlled conditionswith scheduled sleepepisodes, subjective andobjective alertnessfollows a circadian rhythmicity with a nadir (lowest alertness) at approximately 4h in themorning [188]. The presence of (day‐) light intuitively relates to an alert or awake state inhumans [189].Thealertingeffectsof lightaredeterminedbymanyaspects, suchas luminousintensity (illuminance), spectral composition of light and exposure duration [189]. Themechanisms for time‐of‐day dependent effects of light‐induced alertness changes during day‐and nighttime are still unclear [189, 190]. There have been many studies showing greateralertingresponseswithbrightlightexposuresatnightwhencomparedtodimlightconditions[190‐194].Most of these studies concluded that nocturnal light‐induced alerting effectswerealso driven by light‐induced melatonin suppression [12, 189, 193]. The explanation is thatmelatonin can induce sleepiness and therefore its nocturnal suppression by light increasesalertness[12,189,193].However,duringdaytime,melatoninsecretion isalmostzero,eveninvery dim light and thus, light‐induced alerting effects during daytime cannot be conveyed bymelatoninsuppressionalone[189,195].Severalstudiesreportedalertingeffectsbylightduringdaytime as assessed with subjective alertness assessments [190, 196], performance tests(Psychomotor Vigilance Task; PVT) [196], and functionalmagnetic resonance imaging (fMRI)[197]. One of the studies found that nighttime and daytime bright light similarly reducedsleepinessandfatigue[190].TwostudiesanalysedthealertingresponsesassessedbyPositronEmission Tomography (PET)[198] during nighttime and by fMRI during daytime [197]. Theirresultsshowedthatpolychromaticlightwasnotonlyconfinedtohypothalamicareasincludingthepinealgland(wheremelatoniniscontrolled),butalsoextendedtothemodulationofalarge‐scalenetworkofcorticalareas[197,198]. Recently,Rautkyläetal.[195]suggestedthatlight‐induced alertness might involve also other brain mechanisms, which are different from thecircadianpathway.Lightmayconnectalsotolimbicpathways:thelimbicsysteminhumansisimportantforemotionalresponses[195].

Sleepquality

The24‐hsleep‐wakecycleisoneofthemostimportantcircadianrhythms.Poorsleepcanaffectphysical and mental health [199, 200] and induce work‐related problems in cognitiveperformance[201].Improvingsleepqualityincreasesindividualhealthandwell‐beingincludingorganizationalproductivity[201].

Polychromatic white or blue‐enriched white light exposure during the day increases sleepquality the followingnight [134,202]: it is the timingof lightexposurewhichdetermines theimpact on sleep and wakefulness. Bright or blue‐enriched light in the evening has a phasedelaying effect and thus often results in greater problems to fall asleep [203, 204]; evenelectronicdevicessuchaslaptopsinfluencethefollowingsleepepisode[205,206].


76

Mood

Severalcomplexaspectsincludingindividualphysiologyandsocialactivitiescaninfluencemood[207]. It can vary significantly with circadian phase, but there was no reliable effect of theduration of prior wakefulness [208]. However, a recent study [207] considered data frommillions of public online messages across 84 countries for mood expressing words andsentences.Theauthorsfoundthatinthemorning,individualsawokegenerallyingoodmood;inthecourseof theday,mooddeterioratedover time.Moreover, thestudypointedoutseasonalchangesofmood:theyweredependentonthechangesinday‐length[207].

Moodcanalsobeinfluencedbylight,dependingonilluminanceandtimingofexposure.Brightlightduringdaytime,particularlydaylightorbright lightduringwinter, can improvemood inhealthy persons and patients with seasonal affective disorders [209‐213]. Many studiesinvestigatedtheeffectsofelectriclighting(mainlyfluorescentlightsources)onmoodinhealthyvolunteersduringdaytime;however,thereisnoclearconclusionwhetherdifferentlightcolourshaveaneffectornotdue toa lackofstrongevidence,according to thereviewsbyVeitchandMcColl[67,214].Interestingly,somestudiesfoundthatbrightlightovernightdidnotappeartohavethesameeffectonmoodasduringtheday[211,215].Itwasnotedthattheexperimentalset‐upwithacontrolledroutinewithoutsleepduring thenightmighthavealreadyanegativeeffectonmood;thus,short‐termlightexposuresmightnotbesufficienttochangemoodinthatcase [188]. On the other hand, the other field studies reported bright light during nighttimeenhancedthemoodofnightshiftworkers[216‐218],suchasnightshiftnurses[216,217].

Healthandwell‐being

The short‐termeffects of light onhealth andwell‐beingwere recently investigated in severalstudies[17,18,46,79,219‐222].Theterms‘healthandwell‐being’wereusedfordifferentself‐reportedquestionnaires:

Subjectiveperceptions:ocularandeyepain,musculoskeletalsymptomsandsystemicbodysymptoms [220], headache [46, 79], dry throat, dry eyes, irritated skin, sniffles [46] andphysicalclumsiness[221];

Psychological comfort: tiredness, dullness, irritability [46], satisfaction [17, 222],depression[220],vitality,mentalhealth[79]andmood[222];

Subjective performance: self‐reported productivity [17, 79], trouble with memory [221],subjectivedaytimealertness[221,222],jobsatisfaction,jobstressandanxiety[220];

Sleepquality[46,79].

Takentogether,healthandwell‐beingassessmentsarelikelyrelatedto(subjective)physicalandmentalresponsestotheenvironment,includinglightingconditions.Fromthesestudies,severalassociations of health and well‐being with lighting conditions were found; higher correlatedcolourtemperature(CCT)improvedwell‐being[221].Outsideviewthroughwindowsenhancedwell‐beingwhichledtobettersleepquality[46].However,nostatisticallysignificantdifferencesofwell‐beingwerefoundataworkplacefordifferentilluminanceconditions(500lx,1500lxor


77

2500 lx) [222] and different light characteristics (dynamic vs. static lighting [79]/ direct orindirect[220])ordifferentlightingconditionsfordaytimeshifts.

Disorders

The lackof sufficient lightduringdaytime(e.g. in fall/winter)canstimulate theemergenceofseasonal affective disorders (SAD) in humans. Rosen et al found that the prevalence of SADdepended on geographical latitude, but not longitude. The global sensitivity score (GSS) wasratedhigherinincreasinglatitudenorthoftheequator[223].Lighttherapyhasbeenacceptedasatreatment forSADandnon‐seasonalpsychiatricdisorders(forareviewseeTermanetal.2010[224]).

4.1.3 Inter‐individualdifferences

Tooptimize lightingconditionsforNIFeffects, inter‐individualdifferencessuchasage,genderand chronotypes should also be taken into account. The term chronotype refers to subjectivepreferences of early or late (morning type or evening type) habitual sleep/wake time [225].Fromthesesubjectivepreferences,extrememorningandeveningtypesaswellasintermediatechrontoypes can be categorized, based on questionnaires [226]. Other inter‐individualdifferencessuchasgenderandagemayinteractwithacertainchronotype;duringlifespan,thechronotype can even change from a late to an early ‘bird’ as described by Roenneberg et al.[227].Thechangeinchronotypestartsattheageof10year‐old,whenchildrenbecomerathereveningtypesuntiltheyreachtheendoftheiradolescence.Withincreasingage,adultschangetowardsearlierchrontoypes[227].Interestingly,femaleshaveatendencytodevelopthisearlierthanmales:i.e.womenbecomethe‘latest’typesonaverageattheageof19.5years(men:20.9year‐old). By the end of the menopause in women, sex differences in chronotypes seem todisappear[227].

The effects of light on different genders and ages have been also reported. Melatoninsuppression was significantly lower in the elderly subjects following exposure to shortwavelength(456nm)light,whencomparedtotheyoungsubjects[228].Thismightreflectage‐related changes in lens density; however, the differences in gender are still unclear. Somestudiesfoundnosignificantgenderdifferenceinmelatoninsuppression[229,230],butothersshowed that females had greater melatonin suppression in response to light exposure thanmales[231,232].

4.2 LuminouspropertiesmodulatingNIFeffectsinhumans

This section describes the different fundamental characteristics of light, which influencedifferentNIFeffects.


78

4.2.1 Illuminance

Lightenteringtheeyeisthemainstimulusforcontrollingsynchronizationofcircadianrhythmswith theenvironmental24‐h light‐darkcycles [15,182,184,188,233]; it is thereforeevidentthat forNIFeffects,onehastomeasure illuminanceat thehumaneye inaverticalplane.Thisdiverges from standard measures, where illuminance is indicated in lux (lx) on a horizontalplane.Daytimebrightlightexposure(~1000to7000lxatthecornea)hasbeenshowntoreducesubjectivesleepiness[190,196], induceobjectivealertnessintheelectroencephalogram(EEG)[38],andimprovetaskperformance,whencomparedtodimlightconditions[196,197,234]orlowerilluminance(200lx)[235].Somestudiesinvestigatedalsotheeffectsofdaylightonnon‐visualfunctions.Halfanhourexposuretoabrightdaylightfluxnearbythewindows(1000lxtoover4000lx)hasbeenshowntobealmostaseffectiveasashortnapinreducingnormalpost‐lunchtimedrowsiness in healthy subjects [236]. Another field study showed that exposure tobrightlight(usingalightboxinducingapproximately2500lxattheeyelevel)enhancedmoodandvitalityinhealthyofficeworkersduringwintertimeinthenorthernhemisphere[212].

Nocturnal bright light can acutely increase subjective and objective alertness [192‐194, 211].Cajochenetal.[193]reportedthedose‐responserelationofalertnesstodifferentintensitiesofbrightlightexposures.Comparingalertingresponsestopolychromaticwhitelightof9100lx,theresult showed that 50%of themaximal alerting responsewas reachedwith less than 100 lxverticalilluminanceduringnighttime[193].Thesensitivitytolightshowsthattheilluminanceofambientlightingconditionsduringearlynightcanincreasealertness[193].Itwasalsoshownthat high nocturnal illuminance increased performance of complex cognitive tasks andsubjectivearousalofnight shiftworkers [215];however, therewasnodifferenceof luminousintensityonperformanceinsimplecognitivetasksormood[215].

4.2.2 Spectralcompositionoflight

In the early 2000’s, two independent research teams, Brainard et al. [118] and Thapan et al.[119], found similaractionspectrums forhormonalmelatoninsuppression inhumanswithapeakinthebluishpartoftheelectromagneticspectrumoflightat446‐477nm(Figure4.2);thispeak appears to be different from the photopic or scotopic response curves. Thiswork gavefurtherevidence for theexistenceofanovelnon‐rod,non‐conephotoreceptorpigment, called“melanopsin”. Its discovery has been very important for the lighting community since itprovided a deeper understanding of the importance of human circadian phototransduction.These findings also fostered novel approaches for light treatment of circadian and affectivedisordersinclinicalandnonclinicalsettings[237].BasedontheactionspectrumofBrainardetal.andThapanetal.[118,119],thecircadianactionfunctionwasdefinedlaterbyGalletal.andwastermedasC(λ)curve[238].RelatedcircadianactionfunctionsarementionedinSection3.1.


79

Figure 4.2: Spectral sensitivity curves formelatonin suppression in response to light in humans: greysquaresshowthespectralsensitivityformelatoninsuppressionbyBrainardetal.[118],blackcirclesshowthespectralsensitivitybyThapanetal.[119],lightbluelineshowscircadiansensitivitycurvesdevelopedbyGalletal.comparedwithsolidlineanddottedlineindicatespecificsensitivitycurvesofcones(V(λ))androds(V’(λ)); adaptedfromBrainardetal.[118]),Thapanetal.[119]andGalletal.[238]

Lockley et al. found a greater circadian phase‐shifting effect andmelatonin suppressionwithmonochromaticbluelight(460nm)thanwithgreenlightat555nminhumans.Exposuretothesamewavelengthsand luminous intensities resulted ingreatereffectsonsubjectivealertness,thermoregulation,heartrateandsleepwithbluethanwithgreenlight[12,239,240].

The studiesmentioned above demonstrate the effects ofmonochromatic light on NIF effects.Besidesluminousintensity,colourappearanceoflight,suchasthecorrelatedcolortemperature(CCT),thespectralcompositionoflightisalsoimportantforofficelighting.Higherilluminancestimulatedalertnessmoreeffectivelywith ‘daylight’ (5500K) thanwith ’warmwhite’ (3000K)polychromatic fluorescent lamps [241].Another studydidnotonly find stimulatingeffectsonalertness,butalsooncognitiveperformancebyapplyingcompact fluorescent lampsat6500Kduring two hours. These light sources induced greater melatonin suppression and enhancedsubjective alertness, well‐being and cognitive performance, when compared to 2500Kfluorescentlampsand3000Kincandescentlamps[242].Twoappliedofficestudiesshowedtheinfluencesofblue‐enrichedpolychromaticwhite light (17000K)onofficeworkers.Viola et al.[202]reportedhighersubjectivealertnessandperformance,lesssleepinessandeyediscomfort,when compared to polychromatic white light (4000K) during daytime (over several weeks)[202].Mills et al. [221] observed the effects fromblue enrichedpolychromaticwhite light oncaller‐handlers in the UK, and found an improvement of well‐being and work performance[221].


80

4.2.3 Timeofday

Several laboratory–basedstudieshavereported that timingof lightexposure iscrucial for thehumancircadianpacemaker[15,181,182,184]:ocularexposuretolightstimuliinthemorningcansignificantlyadvancethecircadianphase;ontheotherhand, lightstimulipresentedintheevening can delay the circadian phase. Figure 4.3 shows these circadian phase advances anddelaysonaphaseresponsecurvetolightstimuli(10'000lx);phaseadvances(withthepositivevalues), anddelays (withnegativevalues)wereplottedrelative to thecorebody temperatureminimum[183,243].

Figure 4.3: Phase response curve to 6.5‐hour of bright light with respect to the time of day; phaseadvances = positive values and delays = negative values; yellow bar = biological night; vertical solidline=corebodytemperatureminimum;thebest‐fit functiontothedataplots=solidcurve;theassumedaverage phase‐delay drift of the human circadian pacemaker = horizontal dashed line (adapted from[183]).

4.2.4 IssuesofNIFeffectsinarchitectureandbuildingphysics

Thediscoveryofnovelphotoreceptorcellshasalsoopenedthedoorforinnovationsin“healthylighting scenarios” and new approaches in architecture. A field study reported a time‐of‐dayeffect of light on hospitalized patients;. Benedetti et al. found that patients with bipolardepression had a shorter hospital stay in eastern rooms (exposed to direct sunlight in themorning), thanpatients inwesternrooms(exposed todirect sunlight in theafternoon) [244].Pechacek et al. [245] presented a simulated annualmodel of a room in a health care using acomputersimulation.Theysimulatedluminancereceivedatthepatient’seye(inbed),inordertoproposeanabsolutemeasureof circadianefficacy [245].Theannual simulationmodelwasperformed for different room orientations,window sizes and glazingmaterials and based ondaylight(luminousintensity,spectrumanddifferenttimes‐of‐day)receivedatthepatient’seye(inbed), inordertoproposeanabsolutemeasureofcircadianefficacy[245].Wirz‐Justiceand


81

Fournier [246] recently summarized some of the critical questions related to architecturaldesign,lightingandbiologicalaspectsinhumans[246].Somedraftguidelinesweresuggestedasfollowing:

Largerluminousintensityisneededforhomeandworkplacesatacertaintimesofday,differentchronotypesandagegroupshavingtobetakenintoaccount;

‘Intelligent’ and effective daylighting systems are required, since their higherilluminance, spectrum of light, and timing of light exposure can synchronize ourbiologicalclocktotheexternal24‐hhourlight/dark;

Optimizedartificiallightingsystems,suchasspectralcompositionanddynamiclightingsystems, are needed for schools, nursing homes, and offices in order to improve NIFfunctionsandbehaviorofoccupants.

During nighttime, dark conditions are needed in bedrooms: artificial light sources inpublicandprivatespacesshouldbeminimizedtoavoid‘lightpollution’[246].

4.3 Conclusion

Light influences both visual and non‐visual biological functions. Office lighting conditions canimpactmanyNIFeffectssuchasalertness,moodandsleepqualityinofficeworkers.NIFeffectsare therefore to be considered for health and well‐being, and they also contribute to workperformance. The light sensitive photoreceptors (ipRGC) modulate NIF effects; they show adifferentsensitivitymaximuminthevisiblelightspectrum,comparedtorodsandcones.Certaincharacteristicsoflight,suchasluminousintensityandspectralcomposition,whichareoptimalforNIF,maydifferfromthoseforthevisualsystem.Moreover,thetimingoflightexposurealsoinfluencestheNIFfunctions.

Itbecomesessentialtooptimizeofficelight‐darkconditionswithrespecttoNIFeffects;severalstrategiesinlightingpracticehavebeensuggestedtoimprovetheworkenvironment.Daylight,dynamic lighting and individual lighting control are recommended for health, well‐being andproductivity [176, 188, 247]; it also will be important to pay attention to inter‐individualdifferences. Currently, we are still in the early state regarding the optimization of lightingconditions; this is also true for NIF effects. Further research in the area of architecture andlightingwillbeneededinthefuturefor“healthylighting”atworkplaces.


82

83

Chapter5 Experimentalstudiesinofficelightingconditions

Most people spend their days at indoor working places where lighting conditions influencevisual andnon‐visual functions [17, 46, 134, 212, 236, 248]. These functions are time‐of‐day‐dependent, vary under different sky conditions and differ between individuals. Office lightingcan range from bright daylight to pure electric light in windowless office rooms. Twoexperimentalstudieswereset‐upunderrealisticday‐andartificialofficelightingconditionsintheLESOSolarExperimentalBuildingon theEPFLCampus (Lausanne, Switzerland).The firstexperiment aimed to investigate the effects of two different office lighting conditions in theafternoons on subjective alertness, mood, wellbeing and visual comfort in 25 young healthysubjects.Inthesecondexperiment,weinvestigatedthreedifferentlightingconditions,i.e.verydim,verybrightandself‐selectedlightingondifferentphysiologicalandsubjectivevariablesin32youngsubjectswhowereextrememorning‐orextremeeveningtypes.Thefollowingsectionsdescribethemethodsandresultsofthesetwostudies.

5.1 Impactofofficeday‐andelectriclightingconditionsonvisualcomfort,alertnessandmood(ExperimentI)

Several studies have investigated visual comfort of daylight in offices, focusing especially onglaresensations[49,148,152,249].Uptonow,onlyafewstudieshaveinvestigatedtheeffectsofdynamic(day‐)lightingconditionsinofficeroomsregardingdifferentvisualcomfortvariablesandnon‐image formingeffectsover time, andunderdifferent skyconditions.Theobjectiveofour experimentwas to test the acute effects of two different office lighting conditions in thecourseoftheafternoononvisualcomfort,alertness,moodandwell‐being.Thefullexperimentcomprised twoparts: 1) subjectshad to spend the first six hours in the afternoonunder twodifferentlightingconditions(daylightandpureelectriclight)and2),thetwofollowinghoursintheeveningswerespentunderdim lightconditions.Someof thepreliminaryresults fromthefirst study part were previously reported [24, 250]. Results from the second part includinghormonalresponsesandcognitiveperformanceunderdimlightintheeveningwerepublishedrecently [251]. In this section, the results of visual comfort, alertness, mood and well‐beingduringtheafternoonswillbepresented.

Experimentalstudiesinofficelightingconditions

84

5.1.1 Participantsandmethods

Studyparticipants

Healthy subjects between 20 and 30 years old without any medication or drug abuse wererecruitedforthisstudy.Noneofthemwasanextrememorningoreveningtype(asassessedbytheHorne‐Östbergquestionnaire).Subjects,whoperformednightshiftsorhadtravelledacrossmorethantwotimezoneswithin the last threemonths,werenotconsideredtoparticipate.Atotal of 29 subjectswere included in the study. Since the visual comfort scalewas givenonlyonce to the first 4 subjects, these subjectswere not included in the analysis. The data of 25subjectswas analysed: 16men;9women; age=23.5± 2.3 years (mean± standarddeviation;SD).Sevendayspriortothebeginningofthestudy,subjectswereaskedtomaintainaregularsleepdurationofapproximately8hourswithbed‐andwake‐timesataself‐selectedtargettimewithinarangeofapproximately30min.Compliancewascontrolledbymeansofawristactivitymonitor(Daqtix®Süttdorf,Germany)andsleepdiaries.Wake‐timewasonaverageat7:31AM±45min(mean±SD)andbedtimeonaverageat11:25PM±43min(mean±SD).Subjectswerealsoaskedtoconsumealcoholandcaffeinewithmoderationduringthesevendaysbeforethestudy,andtocompletelyabstainonstudydays.Allsubjectsgavetheirwritteninformedconsentduringaninterviewpriortothestudy.ThelocalethicalcommissioninLausanne(Switzerland)approvedthestudyprotocol. Allstudyprocedureswere inagreementwiththeDeclarationofHelsinki.

Roomsetup

The study was carried out in the meeting room of the Solar Energy and Building PhysicsLaboratory(LESO‐PB)ontheEPFLcampus.ThetwolightingconditionsareillustratedinFigure5.1,andsummarizedinTable5.1.

Figure5.1:Twolightingconditionsduringthestudy:a)(mainly)daylightingconditions;b)purelyelectriclightingconditions.


85

Code Lighting Targetvertical

illuminance

Lighting Verticalview conditions system fromwindows

EL Electriclighting 174lx Ceilinglight(4000K) Outsideviewwascompletelycoveredwithopaquecurtains

DL Daylighting 1000‐2000lx Daylight+ceilinglight Thelowerpartofthewindowswascoveredwithtranslucentblinds

Table5.1:Summaryofthetwolightingconditionsduringtheafternoons

Theelectric lighting(EL)conditionsweresetuptomimicawindowlessofficebycoveringtheopeningswithopaque curtainswhichprevented anydiffuse daylight and/or sunrays to entertheroom.Thehorizontalilluminanceatdesklevelinthemiddleoftheroomwassetto400lx(corresponding to174 lx in a verticalplane at theeye level).Thedaylighting (DL) conditionsweresetupinthesameroombyprovidingdaylightthroughAnidolicDaylightingSystems(ADS)locatedatupperpartofthewindows.TheADSre‐directsthedaylightfluxtotheceilingandtotherearoftheroom[252].Thelowerwindowpartswerecoveredwithtextileblindstopreventany outdoorwindow view. The vertical illuminance (Ev), defined as illuminance on a verticalplaneatthesubjects’eyes,wasaimedtobeintherangeof1000to2000lx;theelectriclightingwasswitchedonifEvwasbelow1000lx.Iftheverticalilluminanceexceededthistargetrangeand/ordirect sunlight entered the room, theupperblindswerepartially closed. Theverticalilluminance (Ev), theCorrelatedColourTemperature (CCT), theColourRendering Index (CRI)and the CircadianWeighted Irradiance (Eec) were continuouslymonitored in 5min intervalsthroughout the study. These variables were assessed using a calibrated spectroradiometer(Specbos1201,JETI,Germany),placednexttothesubjects.Thelightsensorofthespectrometerwas at the height of the subject's eye level. Series of snapshots of the scene, reproducing thesubjects’viewat12PM,3PMand5:30PM,weretakenduringtheexperiments.Thephotoswereprocessed by High Dynamic Range (HDR) procedures (see Chapter 3), to determine theluminousdistributionintheroom.ThispartwillbediscussedinmoredetailsinChapter6.

Studyprotocol

The participants spent two consecutive afternoons and early evenings in the test room. Theycametwicetothelaboratoryaroundnoon,oncefortheDLandoncefortheELcondition(inabalanced cross‐over design).During the afternoonhours, theywere allowed to read,work orlisten to music, but not to perform any computer tasks. Regularly (every 30‐60 minutes),subjectshadtoassesstheirvisualcomfort,subjectivealertness,moodandwell‐being.Atrainedassistantwaspresentinthetestroomtoensurecompliancewiththestudyprocedures.

Subjectiveassessments

Duringboth studydays, subjectshad toassessvisual comfort, subjective alertness,moodandwell‐being on paper‐based questionnaires by theway of a visual analogue scale. They had toindicate their current status between two extremes (for example: extremely sleepy andextremelyalert)onahorizontalline(0‐100mm)[253].


86

Visualcomfortratings

Twoquestionnaireswereused forvisualcomfortassessments. In thecourseof theafternoon,subjects completedeachhour seven itemsof theVisualComfort Scale (VCS; seeAppendixA).TheitemswereextractedfromthefullOfficeLightingSurvey(OLS)questionnaire[80,254].Thesevenitemswheremodifiedtobeusedonacontinuousvisualanaloguescale.Oncetowardstheendoftheafternoon,subjectshadto fill inamodifiedOLSversion[250](10items; fordetailsseeTable5.3)oncontinuousvisualanaloguescales.

Non‐visualfunctionsassessments

In the course of the afternoon, subjects assessed every 30minutes their subjective alertness,mood,physicalwell‐beingandrelaxationoncontinuousvisualanaloguescales(seeAppendixA).

Statistics

The software Statistica 6.12 (StatSoft, USA)was used for statistical analyses. The non‐visualfunction assessments were collapsed in hourly bins. Results from subjective ratings wereanalysed by performing repeated analyses of variance (rANOVA) with the factors ‘time’,‘condition’(DL,EL),‘gender’and‘order’(beginwithELorDL).Missingdatafromthesubjectivequestionnaireswere linearly interpolated. For post‐hoc analysis, the Duncan’smultiple rangetest and t‐testswere applied. For correlation analysis, the Pearson’s correlationwas used. Inorder to analyse significant changes in the course of the afternoon, subjective assessments ateachtimepointwerecomparedtothebeginningofthestudybyusingt‐testswithadjustedp‐values formultiplecomparisons.Forp‐valueadjustments,theFalseDiscoveryRate(FDR)wasused[255].

5.1.2 Results

Photometricvariables

Theverticalilluminance(Ev)duringtheDLconditionwasonaverage965.8±94.5lxand174.3±1.0lxfortheELcondition(mean±standarderrorofmean;SEM).Allphotometricmeasures(Ev,CCT,CRIandEec)aresummarizedinTable5.2forbothlightingconditions.TheELconditionwassteady‐state,whereastheDLconditionwasdynamicsuchthatEv,CCT,CRIaswellasEecvariedovertime,asshowninFigure5.2.


87

Light Unit ElectricalLight(EL) Daylight(DL)

Properties n=25 n=25

(mean±SEM) (mean±SEM)

Ev Lx 174.3±1.0 965.8±94.5 CCT K 3707.0±5.8 4284.2±44.3 CRI ‐ 83.4±0.03 89.9±0.6

Eec W/m2 0.13±0.001 1.9±1.0 Table5.2:Averagedverticalilluminance(Ev;lx);CorrelatedColorTemperature(CCT;K),ColorRenderingIndex(CRI;nounits)andcircadianweightedirradiance(Eec;W/m2)duringDLandEL(meanofallhours±SEM).

Figure 5.2: Averaged photometric variables in the course of the afternoon (mean ± SEM) a) Verticalilluminance(Ev)onalogscale;b)Circadianweightedirradiance(Eec);c)CorrelatedColourTemperature(CCT);d)ColourRendering Index(CRI);DLcondition=bluecirclesandELcondition=orangecircles;*indicatesthedifferencesbetweenDLandEL;p<0.05.

The DL conditions between 12PM and 4PM showed significantly higher Ev, CCT, CRI and Eecvalues,when compared to the EL conditions (t‐test; p<0.01; Figure 5.2). Since the studywasconducted during fall and winter, all photometric variables under DL conditions weresignificantlyhigherbetween12PMand4PM,whencomparedtothoseat6PM(t‐test;p<0.01),becauseofthefadingofdaylight.

12PM 1PM 2PM 3PM 4PM 5PM 6PM

E v(lx)

100

500

1000

2000

DL(n=25)EL(n=25)


CCT(K)

3500

3750

4000

4250

4500

4750

Time


CRI[‐]

80

85

90

100

Time


E ec(W/m

2)

0.1

0.25

0.5

1

1.75

a)Ev c)CCT

d)CRIb)Eec

* * * * * *

* * * * * *

* * * * * **

* * * * * **


88

Visualcomfortvariables

Subjective assessments of visual comfort showeda significantlybetter visual comfort, greaterlightpreference,lowerglaresensationandalargerlightingsufficiencyforworkunderDLthanunder EL conditions (4‐way rANOVA; main effect of ‘condition’; p<0.05; Figure 5.3). Allassessments for light preference, visual comfort and subjective glare ratingswerewithin therange of satisfaction (> 50), for both lighting conditions. Light preference and visual comfortratings became lower in the course of the afternoon for both lighting conditions (Duncan’smultiple range test; p<0.05). Subjective glare ratings were significantly higher under ELcomparedtoDLafter2PM,asshowninFigure5.4(2‐wayrANOVA;‘timexcondition’;Duncan’smultiplerangetest;p<0.05),indicatinghigherglareperceptionunderEL.

Figure5.3:Averagedsubjectiveassessmentsofa)visualcomfort;b)lightpreference;c)subjectiveglareratings;andd)sufficiencyforworkrating;DL=bluebarsandEL=orangebars;*=p<0.05(mean±SEM;n=25;0=notsatisfiedatall;100=extremelysatisfied).

a)VisualComfort b)LightPreference c)Sufficientforwork c)SubjectiveGlare50

60

70

80

90

100DL(n=25)EL(n=25)

***

*

MoreSatisfied/imperceptible

LessSatisfied/intolerable

VisualCom

fortScale


89

Figure 5.4: Averaged subjective glare assessments per hour; DL = blue circles and EL = orange circles(mean±SEM);*=significantdifferencesbetweenDLandEL;n=25.

For themodified version of the entireOLS questionnaire (given once towards the end of theafternoon;Table5.3).The luminousdistributionwas ratedasbeing significantlybetterunderDLthanunderELconditions: thesubjectsreportedhaving lessreflections, light flickeringandshadows(p<0.05).Subjectsalsoratedthe lightappearancetobe ‘colder’underELthanunderDLconditions(p<0.05).TheyfoundthatthelightingundertheDLconditionswasbetterthanintheirexistingoffices,andtheyestimatedtobeabletoworkforalongertimeperiodunderDLthanunderELconditions(p<0.01).ThelatterwasalsodependentonwhetherthesubjectshadDLon their first or on their second studyday (main effect of ‘order’; p<0.05).When subjectsstartedtheirtestsessionsunderDLconditions,theyfoundthelightingunderELconditionslessevenlydistributedandjudgedthattheywouldspendlesstimeinthethatlightingenvironmentthanthosewhostartedtheirsessionwithEL(maineffectof‘order’;p<0.05).

ClockTime(h)

12PM 1PM 2PM 3PM 4PM 5PM

Glare

20

30

40

50

60

70

DL(n=25)EL(n=25)In

tolerable

Imperceptible * * * *


90

Non‐repeatedquestionnaire Maineffectof

(0=Yes/100=No) condition order

DL EL Value DL/EL EL/DL Value

Mean SE Mean SE p Mean SE Mean SE p1.Thelightingispoorlydistributed 90.4 2.2 77.2 5.4 ** 74.9 5.4 92.0 2.0 * 2.Thelightingcausesdeepshadows 83.7 3.8 73.2 6.0 * 71.4 5.8 85.0 4.1 ‐ 3.Reflectionsfromthelightfixtureshindermywork 77.9 5.4 64.7 6.5 ** 72.8 5.3 69.8 6.7 ‐ 4.Theceilinglightfixturesaretoobright 79.9 3.8 75.2 4.7 ‐ 79.9 3.0 75.4 5.0 ‐ 5.Myskinhasanunnaturaltoneunderthelighting

71.0 5.9 68.8 5.2 ‐ 63.5 6.1 75.9 4.8‐

6.Thelightingflickersthroughouttheday 88.1 3.6 78.0 5.0 ** 78.6 5.1 87.2 3.7 ‐ 7.Thecolourofthelightingistoo“warm” 88.1 2.3 90.4 1.8 ‐ 87.7 2.3 90.7 1.8 ‐ ‐ 8.Thecolourofthelightingistoo“cold” 67.9 5.2 46.6 6.3 ** 50.1 6.0 63.9 6.0 ‐ 9.Howdoesthelightingofthisoffice 2.4 0.2 3.2 0.2 ** 3.0 0.2 2.6 0.2 ‐comparewiththelightingofotheroffices youhaveworkedin?(1=better…5=worse) 10.Foraworkingday,IimaginethatIcan 3.4 0.2 2.5 0.2 ** 2.6 0.2 3.3 0.2 *workinthislightenvironmentfor..hrs

Table 5.3: Averaged ratings for the Office Lighting Survey (OLS) questionnaire under both lightingconditions(DLandEL);meanandstandarderrorsofmean(SEM);*=p<0.05,**=p<0.01;(3‐wayrANOVA:‘condition’,‘order’and’gender’;n=25).

Non‐visualfunctions

Therewas no significant difference between the two lighting conditions regarding non‐visualfunctions. For both lighting conditions, a significant change over time was observed forsubjective alertness,mood, physical wellbeing and relaxation (4‐way rANOVA;main effect of‘time’;p<0.05).Theyallbecameworse towards theendof theafternoon.Withineach lightingcondition, the relative time course of those subjective assessments was further analysed bycomparingtheirchangestothevaluesatnoon.Afterthefirsthour,subjectsreportedincreasedsleepinessandlessphysicalwell‐beingunderELconditions,whencomparedtothebeginningofthestudy(Figure5.5;t‐testwithadjustedp‐valuesbyFDR;p<0.023).UndertheDLconditions,physical well‐being did not significantly change in the course of the afternoon (Figure 5.5a;p>0.05).AnearlierdecreaseofalertnessduringELthanunderDLconditions(Figure5.5b;t‐testwithadjustedp‐valuesbyFDR;p<0.018)wasobserved.


91

Figure5.5:Changesinphysicalwell‐being(upperFigure)andsubjectivealertness(lowerFigure)duringtheafternoon,whencomparedtothebeginningofthestudy(mean±SEM,n=25);DL=bluecirclesandEL=orangecircles;thehorizontaldashedlineindicatesthevalueatthebeginningofthestudy;*=significantchangessincethebeginningofthestudy(12PM).

Correlationanalysis

Inordertoinvestigatetherelationshipbetweenthephysicallightpropertiesandtheresponsesfromsubjects,severalphotometricvariableswerecorrelatedwithsubjectiveassessments.Theinter‐correlations of different subjective assessments were also analysed. Only statisticalsignificantcorrelationsareshowninTable5.4(Pearson’scorrelation;p<0.05).

a)PhysicalWell‐being

12 PM 1 PM 2 PM 3 PM 4 PM 5 PM

LessCom

fortMoreComfort

-20

-15

-10

-5

0

DL(n=25)EL(n=25)NoonasReference

b)Alertness

ClockTime(h)

12PM 1PM 2PM 3PM 4PM 5PM

LessAlertMoreAlert

-20

-15

-10

-5

0

DL*

EL*

EL*


92

Correlationbetweenphysicallightpropertiesandsubjectiveassessments

Table 5.4a) presents the correlation coefficients (r) between somephotometric variables andvisual comfort variables, including the non‐visual functions. Positive correlation coefficientssuggest significant intra‐individual associationswith greaterphotometric values (Ev, CCT, CRIandEec)andbettervisualcomfortratings,aswellasgreaterlightpreference(r>0.13;p<0.05).Higher CCT and CRIwere associatedwith better physical well‐being and lower glare ratings(r>0.12;p<0.05). Sunshinedurationwasalsocorrelatedwiththesevariablessuchthat longerdaily sunshine hours were related to greater light preference, visual comfort and morerelaxation(r>0.12;p<0.05).

Positivelycorrelationcoefficient

Higherlight

preference

Highervisualcomfort

Moresufficienttowork

Moresubjectiveglare

Morerelaxation

Betterphys.well‐being

Morealert

Bettermood

HigherEv 0.16 0.13 ‐ ‐ ‐ ‐ ‐ ‐

HigherCCT 0.34 0.32 ‐ ‐0.19 ‐ 0.13 ‐ ‐

HigherCRI 0.33 0.31 ‐ ‐0.20 ‐ 0.12 ‐ ‐

HigherEec 0.16 0.13 ‐ ‐ ‐ ‐ ‐ ‐

Longersunshineduration

0.21 0.18 ‐ ‐ 0.13 ‐ ‐ ‐

b)Visualcomfortvariables Higherpreference x 0.90 0.42 ‐ ‐ 0.19 0.28 0.24

Highervisualcomfort

0.90 x 0.35 ‐ ‐ 0.18 0.33 0.19

Moresufficienttowork

0.42 0.35 x 0.17 0.16 0.11 0.16 ‐

Moresubjectiveglare

0.12 ‐ 0.17 x 0.17 0.16 ‐ 0.24

c)Non‐visualfunctions Morerelaxed ‐ ‐ 0.16 0.17 x 0.51 0.18 0.44

Greaterphysicalwell‐being

0.19 0.18 0.11 0.16 0.51 x 0.50 0.56

Morealert 0.28 0.33 0.16 ‐ 0.18 0.50 x 0.47

Bettermood 0.24 0.19 ‐ 0.24 0.44 0.56 0.47 x

Table5.4:Correlationcoefficients(Pearson’scorrelation;r)betweenvisualcomfortvariables,non‐visualfunctions and light properties; only significant correlations (p<0.05) are shown (‐ = p>0.05). Positivecorrelationcoefficientsindicatepositivecorrelationsofhighervisualcomfortvariables,morerelaxation,betterphysicalwell‐being,higheralertnessandbettermoodwithhigherEv,CCT,CRIandlongersunshineduration.

Inter‐correlationsofsubjectiveassessments

Inter‐correlationsbetweenvisualcomfortvariablesandnon‐visualfunctionsaresummarizedinTable5.4b)andc).Greater lightpreferenceswerecorrelatedwithabettervisualcomfort(r=0.90,p<0.001);bothvariableswerealsopositivelycorrelatedwithgreatersatisfactionwiththelighting forwork(r>0.35;p<0.001).Small,butsignificantcorrelationsbetweenvisualcomfortvariablesandnon‐visualfunctionswerealsofound;greatervisualcomfortandlightpreferencewereassociatedwithphysicalwell‐being, greateralertnessandbettermood (r>0.18,p<0.05).


93

Greaterphysicalwell‐beingwasalsoassociatedwithmorerelaxation(r=0.51,p<0.001).Bothwere also related tobettermood (r >0.44,p<0.001) andgreater alertness (r>0.18,p<0.001).Morealertsubjectswereassociatedwithbettermood(r=0.47,p<0.001).

5.1.3 Discussion

TheresultsshowedgreaterlightpreferenceandbettervisualcomfortunderDLthanunderELconditions,whichisinagreementwiththecurrentliterature[70,71,256,257].Similarly,glaresensationwasbettertoleratedunderDLthanunderELconditions,asreportedbefore[23,45,258], although the vertical illuminances were larger under DL conditions. A few studiessuggested that a pleasant view out of the windows may increase visual comfort and inducelower glare ratings [46, 49]. In our study, any outside view through vertical windows wasprevented;hence,viewwasnotthereasonforbettervisualcomfortratingswithlessglareunderDLthanunderELconditionsinthiscase.Itmightnotonlybetheview,butalsothephotometricproperties of daylight per se that enhanced visual comfort. We indeed found that higherilluminance,CCT,CRIandEecaswellaslongersunshinedurationswerecorrelatedwithhigherlightpreferenceandbettervisualcomfort.

Light preference and visual comfort decreased over time, as also shown in a previous report[17],butsubjectiveglareratingsremainedconstantforbothlightingconditionsinthecourseofthe afternoon. This implied that lower visual comfort was not always induced by higherdiscomfortglareratings.

Non‐visualfunctions,suchasrelaxation,physicalwell‐being,subjectivealertnessandmooddonot show any significant differences between the two lighting conditions. However, physicalwell‐being and subjective alertness decreased significantly earlier under EL than under DLconditions.This is inagreementwithaprevious report, showing that sleepinesswas reducedafterexposuretobrightdaylightnearbywindowsforhalfanhourintheafternoon[236].Otherreports showed that higher light intensity during daytime led to greater alertness and thusservedasacountermeasuretoanincreaseofsleepinessduringtheday[190,196].Conversely,moodassessmentsdecreasedovertimeforbothlightingconditions,buttherewasnosignificantdifferencebetweenthem.Therewasnocorrelationbetweenanyphotometricpropertiesoflightand either mood or alertness. The lack of relationship between the two might also be aconsequenceofthechosenilluminanceconstraintsoftheDLcondition(i.e.tobebetween1000and2000lx).

ItmustbenotedthatunderDLconditionsmostofthedisabilityglarewaspreventedbytheuseoftextileblinds.ThatmightalsobethereasonofthelowersubjectiveglareratingsfoundunderDLconditionsandneedstobeconsideredinarealofficesetting.Furtherstudiesarerequiredtobetterunderstand the impactofoffice lightingby thewayofa largersamplesizeandvariouslightsettings.Itisalsonecessarytoinvestigatelongertime‐dependentlighteffectsduringday‐time.

Oursecondexperiment focusedon inter‐individualdifferences.We includedextrememorningandeveningtypestoassesspotentiallydifferentsubjectivelight‘needs’duringworkinghours.


94

5.2 Impactofdifferentlightingconditionsonvisualcomfort,discomfortglare,alertnessmoodandwell‐beinginextremechronotypes(ExperimentII)

As described earlier in this thesis (Section 4.1), light influences several non‐image formingeffects,suchastheentrainmentoftheinternalcircadianclocktotheexternal24hoursofaday.There are large inter‐individual differences in habitual daily light exposure times among thegeneralpopulation.Thesedifferencesarealsoknownfromdifferentchronotypes(i.e.morningand evening types). Extreme chronotypes are healthy persons with strong subjectivepreferencestogetupveryearlyandtogotobedveryearly(morningtypes),ortogetupverylate and to go to bed very late (evening types) [225]. Roenneberg et al. [225] reported theassociationsoflongerexposuretobrightlightinthemorningwithmoreadvancedsleepperiodsin humans under daily life situations. Their results suggest that in addition to genetic andbehavioural factors, daily light exposures might consolidate the existence of differentchronotypes [225]. Morning types showed a longer and more intense habitual exposure tobright light during the day than evening types [225, 259, 260]. The greater daytime lightexposurewentalongwithearliercircadianphases.Conversely, theshorterhabitualexposuresto bright light observed in evening typesmay contribute to their later circadianphases [259,261].Asignificantpositivecorrelationwasfoundbetweennocturnallightlevels(measuredbyquestionnaires) in adolescents’who are later chronotypes [262]. Vollmer et al. [262] showedthat the use of electronic screen media (computer screen or television) at night time is aninfluential determinant of eveningness. Adolescents living in bright illuminated urban areasshowed a stronger eveningness preference than those who are living in darker ruralenvironments[262].

Since light intensity and time‐of‐day‐dependent light exposure modulate circadian phases ofphysiology and behaviour, extreme morning and evening types are known to be exposed tonatural light at different internal circadian phases. Thismay lead to different light effects onvisualandnon‐visualfunctions.Thesecondexperimentwasaimingtotesttheimpactoflightonsubjective and objective visual and non‐visual functions, including cognitive performance inyoungadults,whoareextremechronotypes.Someofthe(preliminary)resultswerepreviouslypresentedatconferences[263‐266].Someofthedataanalyses,suchascognitiveperformanceandhormonalanalysesarestillongoing.Inthissection,theimpactoflightonvisualcomfortandsubjectivealertness,moodandwell‐beinginextrememorningandeveningtypesarepresented.

5.2.1 Subjectsandmethods

Studyparticipants

Extrememorningtypes(MT)andeveningtypes(ET)wererecruitedbyflyersattheUniversityof Lausanne and EPFL (Switzerland). They were selected based on their extreme subjectivesleep‐wake preferences, as identified by two validated questionnaires: the Horne‐Östberg


95

questionnaire [226] and the Munich chronotype questionnaire (MCTQ,[225]). Only healthysubjectsbeingeitherextremeMTorETbetween20and30yearsoldwereincludedinthestudy.Except for the chronotype criteria, the inclusion criteria were the same as in Experiment I(Section5.1.1).Atotalof16MTand16ETsubjectscompletedthestudy(14men;18women;age=22.7±3.5years;mean±SD).Sevendayspriortothebeginningofthelatter,subjectswereaskedtomaintainaregularsleep‐wakecycleofapproximatelyeighthourseachataself‐selectedhabitualbedandwake‐timewithinarangeofapproximately30min.Compliancewascheckedbymeansofawristactivitymonitor(Daqtix®Süttdorf,Germany)andsleeplogs.Subjectswerealsoaskedtoonlymoderatelyconsumealcoholandcaffeineduringsevendaysbeforethestudy,andtocompletelyabstainonstudydays.Allsubjectsgavetheirwritteninformedconsentduringapriorinterview.ThestudyprotocolwasapprovedbythelocalethicalcommissioninLausanne(Switzerland).AllprocedureswereinagreementwiththeDeclarationofHelsinki.

Studyprotocol

Thesubjectshadtoparticipateinthreestudysessions.Eachofthemwasscheduledtobeginonehourafterthehabitualwake‐timeandlastedfor16hours.Thestudytimesandcorrespondingwake‐time/bedtimearesummarizedonTable5.5;onlyonesubjectwaspresent ineachstudysession.Duringthestudysessions,subjectsremainedseatedinthetestroom;theywereallowedto read,work or listen tomusic (including onehour of scheduled computerwork). A trainedassistantstayedwiththemthroughoutthestudy.

Subjects MT ET

Clocktime Clocktime

(mean±SD) (mean±SD)

Wake‐time 06:19±0:36 10:18±0:15

Bedtime 22:13±0:42 02:10±0:57

Studybegin 07:15±0:34 11:14±1:01

Studyend 23:15±0:34 03:14±1:01

Table 5.5: Averaged habitual wake‐ and bedtimes, study begin and end times (mean± SD) of extrememorningtypes(MT;N=16)andextremeeveningtypes(ET;N=16).

Roomset‐upandlightingconditions

ThestudywasperformedinthetestroomoftheSolarEnergyandBuildingPhysicsLaboratoryat EPFL, as described in Section 5.1.1. For the three sessions, all subjects underwent threedifferentlightingconditions:

I. Dimlight(DIM)

II. Constantbrightlight(BL)

III. Self‐selectedlighting(SSL)dependingonthesubject’schoice

Topreventanyoutsideview,alllightingconditionsdidnotallowaviewoutthroughthelowerwindowspart.ThecriteriaforthethreelightingconditionsaresummarizedonTable5.6.


96

Lighting Vertical Daylight Artificial VerticalviewCode conditions illuminance lighting throughwindows

DIM Dimlight <5lx no yes no

BL Brightlight ~1000lx yes yes no

SSL Self‐selectedlight Dependingon available available no

subjects'choices

Table 5.6: Criteria for the three different lighting conditions (DIM, BL and SSL). Set‐points of verticalilluminancevaluesweretargetedatthesubjects’eyelevel.

For DIM and BL conditions, vertical illuminance was below 5 lx and at approx. 1000 lx,respectively.Thestudyassistantwasresponsibleforthelightingcontrolthroughoutthesetwolightingconditions.FortheBLcondition,therewasmainlydaylight,whichwascomplementedwithelectricallight,iftheilluminancethresholddroppedbelow1000lx.DuringSSLconditions,thesubjectswereaskedhourlytochoosebetweendaylightingand/orelectricallighting;thesechoices,illustratedinFigure5.6,comprised:

1) Daylightingconditions:

ThedaylightfluxenteredtheroomviawindowsequippedwithAnidolicDaylightingSystems(as described in Section 5.1.1), located at the upper part of the windows. The façade wasequippedwith:

a) External blinds: translucent blinds located in front of the lower verticalwindowwere fully lowered in order toprevent any outside view; theupperparts of theblindswerecontrollablebythesubjects.

b) Internalcurtain:opaquecurtainswereavailable tocloseallwindows inorder touseelectricallightonly(subjectswerenotallowedtobeincompletedarkness).

c) Internal californian blinds: movable blinds, installed on the upper part of thewindow could be used to control the daylight flux (in front of the AnidolicDaylightingSystem).

2) Artificiallightingconditions:

a) Ceilingmountedluminaries(fluorescenttubes;CCT4000K,Ev16‐1280lx)

b) Indirectlightstandinglamp(fluorescenttubes;CCT3000K;Ev1.1‐52.5lx)

c) Desklamp(incandescentlightsource;Ev95‐1425lx)


97

Figure5.6:Availableday‐andelectric lightingsystemstoregulatethe lighting intheSSLcondition:1a)external blinds; 1b) internal curtains; 1c) Californian blinds; 2a) ceiling light; 2b) indirect light from astandinglamp;2c)desklamp.

Vertical illuminance at the subject eyes’ level was continuously monitored during the studysessions using a spectroradiometer (Specbos 1201, JETI, Jena, Germany). The study alwaysstartedwiththeDIMcondition,followedbyeithertheBLorSSLconditionsinabalancedcross‐over order for both chronotypes. Visual comfort variables were assessed hourly; subjectivealertness,mood,physicalwellbeingandrelaxationwereassessedevery30minutes(seebelow).Imagesofthescenemimickingthesubjects’viewwererecordedhourlyduringthestudyunderBLandSSLconditionsusingtheCamera‐LikeLightSensor(CLLS),asdescribedinChapter3.Thehourlyluminousdistributionofthesceneswasdetermined;theyarediscussedinChapter6.

Subjectiveassessments

During the three study sessions, subjects had to assess visual comfort, subjective alertness,moodandwell‐beingonpaper‐basedquestionnairesusingavisualanaloguescale.Thisrequiredtomark their choice between two extremes (e.g. extremely sleepy and extremely alert) on ahorizontalline(0‐100mm)[253].


This study had slightly different visual comfort assessments, when compared to the firstexperiment(Section5.1.1).Inthecurrentexperiment,subjectsassessedeachhour11itemsonthe Visual Comfort Scale (VCS), which were extracted from the larger Office Lighting Survey(OLS) [80, 254]. The eleven itemsweremodified in order to be used on a continuous visualanaloguescale(seeAppendixAfordetails).

Non‐visualfunctionsassessments

Asinthefirstexperiment(reportedinSection5.1.1),subjectsexpressedevery30minutestheirsubjective alertness,mood, physicalwell‐being and relaxationon a visual analogue scale (seeAppendixA).


98

Statisticalanalysis

TheStatistica6.12 (StatSoft,USA)softwarewasused tocarryout thestatisticalanalyses.Thesubjective assessmentswere averaged per hour.Missing data from subjective questionnaireswere linearly interpolated. Data from subjective ratings were analyzed first by performingrepeatedanalysesofvariance(rANOVA)withthefactors:‘time’(hours1‐16afterhabitualwake‐time),‘condition’(DIM,BLandSSL),‘chronotypes’(MT,ET),‘gender’and‘order’(beginwithBLorSSL).Inanextstep,subjectiveassessmentswereanalyzedseparatelyforeachcondition(BLor SSL) in order to investigate the effects of different sky patterns, such as overcast,intermediate and clear sky, at different times (2‐way rANOVA; ‘weather’ x ‘time’; fordetermination of the sky conditions see Chapter 5.2.2). For post‐hoc analysis, the Duncan’smultiplerangetestwasappliedandPearson’s’correlationwasusedforthecorrelationanalysis.

5.2.2 Results

Photometricvariables

Thevertical illuminance (Ev),CCT,CRIaswellas thecircadianweighted irradiance(Eec)weremonitoredduring16hoursunderthethreedifferentlightingconditions(DIM,BLandSSL).Thevertical illuminanceunderBLconditionswasconstantat1031±64 lux (mean±SD),while itvariedunderSSL(714±958;mean±SD).TheaveragedphotometricvariablesunderdifferentlightingconditionsaresummarizedforallsubjectsinTable5.7.

Lighting Ev CCT CRI Eecconditions (lx) (K) (‐) (W/m2)

(mean ±SD)

(mean ±SD)

(mean ±SD) (mean±SD)

DIM 3.07±0.4 3115±270 85.8±1.0 0.002±0.0004 BL 1031±64 4120±535 85.7±6.1 0.83±0.18 SSL 714±958 4332±989 90.0±6.2 0.67±0.89

Table5.7:Averagedverticalilluminance(Ev),CCT,CRI,andEec(mean±SD)forDIM,BL,SSLconditions(N=32).

Sincedaylightisdynamic,thelightpropertieschangedovertimeunderSSLandBLconditions(Figure5.7).Verticalilluminance(Ev)andEecwereonaveragesignificantlyhigherunderBLthanunderSSLconditions(averagefrom16hours).TheCCTandCRIweresignificantlyhigherunderSSLthanunderBLconditions(2‐wayrANOVA;p<0.05).

Inorder to investigate thedifferencesof lightingenvironmentsbetween the twochronotypes,BL andSSL conditionswere analysed separately. Therewere significant interactionsbetweenthe factors ‘time’ and ‘chronotype’ for all photometric variables, except for illuminanceundertheBLcondition(2‐wayrANOVA;p<0.05Figure5.7).UnderBLconditions,EThadhigherCCT,


99

CRI and EecthanMT in the first 2‐3 hours after habitualwake time. These valueswere lowerfrom5to11hoursinETthanMT(Figure5.7c,e,g).ThedifferencesbetweenchronotypesweresimilarunderSSLconditionsexceptfortheCCT,wherethefirst2‐3hourswerenotstatisticallydifferentbetweenMTandET.(Figure5.7b,d,f,h).

For further analysis of SSL andBL conditions, theweather statuswas characterized by threedifferentskyconditions(clear,intermediateandovercastsky),accordingtotheSwissNorm150911 [267]. Themeteorological datawere issued from the localweather station (Meteosuisse,Pully,VD,Switzerland)[268].Aclearskyreferstoa0%‐25%ofcloudcoveringduringthedaysof the study; an overcast sky reflects to 75% ‐100% of cloud covering. A sky with a cloudcoveringbetween25%and75%isconsideredasanintermediatesky.TheaveragedEv,CCT,CRIandEecvaluesfordifferentskyconditionsareshowninTable5.8.ForBL,theCCT,CRIandEec,where significantly lower for overcast skies,when compared to theother two sky conditions.ForSSLconditions,therewashigherEvandEecforclearskiesthanforovercastskies,andhigherCCTandEecforintermediateskiesthanforovercastskies(Table5.8).

Lighting Sky N Ev CCT CRI Eec

conditions conditions (lx) (K) (‐) (W/m2)

(mean±SD) (mean±SD) (mean±SD) (mean±SD)

BL Overcast 10 1039±56 3879±191+# 83.1±3.3+# 0.76±0.09+#

Intermediate 11 1039±90 4177±542 86.3±6.2 0.86±0.2

Clear 11 1020±64 4276±652 87.4±7.2 0.87±0.2

SSL Overcast 13 457±303# 4064±765+ 88.9±5.5 0.39±0.31+#

Intermediate 6 771±618 4812±1166 90.9±7.2 0.88±1.4

Clear 13 934±1349 4377±1011 90.6±6.1 0.89±1.2

Table 5.8: Averaged vertical illuminance (Ev), CCT, CRI, and Eec (mean ± SD) under DIM, BL, SSLconditionswithdifferentskyconditions(overcast,intermediateandclearsky).N=numberofsubjectsintherespectivecondition;+=significantdifferencesbetweenintermediateandovercastsky(p<0.05);#=significantdifferencebetweenclearandovercastsky(p<0.05).


100

Figure5.7:TimecoursesofaveragedphotometricvariablesforBL(leftcolumn)andSSL(rightcolumn);a‐b:illuminanceonlogscales(Ev);c‐d:ColourCorrelatedTemperature(CCT);e‐f:ColourRenderingIndex(CRI);g‐h:circadianweightedirradiance(Eec);MT(n=16):solidlinesandET(n=16):dashedlines(mean±SEM);*=significantdifferencesbetweenMTandET(p<0.05).Elapsedtimeonthex‐axiswasalignedrelativetohabitualwaketimeforbothchronotypes.

a)EvforBL

Hours since Waketime (hrs)1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

E v(lx)

0

1000

2000

3000 MTET

b)EvforSSL


E v(lx)

0

1000

2000

3000MTET

c)CCTforBL

Hours since Waketime (hrs)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

CCT(K)

3000

3500

4000

4500

5000

5500MTET

d)CCTforSSL


CCT(K)

3000

3500

4000

4500

5000

5500

MTET

e)CRIforBL

HourssinceWaketime(hrs)1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

CRI(‐)

75

80

85

90

95

100

MTET

f)CRIforSSL


CRI(‐)

75

80

85

90

95

100

MTET

g)EecforBL

HourssinceWaketime(hrs)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

E ec(W/m

2 )

0.0

0.5

1.0

1.5

2.0 MTET

h)EecforSSL


E ec(W/m

2 )

0.0

0.5

1.0

1.5

2.0 MTET

** * *

* * ** * *

**

*

* * ** * * *

**

* * * * * * **

*

* **

* * * * **

** * *

* *


101

In order to investigate the dynamics of light under different sky conditions, the photometricvariableswereanalysedacross16hoursforallsubjects(n=32).Significantinteractionsbetweenthe factors ‘time’ and ‘sky conditions’ were found under BL and SSL conditions for mostphotometricvariables,exceptCCTandCRIunderSSLconditions(Figure5.8).Post‐hocanalysesrevealedsignificantdifferencesbetweenthethreeskyconditions,predominantlyduringthefirst8hoursofthestudy,whendaylightwasavailable(Figure5.8b‐c,g‐h;2‐wayrANOVA;Duncan’smultiplerangetest:p<0.05).UnderSSLconditions,CCTandCRIweresimilarforthethreeskytypes(Figure5.8d‐e).


102

Figure5.8:Time coursesunderBLandSSL conditions forphysical lightpropertiesunderdifferent skyconditions for all study sessions (BL: n=32; SSL: n=32); a‐b: illuminance (Ev); c‐d: Colour CorrelatedTemperature(CCT);e‐f:ColourRenderingIndex(CRI);g‐h:circadianweightedirradiance(Eec);overcastsky=dashed‐dottedlines,intermediatesky=dashedlinesandclearsky=solidlines(mean±SEM);+=significant differences between intermediate and overcast sky (p<0.05); # = significant differencesbetweenclearandovercastsky(p<0.05);*=significantdifferencesbetweenclearandintermediatesky(p<0.05).

a)EvforBL


E v(lx)

0

1000

2000

3000

OvercastskyIntermediateskyClearsky

b)EvforSSL


E v(lx)

0

1000

2000

3000


d)CCTforSSL


CCT(K)

3000

4000

5000

6000


c)CCTforBL


CCT(K)

3000

4000

5000

6000OvercastskyIntermediateskyClearsky

f)CRIforSSL


CRI(‐)

75

80

85

90

95

100


*

#*

++ +

+ +

#*

# # #

e)CRIforBL


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

CRI(‐)

75

80

85

90

95


g)EecforBL


E ec(W/m

2 )

0

1

2


+* *# # #

+ + + + + + +

# # # # # #* *

+ + +

# # #

h)EecforSSL


E ec(W/m

2 )

0

1

2

3


+ * *# # #

* *# # #


103


All visual comfort variables showed significant lower values for DIM conditions than for theother two lighting conditions (5‐way rANOVA; main effect of ‘condition’; Duncan’s multiplerange test; p< 0.05; Figure 5.9), except for two items: subjective glare ratings and “toomuchlight towork”ratings.TheBLconditions led tomoreglaresensationsand“toomuch light forwork” ratings compared to DIM and SSL conditions (p<0.05; Figure 5.9 b) and c). Moreover,subjectsratedhigherlightpreferenceandvisualcomfortunderSSLthanBLconditions(5‐wayrANOVA;‘time’x‘conditions’x ‘chronotype’x ‘gender’x‘order’interactions;Duncan’smultiplerangetest;p<0.05;Figure5.9a).

Figure5.9:Averagedvisualcomfortvariablesforalllightingconditions(N=32).DIM=brownbars;BL=bluebarsandSSL=pinkbars(mean±SEM);*=significantdifferencesbetweenDIMandtheothertwo

Lessthan2hours 2‐4hours 4‐6hours Morethan6hours

11.Icanworklonger

myusualworkplace

Worse Ratherworse Same Ratherbetter Better

LessSatisfiedMoreSatisfied

0 20 40 60 80 100

2.Visualcomfort

1.Lightpreference DimBLSSL

10.Comparingto

underthislighting

NoYes

0 20 40 60 80 100

4.Toomuchlightforwork

3.Toodimforwork

IntolerableImperceptible

0 20 40 60 80 100

6.Subjectiveglarerating

5.Lightdistribution

PoorlyWell

9.Lightistoo"cool"

8.Lightistoo"warm"

7.Myskinisunnaturaltone

NoYes

0 20 40 60 80 100

*

*

*

*

*

*

*

*

#

#

+

+*

+

+

a)

b)

c)

d)

e)


104

conditions;+=significantdifferencesbetweenBLandtheothertwoconditions;#=significantdifferencesbetweenSSLandtheotherconditions(p<0.05).

Overall,therewasasignificantdecreaseoflightpreferenceandvisualcomforttowardstheendof the study, regardless of the ‘chronotype’ as a variable (between12‐14hours and16hoursafterhabitualwake‐time;maineffectof‘time’;Duncan’smultiplerangetest;p<0.05;notshown).OnlyforBLconditionsthereweresignificantdifferencesbetweenMTandET(5‐wayrANOVA;‘chronotype’x’time’;Duncan’smultiplerangetest;p<0.05);ETshowedhigherlightpreferenceunderBLconditionsfivehoursafterwake‐time,whencomparedtoMT(Figure5.10a).Theyalsoexpressed a better visual comfort (under BL), one and four hours after habitual wake‐time(Figure5.10c).

Nosignificantdifferencewiththefactors´time’or‘chronotype’wasfoundfortheSSLconditions(p>0.50;Figure5.10bandd).Overall,therewerenosignificantdifferencesregardingthefactor‘gender’ or ‘order’ for visual comfort variables for the three different lighting conditions(p>0.25).

Figure5.10: Averagedtimecoursesofbothchronotypesfor: lightpreference(a‐b);visualcomfort(c‐d)underBL(leftcolumn)andSSLconditions(rightcolumn).MT(N=16):solidlinesandET(N=16):dashedlines(mean±SEM);*=significantdifferencesbetweenMTandET(p<0.05).

The subjective glare ratings varied also over time (5‐way rANOVA; main effects of ‘time’;p<0.05);howevertheydidnot increasetowardstheendof thestudyaswasthecase for lightpreference and visual comfort. Compared to the first hour, lower glare sensations wereexpressed8and10hoursafterwake‐time.Lateron,nodifferencecomparedtothebeginningof

BLCondition


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

LessPreferredMorePreferred

0

10

20

30

40

50 MT(n=16)ET(n=16)

*

BLCondition


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

LessCom

fortMoreComfort

0

10

20

30

40

50 MT(n=16)ET(n=16)

**

SSLCondition


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

LessPreferredMorePreferred

0

10

20

30

40

50 MT(n=16)ET(n=16)

SSLCondition


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

LessCom

fortMoreComfort

0

10

20

30

40

50 MT(n=16)ET(n=16)

a)LightPreference

c)VisualComfort

b)LightPreference

d)VisualComfort


105

the studywas founduntil the study end (5‐way rANOVA; ‘time’ x ‘condition’ x ‘chronotype’ x‘gender’x‘order’interactions;Duncan’smultiplerangetest;p>0.05).

Inanextstep,BLandSSLconditionswereanalysedseparatelyfordifferentskyconditions. InBL,subjectsexpressedlowerratings fordiscomfortglareunderintermediateskiesthanunderovercast skies (2‐way rANOVA;maineffectof ‘sky conditions’,p<0.05).Compared toovercastskies, subjects rated less discomfort glare under an intermediate sky5, 8‐9, 14 and 16hoursafter their habitualwake‐times; they also rated less discomfort under clear skies three hoursaftertheirwake‐time(2‐wayrANOVA; ‘time’and ‘sky ’ ;Duncan’smultiplerangetest;p<0.05;Figure 5.11). There was no impact of sky conditions on subjective glare ratings under SSLconditions(2‐wayrANOVA;p>0.50),andnoeffectofskyconditionsontheothervisualcomfortvariables(p>0.08).

Figure5.11:EffectsofweatheronsubjectiveglareovertimeunderBLconditions;overcastsky=browncircles;intermediatesky=greencircles;clearsky=bluecircles(mean±SEM);+=significantdifferencesbetweenintermediateandovercastskies(p<0.05);#=significantdifferencebetweenclearandovercastskies(p<0.05);n=numbersofsubjectsduringtherespectiveweathercondition.


Subjectivealertness,mood,physicalwell‐beingaswellasrelaxationvariedovertimeduringthestudy sessions (5‐way rANOVA; main effect of ‘time’; p< 0.05). The subjects reported loweralertness,aworsemoodandworsephysicalwell‐beingafterninehourssincetheirwake‐time,and they felt less relaxedafter12hours (Duncan’smultiple range test;p<0.05).They felt lessalert,lessphysicallywellandlessrelaxedunderDIMconditions,whencomparedtoBLandSSLconditions (5‐way rANOVA; main effect of ‘condition’; Duncan’smultiple range test; p<0.05).

Hours since Waketime (hrs)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

Subjective Glare

Imperceptible Intorelable

0

20

40

60

Overcast sky (n=10)

Intermediate sky (n=11)

Clear sky (n=11)

# + + + + +

BL Condition


106

Therewas,however,nosignificant ‘condition’effectonmood(p>0.78).Overall, therewerenosignificant differences between chronotypes on subjective assessments for BL and SSLconditions.

Thetimecoursesofthoseassessmentswereanalysedseparatelyforeachlightingconditionandthethreedifferentskies.UnderSSLconditions,thesubjectsfeltsignificantlysleepier(lessalert)underanovercastskycomparedtointermediateorclearskies(2‐wayrANOVA;maineffectof‘sky condition’, p<0.05). Compared to overcast skies, they rated to be significantlymore alertunderanintermediatesky9to13hoursaftertheirhabitualwake‐time,aswellas11hoursaftertheirwake‐timeunderclearskyconditions (2‐wayrANOVA; ‘weather’ and ‘time’ interactions;Duncan’smultiple range test, p<0.05; Figure 5.12 a). A significant interactionwas found formood:thesubjectswereinabettermoodunderaclearskyduringthelasthoursofthestudy,when compared to an intermediate sky (2‐way rANOVA; ‘weather’ and ‘time’ interactions;Duncan’smultiplerangetest,p<0.05),asshowninFigure5.12b).Therewasnoimpactof ‘skycondition’onalertnessandmoodunderBLconditions.


107

Figure5.12: TimecoursesunderSSL(N=32)conditionsfor:a)subjectivealertness;b)mood;overcastsky= brown circles; intermediate sky=pink circles; clear sky= blue circles (mean± SEM); + = significantdifferences between intermediate andovercast skies (p<0.05);#= significantdifferencebetween clearandovercastskies (p<0.05); *= significantdifferencebetweenclearand intermediatesky(p<0.05);n=numbersofsubjectsduringtherespectiveskycondition.

Correlationsanalysis

ThephysicallightpropertiesunderBLandSSLconditionswerethencorrelatedwithsubjectiveassessments, suchasvisual comfort variables andnon‐visual functions.Photometric variablesshowed small but significant correlations with both visual comfort variables and non‐visualfunctionsunderall lightingconditions(Pearson’scorrelation;p<0.05).Itmustbeoutlinedthat


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17

SubjectiveAlertness

SleepierMoreAlert 0

20

40

60

80 Overcastsky(n=13)Intermediate(n=6)Clearsky(n=13)

### # # # #

+ + + + +

a)SubjectiveAlertness


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17

Mood

WorseMoodBetterMood

40

50

60

70

80

90

Overcastsky(n=13)Intermediatesky(n=6)Clearsky(n=13)

*

b)Mood

SSLConditions

SSLConditions


108

verticalilluminanceunderBLconditionscouldnotbeusedforthecorrelationanalysiswiththeothervariables,sinceitwasmaintainedconstantat1000lx.

Correlationsbetweenphysicallightpropertiesandvisualcomfortvariables

Table5.9a)showsthecorrelationcoefficients(r) forvisualcomfortvariables.HigherEv,CCT,CRIandEecwerepositivelycorrelatedwithgreaterlightpreferenceandvisualcomfortaswellaslongerworking hours. Higher values of photometric variableswere also associatedwith lessfeelingof“toocold”and“toowarm”lighthues.LowerCCT,CRIandEecvalueswereassociatedwith significantly larger glare sensations, more feeling of “too much light to work” and lessnaturalskintone.

For the lightingdistribution,higherCRIandEecwereassociatedwithapoor lightdistributionunderBLconditions.On theotherhand,higher illuminanceandEecwere linkedwithabetterlightdistributionunderSSLconditions.VisualcomfortvariableswereassociatedwiththecolourappearanceoflightunderBLconditions,suchthathigherCCT,CRIandEecwerecorrelatedwithbetter visual comfort and light preferences; under SSL conditions, neither CCT nor CRIwerecorrelatedwiththelatter.BetterlightpreferenceandvisualcomfortwererelatedinthiscasetolargerEvandEecvalues.Interestingly,onlyEecwaspositivelycorrelatedwithvisualcomfortandlightpreferenceunderbothBLandSSLconditions.

Correlationsbetweenlightpropertiesandnon‐visualfunctions

The significant correlation coefficients (r) are shown in Table 5.9 b) (Pearson’s correlation;p<0.05).LargervaluesofEv,CCT,CRIandEecwerepositivelycorrelatedwithmorerelaxation,betterphysicalwell‐beingandgreateralertnessunderbothconditions.Moodwasinfluencedbythecolourappearanceof light,such thatpositivecorrelationsofbettermoodwith largerCCT,CRIandEecvalueswerefound,butnosignificantcorrelationsofmoodwithEv.


109

Correlation Higher Higher Higher Higher

Coefficient Ev CCT CRI Eec

BL SSL BL SSL BL SSL BL SSL

a)Visualcomfortvariables

1.Greaterlightpreference x 0.12 0.23 ‐ 0.20 ‐ 0.20 0.11

2.Bettervisualcomfort x 0.11 0.26 ‐ 0.25 ‐ 0.24 0.11

3.Toodimlighttowork x ‐0.13 ‐ ‐ ‐ ‐ ‐ ‐0.12

4.Toomuchlightforwork x ‐ ‐0.32 ‐ ‐0.29 ‐0.10 ‐0.28 ‐

5.Better‐distributedlight x 0.12 ‐ ‐ ‐0.15 ‐ ‐0.13 0.11

6.Moresubjectiveglarerating x 0.09 ‐0.37 ‐0.11 ‐0.30 ‐0.10 ‐0.30 ‐

7.Myskinismorenaturaltone x ‐ 0.18 ‐ ‐ 0.10 0.11 ‐

8.Lightistoo"warm" x ‐0.12 ‐0.15 ‐0.19 ‐ ‐0.16 ‐0.11 ‐0.13

9.Lightistoo"cool" x ‐0.13 ‐0.13 ‐ ‐0.11 ‐0.09 ‐0.09 ‐0.12

10.Thislightingisworsethanmy x ‐ 0.12 ‐ 0.16 ‐ 0.12 ‐

usualworkplace

11.Icanworklongerunder x 0.15 0.11 ‐ ‐ ‐ 0.10 0.13

thislighting

b)Non‐visualfunctions

1.Morerelaxed x ‐ 0.16 0.17 0.10 0.25 0.18 0.10

2.Betterphysicalwell‐being x 0.10 0.24 0.25 0.24 0.33 0.26 0.15

3.Morealert x 0.19 0.20 0.20 0.22 0.27 0.24 0.22

4.Bettermood x ‐ 0.18 0.20 0.11 0.18 0.15 0.10

Table5.9:Correlationcoefficients(r;Pearson’scorrelation)ofphotometricpropertieswithvisualcomfortvariables, and non‐visual functions; only significant correlation coefficients are indicated (p<0.05) areshown(‐=p>0.05);

Inter‐correlationsofvisualcomfortvariables

Inanextstep,theintercorrelationswithdependentvisualcomfortvariableswereinvestigated(Table5.10a).Asexpected,visualcomfortandlightpreferencewerehighlycorrelated(r>0.88;p<0.001).Bothvariablesshowedhighcorrelationcoefficientswithsubjectiveglareratingssuchthat greater light preferences and visual comfortwere associatedwith lower subjective glareratings and longer working hours under these conditions (r>0.41; p<0.001). Lower visualcomfortandlowerlightpreferenceswererelatedtomorefeelingsof“cold”and“warm”light,aswellasa“toomuchlighttowork”(r>0.58;p<0.001).

Inter‐correlationsbetweenvisualcomfortvariablesandnon‐visualfunctions

Furthermore,thevisualcomfortvariableswerecorrelatedwithnon‐visualfunctions,asshowninTable5.10c).Severalpositivecorrelationswerefound:morerelaxation,betterphysicalwell‐being,greateralertnessandbettermoodwererelatedtobettervisualcomfortandgreaterlightpreferencesunderBLconditions;nosignificantcorrelationswerefoundunderSSLconditions.Theaforehandmentionednon‐visualfunctionswerealsonegativelycorrelatedwiththefeelingof too “cold” or too “warm” light. Better moods were associated with a better luminous


110

distribution, more natural skin colour appearance and sufficient light to work under BLconditions (Table 5.10 b; r>0.34; p<0.001). Less relaxation and worse physical well‐beingshowedsmall,butsignificantcorrelationswithhigherratingsof“toomuchlightforwork”andmoresubjectiveglareunderbothlightingconditions(Table5.10b;r>0.09;p<0.05).

Inter‐correlationsofnon‐visualfunctions

Thedependentnon‐visualfunctionswerealsointer‐correlated,asshowninTable5.10d).Highcorrelationswerefoundbetweengreaterrelaxationandbettermoodandbetterphysicalwell‐being (r>0.72; p<0.001). The latter showed also high correlationswith greater alertness andbettermood(r>0.48;p<0.001).

Positive Higher Better More Better More Better

Correlation light visual relaxed physical alert mood

Coefficient preference comfort well‐being

BL SSL BL SSL BL SSL BL SSL BL SSL BL SSL

a)Withinvisualcomfortvariables b)Visualcomfortvariablesandnon‐visualfunctions1.Higherlightpreference

x x 0.92 0.88 ‐ 0.17 ‐ 0.09 ‐ 0.17 ‐ 0.17

2.Bettervisualcomfort 0.92 0.88 x x ‐ 0.17 ‐ 0.09 ‐ 0.13 0.11 0.18

3.Toodimlightforwork

‐ ‐0.23 ‐ ‐0.25 ‐0.15 ‐ ‐0.17 ‐ ‐ ‐ ‐0.43 ‐0.17

4.Toomuchlightforwork

‐0.58 ‐0.18 ‐0.66 ‐0.24 ‐0.12 ‐0.24 ‐0.15 ‐0.20 ‐ ‐0.09 ‐0.32 ‐0.17

5.Better‐distributedlight

0.13 0.29 0.10 0.29 ‐ 0.17 0.11 0.26 ‐ ‐ 0.36 0.28

6.Moresubjectiveglare

‐0.58 ‐0.43 ‐0.63 ‐0.44 ‐0.09 ‐0.20 ‐0.13 ‐0.20 ‐ ‐ ‐ ‐0.09

7.Colourskinismore ‐0.11 0.17 ‐ 0.20 0.16 ‐ 0.23 0.16 ‐ ‐ 0.34 0.32

natural

8.Lightistoo"warm" ‐0.17 ‐0.21 ‐0.24 ‐0.23 ‐0.23 ‐0.20 ‐0.22 ‐0.23 ‐ ‐0.09 ‐0.32 ‐0.23

9.Lightistoo"cold" ‐0.40 ‐0.43 ‐0.44 ‐0.45 ‐0.13 ‐0.19 ‐0.16 ‐0.15 ‐0.12 ‐0.09 ‐0.32 ‐0.19

10.Thislightingisworse

‐0.23 ‐0.30 ‐0.21 ‐0.30 0.11 ‐0.14 0.09 ‐ 0.17 ‐ ‐ 0.10

thanmyusualworkplace

11.Icanworklongerunderthislighting

0.41 0.54 0.43 0.53 ‐ 0.13 ‐ ‐ ‐ 0.16 0.09 0.13

c)Non‐visualfunctionsandvisualcomfortvariables d)Withinnon‐visualfunctions

1.Morerelaxed 0.17 ‐ 0.17 ‐ x x 0.78 0.72 0.45 0.34 0.54 0.51

2.Betterphysical 0.09 ‐ 0.09 ‐ 0.78 0.72 x x 0.53 0.48 0.59 0.52

well‐being

3.Morealert 0.17 ‐ 0.13 ‐ 0.45 0.34 0.53 0.53 x x 0.34 0.34

4.Bettermood 0.17 ‐ 0.18 0.11 0.54 0.51 0.59 0.59 0.34 0.34 x x

Table5.10:Correlationcoefficients(Pearson’scorrelation;r)betweenandwithinvisualcomfortvariablesandnon‐visualfunctions;onlysignificantcorrelationcoefficientwherep<0.05areindicated.(‐=p>0.05).


111

5.2.3 Discussion

Thisstudyshowed(asexpectedfromtheliterature)thatallsubjectsexperiencedabettervisualcomfort and greater lighting preference under SSL than under BL conditions. Under BLconditions,mostofthevisualcomfortvariablessignificantlydecreasedattheendofthestudy.SignificantdifferencesbetweenchronotypeswereshownunderBLconditions:MTchronotypesexpressedlowervisualcomfortandlowerlightpreferencethanETchronotypesatthebeginningofthestudy.UnderBLconditions,subjectiveglarewasratedasmoreintolerableunderovercastskythanintermediateandclearskiesatdifferenttimes.Theglareratingsvariedovertimebutdid not decrease towards the end of the study. For non‐visual functions, subjective alertness,mood and well‐being decreased over time. Most subjective assessments such as alertness,relaxation andphysicalwell‐beingwere rated significantly lowerunderDIM conditionswhencompared to BL and SSL conditions (except mood). Furthermore, under SSL conditions, thesubjectsexpressedgreatersleepinessunderovercastskiesthanunderintermediateorclearskyconditions.Moodwasalsodifferentlyassessedunderthoseskies:moodwasratedbetterunderaclearskythanunderanintermediateskyduringthelasthourofthestudy.

AssumingthattheBLandSSLlightingconditionswererealisticofficelightingscenarios,theSSLconditionswouldcorrespondtoofficelightingwithindividualcontrolledsystems,whiletheBLconditionsmimicked realisticoffice lightingwith centrally controlled systems.The automatedcontrol systems were targeting a 1000 lx threshold at subjects’ eye; this lighting conditionprovidesasaturatingstimulationforthenon‐visualsystemaccordingtothestudyofCajochenetal. [193]. The effects of visual comfort between chronotypes were only observed under BLconditions:MTchronotypes rated lowervisual comfort (oneand fourhours after theirwake‐time)andindicatedlesslightpreference(fivehoursaftertheirwake‐time)thanET.ThiscannotbeexplainedbyilluminancedifferencessinceilluminancewaskeptconstantinBLconditionsforboth chronotypes. Lower CCT and CRI values for MT, under BL conditions, compared to ETmighthavecontributedtolowervisualcomfortratingsforMTatthebeginningofthestudy,butnotafter4hours,whenphysical lightpropertieswere thesame forbothchronotypes (Figure5.7).

Interestingly, subjects expressed more glare sensations under overcast sky conditions thanunder the other sky types at different times; although daylight availability was higher forintermediate and clear skies compared to an overcast sky. It is very likely that somecomplementaryelectric lightingwasaddedunderovercastskies inorder toreachthe1000 lxvertical illuminance target forBLconditions. In fact, therewas someevidence from the studypresented in Section 5.1 aswell as from previous studies [23, 45, 49, 258] that the subjectsusuallyratedhigherglaresensationsunderelectriclightingthanunderdaylighting.Thismightexplainthehigherglaresensationsexpressedunderovercastskyconditions.Itisalsopossiblethat the different skies provided varying daylight fluxes into the room, which contributed todifferentluminousdistributions;thispointwillbediscussedfurtherinthenextchapter.

Under SSL conditions, which are considered as individually controlled lighting systems, thesubjects chose vertical illuminance values mostly depending on daylight availability acrossdaytime hours. This is in agreement with previous studies showing that under individuallycontrolled systems, the artificial lighting is switched on only when the (indoor) daylightworkplane illuminance isminimal [76,269‐271]. Photometricpropertiesof lightvariedover


112

time, and under SSL also according to the subjects’ choices. They rated SSL conditions asextremely high regarding visual comfort and light preference; those assessments did notdecreaseovertime.PreviousstudiesbyMooreetal.alsoreportedthewiderangeofilluminanceandhigh satisfaction ratingsunder individually controlledoffice lighting [76, 272].Regardingthe sky conditions, no influence of weatherwas found for visual comfort or light preferenceeither under BL or under SSL conditions. This is not in agreementwith a previous study byLaurentin et al. [248], showing that the subjects rated visual comfort asmorepleasantunderclearskycomparedtoovercastsky.Itmightbethatinthatstudysubjectswereexposedtotheviewoutofthewindows,whileinourstudy,nooutsideviewwasallowed.Itisalsopossiblethatunpleasantlightingsituationswerepreventedbytheindividuallightconditionsofthesubjects.

ComparedtoDIMconditions(withalmostnolight),thetworealisticofficelightingscenariosBLandSSLindicatedabsoluteeffectsof‘light’ontheofficeworkers.Lightaffectedclearlythevisualcomfort variables aswell as subjective alertness andwell‐being; however, nodirect effect onmoodwasfound.Fromtheliteratureisknownthatbesidespsychiatricdisorders[224],lightcanalsodirectlyimpactonmoodinhealthysubjects[211,215].Inapreviousstudy,abettermoodwasfoundduringsunnydayscomparingtocloudydays;thiswastrueevenforpeopleworkingwithin buildings [210]. Similarly, better mood was observed under clear skies significantlylongerthanunderintermediateskiesattheendoftheday[210].

Forsubjectivealertness,althoughtherewasnosignificanteffectbetweenBLandSSLconditions,changesofsubjectivesleepinessunderSSLthanBLconditions(incomparisontothefirsthour)werepreviouslyreported[266];thiswasexplainedbylowerilluminanceinthesecondhalfofthe SSL sessions, when lower lighting conditions were chosen according to subjects’ choice[266].SubjectsratedthemselvestobelessalertunderSSLconditionsforovercastskiesthanforthe other two sky types. The greater daylight availability under intermediate and clear skyconditionsaccordinglypreventedsleepinessfromthemiddleuntiltheendoftheday.Thisisinagreementwithmany studies showing that exposure tobrightdaylight can reduce sleepiness[236]andthathigherlightintensityduringdaytimeledtogreateralertnessduringtheday[190,196].Interestingly,whenweexaminedtheaverageilluminanceandEecfortheseskyconditions,overall intermediate and clear sky provided larger Eec values (not illuminance in the case ofintermediate sky; Table 5.6) than overcast sky. This shows the impact of light on non‐visualfunctions, which obviously respond better to circadian metrics (according to C‐λ) than tophotometricquantities(accordingtoV‐λ).

Overall, fewer correlationsbetweenphotometric lightproperties andvisual comfort variableswere foundunderSSL thanunderBLconditions; thismightbedue to thevarietyof light fluxunder SSL thanBL conditions, aswell as due to subjects expressing very high visual comfortratingsunderSSL.ThosehighassessmentsdidnotchangeovertimeunlikeunderBLconditions.It appears that the perception of visual comfort varied throughout their preferred lightingconditions, not the physical properties of light,perse. However,when considering non‐visualfunctions, most of the variables were positively correlated with photometric variables underbothconditions.Interestingly,thismightimplythatthesubjects’non‐visualfunctionswerenotinfluencedby their light preference; subjective alertness,well‐being andmoodwerehoweverdirectly influenced by the physical properties of light. Asmentioned above, visual discomfortmightbepreventedbyindividuallightingpreference;however,subjectivealertness,well‐beingandmoodmightnotbeimprovedbylightingpreference.


113

Lastly,visualcomfortandlightpreferenceshowedsignificantcorrelationswithbothilluminanceandcolourappearance, suchasCCTandCRI.UnderBL lightingconditions,where theverticalilluminancewasmaintainedconstantaround1000lx,bettervisualcomfortandlightpreferencecorrelatedwithhigherCCT,CRIandEec.Conversely,underSSLconditions, thesamevariableswerecorrelatedpositivelywithhigherilluminanceandEec;nocorrelationswithCCTorCRIwerefoundhowever.ItisinterestingtofurtherinvestigatewhetherEeccanbeanaccuratepredictorofsubjectivevisualcomfortinfurtherstudies.Moreover,abetterlightingdistributioncorrelatedwithahigherEecunderSSLconditions.TheeffectsunderBLconditionsweretheopposite:lowerilluminance and lower Eecwere associated with a more optimal luminous distribution. Theopposite correlations under SSL and BL conditions may also be explained by the luminousdistribution;thisistheobjectiveofstudyofChapter6.

5.3 Conclusion

Thefirstexperimentconfirmedtheimpactoflightonvisualcomfort,subjectivealertness,moodandwell‐being.Thesubjectsbecamesleepierandexpressedalowerphysicalwell‐beingearlierunderpureelectricallightwhencomparingtodaylightingconditions.Theseresultsshowedthatphotometric lightpropertiesperse, canenhanceourvisualcomfortsincecorrelationsshowedan improvement of visual comfort and well‐being with a larger daylight availability. Higherilluminance,CCT,CRI,Eecandlongersunshinedurationwerealsocorrelatedwithgreaterlightpreference,abettervisualcomfortandphysicalwell‐being,aswellasmorerelaxation.

Forthesecondexperiment,underbrightlightconditionswheretheverticalilluminanceateyelevelwasmaintainedconstantat1000lx,visualcomfortdecreasedinthecourseofthestudy,aswell as light preference. Different sky conditions influenced also the subjective assessments,suchasglareratings: the latterwereoftenobservedunderovercastskyconditionsdueto thecomplementaryaspectofelectriclighting.

When subjects were allowed to select their own lighting systems, they chose the one whichinvolved daylighting all day. Visual comfort and light preference were rated very high; thecorresponding assessmentsdidnot change over time.Different sky conditionshad significantimpacts on subjective alertness and mood. Under overcast sky conditions the subjects feltsleepierthanunderintermediateandclearskyconditions.Underclearsky,theyexpressedalsobetter mood at the end of the study compared to intermediate sky conditions. Subjectivealertness,well‐beingandmoodhadsignificant andpositive correlationswith illuminanceandcolourappearance,suchasCCTandCRI,aswellascircadianweightedilluminanceEec.

Besides the photometric light properties, such as intensity and colour appearance, luminousdistribution played also an important role regarding visual comfort. Assessments of overallluminance distribution, as well as objective glare sensations over time will provide a bettercomprehensionofthelightimpactonvisualcomfort:thiswillbediscussedinthenextchapter.


114

115

Chapter6 Discomfortglareandsubjectiveassessments

Severalglareindiceshavebeendevelopedtoassessvisualcomfort[45,51,148,273,274](seealsoSection3.2).Theavoidanceofdiscomfortglareishoweveronlyoneaspecttoachievehighvisualcomfort.AccordingtopreviousstudiesbyHopkinson[144],Chauvel[45],andBoyceandBeckstead [275], Velds [39] suggested that there might be several other factors involved indiscomfortglare,suchasphotometricvariables,weatherconditions,timeofday,outsideview,priorexperiencesofsubjects,andthesubjects’stateofmind.Forthisthesis,someoftheabovementionedvariableswhichareinvolvedindiscomfortglarewereexaminedtogetherwithglareindices.Theaimwastoenlargeourcurrentunderstandingofvisualcomfortmechanisms.Itwasfurthertestedwhethersomenon‐visualfunctionswereinfluencedbydiscomfortglare.

The datawas obtained from two experimental studieswith human subjects (see Chapter 5).Photometric features, such as vertical illuminance at the eye level, colour temperature (CCT),colour rendering index (CRI) and circadian weighted irradiance (Eec), were correlated withsubjectivevisualcomfortassessments.Besidesthelightintensityandspectralcomposition,thischapterpresentsthediscomfortglareratingsachievedduringthetwoexperimentalstudies(seeChapter 5). The high dynamic range imaging (HDR) technique, described in Chapter 3, wasappliedtocapturealitsceneandtoassessitsluminancedistribution.AnordinaryCCDcamerawas used for that purpose in the first experiment (Section 6.1); a Camera‐Like Light Sensor(CLLS;seeSection3.4)wasused in thesecondexperiment(section6.2).Theoverallobjectivewastomonitortheluminousdistributioninanindoorenvironmentovertime,byusingdifferentglare indices, aswell as background luminance determined byHDR images. Finally, the glareindiceswereassociatedwithvisualcomfortvariablesandnon‐visualfunctionsatdifferenttimesofday.

Discomfortglareandsubjectiveassessments

116

6.1 Impactofdiscomfortglareonvisualcomfort,alertnessandmoodinrealisticofficeday‐andelectriclightingconditions(ExperimentI)

DiscomfortglarewasassessedinanindoorenvironmentduringtheExperimentI,asdescribedinChapter5 (Section5.1): the corresponding results arepresented in thisChapter.The glareindices determined during the afternoon as well as their associations with visual comfortparameterswere investigated. Preliminary resultswerepresented at a scientificmeeting andpublishedasconferenceproceedings[143].

6.1.1 Experimentalmethods

Studydesign

Thestudyparticipants,theroomsetupandthestudydesignaredescribedinSection5.1.Inthisstudy,HDRimagesfromtheindoorscenesunder(mainly)daylightingconditions(DL)involved22 subjects (for technical reasons, data from three subjects out of 25 couldnot be analysed).Subjectswereexposedtodaylightwithavertical target illuminanceat theeye in therangeof1000and2000 lx (mean:966 lx; SEM±94.8 lx).The lowerwindowpartof the test roomwascoveredwithtextileblindsinordertopreventanyoutsideviewforthesubjects.Theupperpartof the windows was equipped with Anidolic Daylighting Systems (ADS) which re‐direct thedaylightfluxtotheceilingandtotherearoftheroom[252];theskyvaultwaspartlyvisibleforthe subjects through the corresponding upper window. Electric lighting (from six ceilingmountedluminaires:58W,3000K)completedtheofficeroomlighting,whenverticalilluminancedroppedbelow1000lx.Subjectscame4‐5hoursaftertheirhabitualwaketimetothelaboratoryandspenttheafternoon(fromnoonto6PM)inthetestroominasteadysittingposition.Theywereallowedtoread,workorlistentomusic;buttheywerenotallowedtoperformcomputerwork.Figure5.1a)inSection5.2showstheviewofthetestroomwiththewindows.

HDRimaging

Asequenceof14pictureswastakenthreetimesduringeachstudyafternoon(at12PM,3PMand5:30PM),byusingacalibratedcamerawithafisheyelens(NikonCoolpix5400;FCE‐9lens).Thecamera was mounted on a tripod and placed next to the subjects. The camera sensor wasoriented in thegazingdirectionof thesubjects.Exposure timeswereset in therangeof2s to1/4000thofasecondwith4.0F‐stopsforabunchofsnapshots,accordingtotheHDRimagingtechnique (see Section 5.1). The 'white balance' was set to the 'daylight' position. Verticalilluminancewascontinuouslyrecordedin5min intervalsthroughoutthestudybymeansofaspectroradiometer (Specbos 1201, JETI, Germany). HDR imageswere created as described inSection3.3,byusingthesoftware'Photosphere'(v1.8.7U,availableonwww.anyhere.com[160]).


117

Glarerisksassessments

After generating a single HDR image, glare indices were determined by using the software'Evalglare' (v0.9, Fraunhofer Institute, Germany) [148, 162]. Figure 6.1 shows a final image(obtained from the HDR imaging technique), with potential glaring sources (coloured areas)detectedby'Evalglare'.Threedifferentglareindiceswereusedasobjectiveglareassessments:theDaylightGlareProbability(DGP)[148],theDaylightGlareIndex(DGI)[45]andtheUnifiedGlare Index (UGR) [276]. The vertical illuminance of each scene was also calculated by thesoftware‘Evalglare’.

Figure6.1:SampleHDRimagecorrespondingtosubject‘Liper29’at12PMwithpotentialglaringsourcesidentifiedbythesoftware'Evalglare'[162].

Statisticalanalysis

ThesoftwareStatisticav6.12(StatSoft,USA)wasusedforstatisticalanalyses.Firstly,Pearson’scorrelationswere used to relate the vertical illuminancemeasured by the spectroradiometerandthecalculatedverticalilluminance.Inthenextstep,glareindicesweresubjectedtorepeatedanalysesofvariance(rANOVA;one‐wayrepeatedanalysisofvariance);post‐hoctests(Studentt‐tests)wereappliedtoanalysepotentialvariationsover time. In the finalstep, thesubjectivenon‐visualfunctionratings,describedinSection5.1,wereassociatedwiththoseglareindicesbyusingPearson’scorrelations.

6.1.2 Results

Objectiveglareassessments

Vertical illuminance (as measured with a spectroradiometer), as well as the correspondingilluminance computedbyEvalglare showeda significant correlation (r =0.84; p<0.001).Bothvariablesaswellastheglareindices(DGP,DGIandUGR),determinedbypost‐processingofthe


118

HDRimages,variedinthecourseoftheafternoon(one‐wayrANOVA;p<0.001;Figure6.2).Thevariationofdiscomfortglarewassuchthatallglareindices(DGI,UGR,DGP)weresignificantlylowerat5:30PM,whencomparedtothoseearlierintheafternoon(t‐tests;p<0.05).

Figure 6.2: Dynamics of averaged glare indices (mean ± SEM) in the course of the afternoon; DGP =DaylightGlareProbability,DGI=DaylightGlareIndex,UGR=UnifiedGlareIndex;N=22.

Associationsofobjectiveglarewithsubjectiveassessments

AlowersubjectivevisualcomfortwassignificantlycorrelatedwiththeDGIinthemiddleoftheafternoon,andshowedatendencywiththeUGR(3PM;r=‐0.44forDGI,p<0.05;forUGR:r=‐0.40,p<0.1;Table6.1).Subjectivelightpreferencedidnotsignificantlycorrelatewiththethreeglareindices(DGP,DGIandUGR),butshowedtendenciesatthebeginningofthestudy(r>0.37;p<0.1;Table1).Therewasnocorrelationbetweenobjectiveandsubjectiveglareassessments(p>0.1;notshown).

Higher objective glare ratings (DGP, DGI and UGR) were significantly associated with lowersubjectiverelaxationat3PM(r<‐0.46;p<0.05),andshowedatendencyat5:30PM(r<‐0.36;p<0.1).Therewereslightpositivecorrelationsatthebeginningoftheafternoon,suchthathigherDGPandUGRtendedtorelatetobetterphysicalwell‐being(r>0.38;p<0.1);however,significantassociationsbetweenthethreehigherglareindicesandlowerphysicalwell‐beingwerefoundat5:30PM(r<‐0.48;p<0.05).Subjectivealertnesswasnegativelycorrelatedwith theDGP in themiddle of the afternoon (3PM), such that larger glare indices were associated with lowersubjective alertness (r<‐0.36; p<0.1). Finally,we did not find significant correlations betweenthethreeglareindicesandsubjectivemoodassessments(p>0.1).

Time

12PM 3PM 5:30PM

DG

I / U

GR

(-)

10

15

20

25

DG

P (

-)

0.10

0.15

0.20

0.25

0.30

DGI UGR DGP


119

Subjective Glareindices Glareindices Glareindices

Assessments

12PM 3PM 5:30PM

DGP DGI UGR DGP DGI UGR DGP DGI UGRVisual ‐ ‐ ‐ ‐ −0.44* −0.40# ‐ ‐ ‐comfort ↑glare↓comfort Light 0.37# 0.40# 0.38# ‐ ‐ ‐ ‐ ‐ ‐

preference ↑glare↑preferred Relaxation ‐ ‐ ‐ −0.60* −0.46* −0.47* −0.41# −0.36# −0.48*

↑glare↓relaxed ↑glare↓relaxedPhysical 0.38# ‐ 0.40# ‐ ‐ ‐ −0.50* −0.48* −0.59*well‐being ↑glare↑comfort ↑glare↓comfortAlertness ‐ ‐ ‐ −0.42* −0.36# −0.37# ‐ ‐ ‐

↑glare↓alert

Table6.1:Correlationcoefficients(r)betweenglare indicesandsubjectiveassessments(visualcomfort,lightpreference,subjectivewell‐beingandalertness)atdifferenttimesduringtheafternoon;*=p<0.05,#=p<0.1(tendency),and–=p>0.1(notsignificant).Forsignificantvaluesortendencies,thedirectionsofthecorrelationcoefficientsareindicatedwitharrows.

6.1.3 Discussion

Significant associations between less subjective visual comfort, less relaxation and loweralertness with higher glare indices were only found at 3PM. For physical well‐being andrelaxation there was a negative correlation also at 5:30PM. Since glare indices did notsignificantly vary between noon and 3:00PM (see Figure 6.2), it was rather the change ofsubjective assessments, not the change of indoor luminous distribution or the presence ofglaring sources, which caused the opposite direction of the correlations between noon and3:00PM.Thismayindicatethatinthecourseoftheafternoonsubjectsbecamemoreperceptivetotheincidenceoflightdistribution.Thenegativecorrelationat5:30PMcanalsobeexplainedbyachangeinilluminanceduetothefadingofdaylight.Thevariationsofsubjectivecomfortinthe course of a day were also described in Section 5.1 and are in agreement with previousresultsfromBoyceetal.[17].

Wecouldnotfindsignificantcorrelationsbetweensubjectiveglareratingsandglareindices,asreportedbyWienold,eventhoughsimilarglareassessmentswereused[249].Oneexplanationmight be that the subjects could avoid discomfort glare caused by direct sunlight during theentirestudy:textileblindswereloweredtopreventanyoutsideview,butalsotoavoidsunraysenteringthetestroom.Secondly, thestudieswereundertaken inaroomwhichwasequippedwith Anidolic Daylighting Systems [277, 278]. These systems provide a more even luminousdistributionontheworkplaneanddeeperintheroomwhichreducesanyglarerisksassociatedto excessive luminance contrasts [151]. This is in agreement with the statement by Velds,indicating that glare indices were developed to assess discomfort glare. They are notappropriateforanywindowsordaylightingsystemsduetothedifferentluminousdistribution[39]; most of the glare indices were developed for electric lighting systems and point light


120

sources. Finally, the sample size of the study might be too small. Further investigationsregardingtheeffectsof indoor luminousdistributiononsubjectiveassessmentsareneededtoextendthestudytolargerpopulationsandlongertimeepisodes.

6.2 Impactofdiscomfortglareonvisualcomfort,subjective,alertnessmoodandwell‐beingforvariouslightingconditionsinextremechronotypes(ExperimentII)

6.2.1 Subjectsandmethods

Studydesign

The inclusionof extrememorning types (MT) andextremeevening types (ET)wasdescribedearlier in this thesis (see Chapter 5). Demographic details, the room setup and study designwere described in details in Section 5.2.1. This section presents subjective assessments andindoorluminousdistributionsof32subjects(16MTand16ET).

ThestudywasconductedinatestroomlocatedintheLESOSolarEnergyExperimentalBuildingequipped with Anidolic Daylighting Systems; the lighting conditions were similar to thosedescribed in Section 5.2. Only bright light conditions (BL) and self‐selected conditions (SSL)wereusedforobjectiveglareassessment,sincedimlightconditions(DIM)werenotconsideredasrealisticofficelightingconditions.UnderBLconditions,theverticalilluminancewastargetedat ~1000 lx throughout the entire session. Under SSL conditions, the subjectswere asked tochoosetheirlightingpreferenceseachhour.Theirchoicescompriseddaylightingand/orelectriclighting,asshowninFigure5.5anddescribedinTable5.6(seeChapter5).

Imagesofthescenesmimickingthesubjects’viewwererecordedhourlyduringthe16hoursofthestudyunderBLandSSLconditionsusingtheCamera‐LikeLightSensor(CLLS)describedinChapter3.TheimagesequencesofonesubjectcouldunfortunatelynotberecordedduringtheBL session, due to a technical problem (BL: n=31; SSL: n=32). For all study sessions, verticalilluminanceat theeye level (Ev), correlatedcolour temperature (CCT), colour rendering index(CRI) and circadianweighted irradiance (Eec) were continuously recorded in 5min intervalsthroughoutthestudy,usingaspectroradiometer(Specbos1201,JETI,Jena,Germany).

Visual comfort variableswere assessed hourly,while subjective alertness,mood and physicalwellbeingwereexaminedevery30minutes(seeSection5.2.1).

CLLSimaging

Adigitalimagewascapturedandavideomoviewasrecordedhourly,alwaysafterthesubjectivevisual comfort assessments, using the software “DeviseIcyLightGUI” (which is an executableJAVAGraphicalUser's Interface for “IcyCAM”)developedbyCSEM.A total of 16 sequencesof


121

hourly images for each subjectwas analysed to assess the luminance distribution in the testroom. In a next step, the software “ImageCAM”, developed by LESO‐PB/EPFL was used toproduceanHDRimagefromeachdigitalpicturetakenbythecalibratedCLLS(seeSection3.2).

Glareratings

Glare indiceswerecalculatedusingEvalglare(Evalglarev1.11,FraunhoferInstitute,Germany)[148,162], asdescribed inSection6.1.Background luminanceaswell as threedifferentglareindiceswereusedasglareriskassessments:DaylightGlareProbability (DGP) [148], DaylightGlare Index (DGI) [45] and Unified Glare Index (UGR) [276]. Figure 6.3 shows three imagescaptured during a study session at different times of the day. Potential glare sources asidentifiedbythesoftwareEvalglareareshownwithcolouredareas.

Figure6.3:HDR images of subject ‘Chroli 22’, takenwith theCLLS in themorning (left picture), in theafternoon(middle)andintheevening(right).Thecolouredareasindicatepotentialglaresources.

Statistics

ThestatisticalanalyseswerecarriedoutusingthesoftwareStatistica6.12(StatSoft,USA).Theanalysisfromsubjectiveassessmentsofnon‐visualfunctionswasdescribedinSection5.2.Datafrom objective glare assessments were analysed by first performing repeated analyses ofvariance (rANOVA)with two factors (2‐way rANOVA): ‘time’ (16 time points) and ‘condition’(BLandSSL).Inanextstep,subjectiveassessmentswereanalysedseparatelyforeachlightingconditiontoinvestigatetheeffectsofdifferentskytypes(overcast,intermediateandclearsky)at different times (2‐way rANOVA; ‘weather’ x ‘time’). Duncan’s multiple range tests wereapplied for post‐hoc analysis. Lastly, correlation analysis was performed using Pearson’scorrelationsfornormallydistributedvalues.

6.2.2 Results

Objectiveglareassessmentsovertime

Figure6.4illustratesthecalculatedglareindices(DGP,DGIandUGR),aswellasthebackgroundluminanceovertimeforbothconditions(BLandSSL)acrossallsubjects.Generally,glareindiceswere rather low (low glare); they were mostly below the borderline between comfort anddiscomfort (grey dashed line in Figure 6.4). All glare indices (DGP, DGI, UGR) as well as


122

backgroundluminancevariedovertime(2‐way;ANOVA;maineffectof‘time’;p<0.001),exceptEv forBLconditions. Allglare indicesandbackgroundluminancesweresignificantlydifferentbetweenBLandSSLconditions,andseveralstatisticalinteractionswerefound(2‐way;ANOVA;maineffectof‘condition’;p<0.001;Figure6.4).TheDGP,DGIandbackgroundluminancewerelowerunderSSLthanBLatthebeginningofthestudy(duringthefirsthour).Duringtheentiresecondhalfofthestudy,allglareindices(DGP,DGIandUGR)andbackgroundluminanceweresignificantlylowerunderSSL,comparedtoBLconditions(2‐way;ANOVA;p<0.001;Figure6.4).Itisnotedthattimeisrelativetothehabitualwaketimesofbothsubjectgroups(MTandET)onFigure5.4.

Figure 6.4: Averaged glare indices indicated relative to elapsed time since habitual wake time(mean±SEM)forDGP(a);DGI(b)UGR(c)andbackgroundluminance(d);BLconditions=solidline;SSLconditions = dashed line; *= p<0.05; dashed grey line=exceeding “disturbing” glare sensations. DGP =DaylightGlareProbability;DGI=DaylightGlareIndex;UGR=UnifiedGlareIndex(UGR)

For further analysis of the lighting environment, different weather conditions werecharacterizedbydetermining three sky types (clear, intermediate andovercast), according totheSwissNorm150911[267],asdescribedinSection5.2.2.

The averaged DGP, DGI, UGR and background luminance for different skies are presented inTable 6.2 for BL and SSL conditions. Overall, glare indices were not significantly differentbetween the three different weather conditions in BL. Only for SSL, background luminancedifferedbetweenweatherconditions; itwaslargerforclearskiesandlowerforovercastskies(Table6.2showsthecorrespondingaveragedvalues;2‐way;ANOVA;maineffectof ‘weather’;p<0.05).Aninteractionbetweenthefactors‘time’and‘weather’wasfound,suchthatfromfiveto six hours after habitual wake time, the DGP in SSL conditions was significantly lower forovercastthanforclearskies(Figure6.5a).Backgroundluminancewassignificantlyhigherwith

b)DGI


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

DGI

10

12

14

16

18

20

22

24

BLSSLDisturbing

d)Backgroundluminances


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

Luminance(cd/m

2 )

0

50

100

150

200

250

BLSSL

a)DGP


DGP

0.10

0.15

0.20

0.25

0.30

0.35

BLSSLDisturbing

c)UGR


UGR

12

14

16

18

20

22

BLSSLDisturbing

**

**** * *

* * ** *** * ***

* **

**** * *

* * ** *** * **


123

clear thanwith overcast skies (and partly intermediate skies), up to the first six hours afterhabitualwaketimeduringtheSSLsessions(Figure6.5b).TheDGIwassignificantlylowerwithintermediate than with clear skies after nine hours in BL conditions (Figure 6.5 c; 2‐wayrANOVA;p<0.05).

Lighting Sky N DGP DGI UGR Background

conditions conditions luminance

(‐) (‐) (‐) (cd/m2)

(mean±SD) (mean±SD) (mean±SD) (mean±SD)

BLOvercast 10 0.232±0.00 16.19±1.27 19.09±1.46 144±24

Intermediate 10 0.228±0.02 15.18±2.75 18.05±3.31 133±57

Clear 11 0.232±0.02 15.93±2.27 18.97±2.64 152±51

SSL

Overcast 13 0.181±0.06 13.94±3.46 16.73±5.20 64±51+#

Intermediate 6 0.206±0.06 13.99±2.56 17.17±3.30 124±120+

Clear 13 0.201±0.11 14.40±3.77 17.16±4.83 136±151#Table6.2:Averagedglareindicesandbackgroundluminance(mean±SD)fordifferentlightingconditions(BLandSSL)anddifferent sky types;DGP=DaylightGlareProbability;DGI=DaylightGlare Index;UGR=UnifiedGlareIndex(UGR);+=significantdifferencebetweenintermediateandovercastsky(p<0.05);#=significantdifferencebetweenclearandovercastsky(p<0.05);


124

Figure 6.5: Time courses under BL and SSL conditions of average a) DGP; b) background luminancesunderSSLconditions;c)DGIunderBLconditions;overcastsky=dashedanddottedlines, intermediatesky = dashed lines and clear sky = solid lines (mean ± SEM); + = significant differences betweenintermediate and overcast sky (p<0.05); # = significant difference between clear and overcast sky(p<0.05);*=significantdifferencebetweenclearandintermediatesky(p<0.05;N=32).

Associationsbetweenvisualcomfortandnon‐visualfunctions

Glareindicesandbackgroundluminanceswerecorrelatedwithsubjectiveassessments,suchasvisualcomfortvariablesandnon‐visualfunctions.Table6.3showsthecorrespondingsignificantcorrelationcoefficients(r).UnderBLconditions,higherglareindices(DGP,DGIandUGR)werecorrelated with lower lighting preference and poor visual comfort (r<‐0.13; p<0.05). Largerglareindiceswerealsoassociatedwithgreatersubjectiveglareandmorefeelingsof“toomuchlight to work” (r<‐0.13; p<0.05). DGP, DGI and UGR were not correlated with subjective

a)DGPSSLconditions


DGP

0.0

0.1

0.2

0.3

0.4


b)BackgroundluminanceSSLconditions


Luminances(cd/m

2 )

0

100

200

300

400


c)DGIBLconditions


DGI

10

12

14

16

18

20

22 OvercastskyIntermediateskyClearsky

# #

*+ +

*

# # # # #


125

luminousdistributionintheroom,exceptforhigherbackgroundluminance.Interestingly,underSSLconditions, the correlationswent in theoppositeway,whencompared to thoseunderBLconditions.Largerglare indiceswerecorrelatedwithhigher lightingpreferences,bettervisualcomfort,‘sufficientlighttowork’andamoreevenluminousdistribution(r>0.12;p<0.05).OnlyhigherDGIcorrelatedwithmoresubjectiveglare (r>0.09;p<0.05).Background luminancedidnot correlate at allwith any of the subjective visual comfort or glare assessments under SSLcondition.

Significant correlations between glare indices and non‐visual functions were found mostlyunderBLconditions.Higherobjectiveglareindiceswerecorrelatedwithlessrelaxation,worsephysicalwell‐being,loweralertnessandworsemood(Table6.3).TheDGIandUGRshowednosignificant correlationswith non‐visual functions (p>0.05). For SSL conditions, a higherDGPwasrelatedtogreateralertness.

Lighting Subjective Higher Higher Higher Higher

conditions variables DGP DGI UGR Background luminance

BL

Lowerlightpreference 0.13 0.25 0.24 ‐

Lowervisualcomfort 0.14 0.27 0.27 ‐

Moresufficienttowork ‐ ‐ ‐ ‐

Toomuchlighttowork 0.19 0.38 0.37 ‐Worseluminousdistribution ‐ ‐ ‐ 0.13

Moresubjectiveglare 0.23 0.41 0.39 ‐

Lessrelaxed ‐ 0.17 0.18 0.15

Worsephysicalwell‐being ‐ 0.21 0.21 ‐

Loweralertness ‐ 0.22 0.23 ‐

Worsemood 0.10 0.19 0.21 0.16

SSL

Higherlightpreference 0.13 0.12 0.13 ‐

Highervisualcomfort 0.12 0.12 0.12 ‐

Moresufficienttowork 0.17 0.21 0.22 ‐

Toomuchlighttowork ‐ ‐ ‐ ‐Betterluminousdistribution 0.09 0.11 0.11 ‐

Moresubjectiveglare ‐ 0.09 ‐ ‐

Morerelaxation ‐ ‐ ‐ ‐

Betterphysicalwell‐being ‐ ‐ ‐ ‐

Greateralertness 0.14 ‐ ‐ ‐

Bettermood ‐ ‐ ‐ ‐

Table6.3:Correlationcoefficientsbetweenglareindices(DGP,DGI,UGR)andvisualcomfortvariablesandnon‐visualfunctions(alertness,relaxation,physicalwellbeingandmood);p<0.05,‐=p>0.05(N=31).


126

6.2.3 Discussion

Theresultsshowthatglareindicesdidnotexceedthe'disturbing'leveloftheglareindexscales;theywere,however,differentunderBLandSSLconditions.UnderSSLconditions,adecreaseofallglareindicesandbackgroundluminancewasobservedfromthebeginningtowardstheendofthe study. Slight effects of weather conditions were found for DGP, DGI and backgroundluminancesatdifferenttimesoftheday.Thedifferentlightingconditions(BLandSSL)showedopposite directions in their correlations between glare indices with subjective assessments.Largerglareindiceswereassociatedwithpoorervisualcomfort,lessalertness,worsewell‐beingandmoodunderBL conditions.On theotherhand, under SSL conditions, larger glare indiceswereassociatedwithbettervisualcomfortandgreateralertness.

Inthesecondstudy,severalcorrelationsbetweenglareindicesandsubjectiveglareratingswereobservedeventhoughthefirststudydidnotshowanycorrelationsbetweenthesevariables.Itwasassumedearlierthatglareindicesreflectlessglaresensationsinthepresenceofdaylightingsystems (such as Anidolic Dayligting Systems and/or textile blinds). This assumption can bedebated since the second study was conducted in the same roomwith the same daylightingsystems. The most likely reason for the somehow inconsistent findings might be the studyduration,whichlastedfor16hours.Thiswasmorethandoubleaslongasthestudydurationinthefirstexperimentandshowedthedynamicsofdaylightinthecourseofanentireworkingday.Duringthistime,subjectiveassessmentsalsovaried.Relatedtothelattermightbethatthestudywas performed with extreme chronotypes. Consequently the inter‐individual range ofassessments was larger and the study was designed to scrutinise some of those differences.Whether their internal different circadian phasemight also have contributed to the observedassociationswithvisualcomfortandnon‐visualfunctionswillbefurtheranalysed.

AnewandinterestingfindingofthestudywasthatoppositecorrelationswerefoundbetweenBL and SSL conditions for visual comfort and subjective glare ratings. Individually controlledlighting reflected subjectively preferred lighting conditions and might prevent any negativeglare sensations, regardless of access to an outside view. A second reasonmight be that SSLconditions enabled a situation where subjects could actively adjust the lighting to becomfortable.

Itisimportanttonotethattheglareindicesweredevelopedfordifferentlightingsituations:theUGRwasrecommended inparticular forartificial lighting[51,273].DuringBLconditions, theDGPdidnotvaryovertime;itispossiblethatDGPisnotsostronglyinfluencedbythevariationof luminance [148]. Another reason might be that the vertical illuminance was keptapproximatelyat1000lxthroughoutthestudy. TheotherparametersofDGPwerenotampleenoughtoalterthelatterinthisstudy.

DuringBL conditions, electric lightwas addedover time, particularly at the nighttime: itwaspresumably leading to stronger correlations of UGR with subjective glare ratings under BLconditions.Ontheotherhand,DGPandDGIweredevelopedfordaylightingsituationsinordertoevaluate“glarefromwindows”[45,148].Thisstudywasconductedinatestroomequippedwith Anidolic Daylighting Systems; the assessments of “glare from anidolic systems” can beapplicablewith“glarefromwindows”.Thecorrelationsofsubjectiveandobjectiveglareratingsweresound.DGIwasoriginallybasedonexperimentswith“uniformlightsources”[144][279];


127

italsocorrelatedratherwellinthisstudywiththesubjectiveglareratingsissuedfromauniformlightdiffusersuchasthetextileblinds.Conversely,DGPwasdevelopedfromarealdaylightingsituationwithwindows (representing non‐uniform light sources from sunrays),which is lesssimilar to this case: significant correlationsofDGPand subjectivevisual comfort assessmentswere found, but the correlationswere smaller than the correlationswith the other two glareindices(DGIandUGR).

For non‐visual functions significant correlations were found however, especially for BLconditions(butnotforSSLconditions).Higherglareindiceswerecorrelatedwithlessalertness,poorerwell‐being andmood. Interestingly, thenon‐visual functionshad stronger correlationswith photometric properties, such as illuminance, CCT, CRI and Eec, than glare indices (seeChapter5).ItseemsthatthehighluminousintensityduringbrightlightconditionswasoverallperceivedasratheruncomfortableandviceversafortheSSL.Furtherinvestigationsaddressingtheseissuesaredefinitelyneeded.Correlationsofthethreeglareindiceswithsubjectivevisualcomfortassessmentsshouldbeinvestigatedhourlyand/orwithdaylightavailabilityinordertobetterunderstandtherealcauseofdiscomfortglare.

6.3 Conclusion

A favourable objective visual comfort assessment might not always go along with subjectiveglare ratings in the caseof individually controlled lighting systems. It ispossible that lightingpreference might be able to cancel or reduce visual discomfort, similar to the outside viewreported before [49, 50]. A new finding from this thesis is that the glare indices have alsosignificant correlations with non‐visual functions, even though these associations not alwayswentinthesamedirectionsandweredependingonthetimeofday.Interestingly,theywerenotwellreflectedbytheabsolutevaluesofglareindices.

Taken together, several glare indices, developed for evaluation of discomfort glare,correspondedwithsubjectiveassessmentsofvisualcomfort.Itisbelievedthatthisistherightway todevelop tools for theassessmentofdiscomfortglare.Atpresent it isnotyetverywellunderstood how an appropriate glare index should be chosen to evaluate a given lightingsituation (for different populations and environments). Evaluations of discomfort glare arecomplex,sincetheydealwithsubjectivesensations;forsure,thesurroundingsandenvironmentsuchasview,pleasantness,timeofday,daylightavailability,etc.playanimportantroleinthisprospect. Further investigations are needed; photometric light properties, such as colourappearance and/or circadianmetricsmust be taken into account, since correlations betweenthesepropertiesandsubjectivevisualcomfortassessmentswerefound.


128

129

Chapter7 Conclusion

7.1. Statementsandresearchquestions

Thisdoctoralthesisaimedatinvestigatingtheimpactoflightonvisualcomfort,includingbothvisual and non‐image forming (NIF) effects. Two relevant questions were addressed in thiswork, as mentioned in Section 2.3: i) how can the luminous distribution of environmentallightingbeassessedmoreefficientlyinsideofbuildings?ii)whatistheimpactoflightonhumanvisualcomfortincludingvisualandNon‐ImageForming(NIF)effects?Inthisconclusivesection,itisshownthatthethesisresultsprovidedsoundanswerstobothquestions.

7.1.1 ValuationofaCamera‐LikeLightSensor

ACamera‐LikeLightSensor(CLLS)hasbeenintroducedinthecourseofthisthesistofostertheuseofHighDynamicRange (HDR) imaging techniques for luminancemappingandglare risksassessment in aworking space. The CLLSwas calibrated for spectral sensitivity, photometricresponse and corrected for vignetting effects. Implementation of this novel device in realsettings can markedly reduce the time for image capturing when compared to HDR imageacquisition:itcancaptureascenewithinasinglesnapshotorrecorditinrealtime.Luminancemapping and glare risk assessments were carried out using the CLLS. The results werecompared to measurements issued from conventional, but definitively slower, HDR imagingtechniques.

In a next step, the CLLS was calibrated for photometric assessments in circadian metrics. Adifferent spectral sensitivitywas applied for thatpurposeby appropriate colored filters.As aresult, the circadianweighted radiance could bemonitored. The CLLSwas also validated forassessmentsofluminouspropertieswithrespecttoNIFfunctionswithinarealisticofficeroomequipped with Anidolic Daylighting Systems in the LESO experimental building at EPFL.Furthermore, it was used for comparisons of luminance distribution and circadian weightedradiance (Lec) distribution between two test rooms, equipped with different daylightingsystems, at the “BerkeleyLabAdvancedWindowsTestbedFacility” at theLawrenceBerkeleyNationalLaboratory(LBNL,CA,USA).

Finally, the practical implementation of the CLLSwas successfully validated in realistic officelighting conditions for both photometric and circadian sensitivity functions. It was efficiently

Conclusion

130

used as a tool to assess luminance distributions within an experimental study with humansubjects,aspresentedbelow.

7.1.2 ImpactofofficelightingonvisualcomfortincludingNon‐ImageForming(NIF)effects

Inthecourseofthisthesis,twoexperimentswithyoungparticipantswereconducted,aimingattesting the acute effects of different lighting conditions on subjective variables in a realisticofficeroom.Subjectiveassessments,suchasvisualcomfortvariables,alertness,moodandwell‐being were assessed by the participants every 30 and 60minutes, respectively. Photometricvariables were continuously monitored. To systematically assess also inter‐individualdifferences on visual and NIF functions, the second study comprised extreme chronotypes(extreme morning and evening types). They served as ‘natural’ examples for such inter‐individual differences due to their very different habitual working hours and presumablydifferentlightingrequirements.


Theresults fromthestudiesconfirmed thatoffice lightingconditionswithan intensiveuseofdaylight led to a better visual comfort andwerepreferredby the subjectswhen compared topureelectric lighting conditions, althoughanyoutsideview for the subjectswasprevented. Itwasalsoconfirmedthatsubjectiveglareratingswerehigher(more intolerable)underelectriclighting thanunderdaylight, regardlessofoutsideview,although thevertical illuminancewashigherunderdaylightingconditions.

Under the bright light conditions,where the vertical illuminancewasmaintained constant at1000 lx(thankstocomplementaryelectric lighting),visualcomfortdecreased in thecourseoftheday,aswellas lightpreference.Differentskyconditionsalsoinfluencedthevisualcomfortvariables,suchasglaresensations.Highersubjectiveglareratingswerefoundatcertaintimesunderovercastskyconditionswhencomparedtointermediateandclearskies.Potentialeffectsoflightoninter‐individualdifferencesofvisualcomfortandNIFwerealsoidentifiedindifferentchronotypes.Extrememorningtypesassessedlowervisualcomfortandlightpreferenceatthebeginningofthedaythanextremeeveningtypesubjects,whentheywereexposedtoconstantbright light. This is an interesting finding, because bright morning light can advance thecircadian phase to an earlier time. Since morning types have already very early wake‐ andbedtimes,verybrightlightinthemorningwouldreinforcethisadvance.Whentheycouldself‐select their lighting, extrememorning types chose a lower luminous intensity in themorningwhichmightbeanintuitivelycountermeasurebythemtopreventamoreadvancedtiming.

Inthestudysessionwherethesubjectscouldhavetheirowncontrolledlightingsystems,theypredominantlychosedaylight,ifavailable,asthepreferredlightsource.Thevisualcomfortandlightpreferenceappraisalswereverypositive;moreover, theirassessmentsdidnotvaryovertime. Different sky conditions did not alter the visual comfort appraisal under these lightingconditions. Overall, positive correlations between the photometric properties and the visualcomfortvariableswerefound:largerverticalilluminance,correlatedcolourtemperature(CCT),

Conclusion

131

colour rendering index (CRI), and circadian weighted radiance (Eec) correlated with greaterlightpreference,visualcomfortandlowersubjectiveglareratings.


Thestudiesoutlined the impactof realisticoffice lightingconditionsonsubjectiveNIFeffects,suchasalertness,moodandwell‐being.Thesubjectsbecamedefinitivelysleepierandindicatedpoorerphysicalwell‐beingearlierunderpureelectrical lightingthanunderdaylighting.Underself‐selectedlightingconditions,varyingskyconditionsdifferentlyaffectednon‐visualfunctions,suchasalertnessandmood.Anovercastskyledtosleepiersubjectsthanintermediateandclearskies. The subjects were also in a better mood under clear skies later in the evening, whencomparedtointermediateskyconditions.

Thephotometricvariableshadalsoasignificantimpactonnon‐visualfunctions;largerverticalilluminance, CCT, CRI and Eec correlatedwithmore relaxation, betterwell‐being and a bettermoodaswellasgreateralertnessunderbothbrightlightandself‐selectedlightingconditions.

Associationsbetweenvisualandnon‐visualfunctions

Strong correlations between visual comfort and light preference were found; visual comfortvariables also showed correlationswith subjective glare ratings.Moreover, the latter showedslightbutsignificantcorrelationswithnon‐visualfunctions:highersubjectiveglareratingswerecorrelated with less relaxation and worse mood as well as with worse physical well‐being.Greater visual comfort and light preference correlated with a more relaxed attitude, higheralertness, better physical well‐being and a better mood. For non‐visual functions, strongcorrelationswerefoundbetweenmorerelaxationandbetterphysicalwell‐being;bothwerealsocorrelatedwithabettermoodandhigheralertness.

7.1.3 Impactofluminousdistributiononvisualandnon‐visualfunctions

TheCLLSwasappliedtoassesstheimpactofluminousdistributiononsubjectiveassessments.Glareindicesandbackgroundluminanceswereusedtodeterminetheluminousdistributionofascene. Luminance mappings issued from a conventional CCD camera and HDR imagingtechniqueswereemployedinthefirststudy;thesecondusedluminancemappingsachievedbythe CLLS. Sequences of images were recorded over time for both studies; the impact of theluminousdistributiononvisualandNIFeffectswereinvestigated.

Significant correlations between subjective and objective glare assessments were found,especiallyunderbrightlightconditions.However,itmustbeoutlinedthatunderrealisticofficelighting,negativeobjectiveglareratings(largerglareindices)donotalwaysleadtopoorvisualcomfortsubjectiveassessments.Preferredlightingconditions,accordingtothesubjects’choice,mayprevent against intolerable glare ratings, even if the objective assessments donot differ.

Conclusion

132

Whether light preference bias and/or the spectrum of light influenced the subjective visualcomfortassessments,couldnotbedeterminedinthisstudy.

A novel finding from this thesis is the association between luminous distributions and non‐visualfunctions,suchthathighobjectiveglareratingswereassociatedwithlessrelaxation,poorphysicalwell‐beingandmood,aswellasloweralertness.Preferredlightingconditionsseemedto mitigate the negative appraisals of these non‐visual functions, as was shown in extremechronotypes.

It remains necessary to further investigate the variations of discomfort glare over time tointegrate results frombothobjective glare ratings and subjective assessments.Both variablesvariedduring thedaydue todaylight fluctuationsaswell as thecircadianmodulationsofourinternal clock. To determine optimal lighting conditions, the modulations of environmentalfactorsandinternalphysiologicalstatesneedtobeconsideredaccordingly.

Novelfindingsregardingvisualcomfort

In the two studies, subjective visual comfort assessments showedvery high correlationswithlight preference; they had a similar time course and decreased during the day. A moderatecorrelationbetweenvisualcomfortandsubjectiveglareratingswasalsofound.Visualcomfortassessments showed also slight but significant correlations with colour appearance andbrightnessperceptionof light,aswellasglareindexes(DGP,DGIandUGR).Thiswasalsotruefor non‐visual functions such as alertness, well‐being and mood: correlations between thosevisualcomfortvariablesandnon‐visualfunctionswerefound.

To conclude thismeans that a positive appraisal of visual comfort does not always imply anabsenceofdiscomfortglare,ascommonlydefined.Apartfromtheabsenceofluminoussourcesleading to discomfort glare, lighting preference as well the other visual comfort variables,includingnon‐visualfunctions,arealsoinvolvedinthesubjectiveassessmentofvisualcomfortforgivenlightingconditions.

7.2. Recommendationsforofficelighting

Theintensiveuseofdaylightingwithinbuildingscanbepromotedasapreferablelightsourceinoffice rooms,notonly forenergy savingpurposes,but also for the sakeof visual comfortandnon‐visualfunctions.Accordingtothestudiescarriedoutinthisthesis,lightaffectsbothvisualandnon‐visualfunctionsunderrealisticlightingconditionsOffice rooms with daylight input provide better visual comfort, light preference and lesssubjective glare. These office lighting conditions can also prevent sleepiness and discomfortglareinthecourseoftheafternoon.Moreover,individuallycontrolledlightingsystemsplayanimportant role regarding subjective assessments of visual comfort by users. The individualchoiceof lightingscenariosdoesnotonlyprovideapreferredlightingenvironmentandbettervisualcomfortconditions,italsoreducesnegative(subjective)assessmentsofdiscomfortglare.

Conclusion

133

7.3. Futureoutlook

This thesis pointed out some additional open questions. Further research is required toinvestigateandprovideknowledgeaboutthefollowing.

Assessmentofdaylightingandelectriclightingsystemsincircadianmetrics

For the first time, a successful implementation of the CLLS for circadian weighted radiancemappingofadvanceddaylightingsystemswasachievedinthisthesis. Itshowedthatdifferentdaylighting systems provide different luminous distributions at different times of day due todifferentdaylightinputfluxes.Interestingly,thedifferentdaylightincidenceanglesdidnotonlyprovidedifferent illuminance levels, but alsomodified the spectral compositionof daylight. Itwouldbe interesting to assess further the circadianweighted radiancemapsof other lightingsystems ‐ combining daylighting and electrical lighting systems ‐ at different times of day inordertoassessthedifferenteffectsoflightwithrespecttonon‐visualfunctions.

UseofCLLSforcircadianmetricsassessmentofvisualcomfortandnon‐visualfunctions

Photometric properties, such as the colour appearance (CCT and CRI) were correlated withsubjectiveassessmentsofvisualcomfort.Thismightthereforebeaparameterofvisualcomfort,liketheglare index,whichcanbeanalysed incombinationwith luminousdistributions.Basedonthefindingsfromthesecondstudy,circadianweightedirradiance(Eec)showedcorrelationswith several visual comfort variables under various lighting conditions. Since the CLLS wassuccessfullycalibratedandvalidatedincircadianmetrics,itwouldbeinterestingtoimplementitinastudyinvolvingacomparisonwithsubjectiveassessments.

Furtherobjectiveassessmentsofnon‐visualfunctionsandcognitiveperformance

Accordingtothisthesis,impactsoflightwerefoundonnon‐visualfunctions,suchasalertness,moodandwell‐being.Besidesthesesubjectiveassessments,interestingobjectiveassessmentsofnon‐visual functionswerealsomonitored, such as thehormonal onsetofmelatonin secretionandcortisol salivaryconcentrationsaswellasotherphysiologicalvariables thatarecurrentlybeinganalysed.Moreover,performancesofvisualandcognitivetasksmustalsobeconsideredtoinvestigate the effects of light. Thismight further explain the luminous impacts on visual andnon‐visualfunctionsandmayallowanimprovementoftaskperformancerelatedvariablesinarealisticworkingenvironment.

Conclusion

134

135

References

1. Roulet, C.‐A., et al., Perceivedhealthandcomfort inrelation toenergyuseandbuildingcharacteristics.BuildingResearchandInformation,2006.34(5):p.467‐474.

2. Fisk, W.J. Review of health and productivity gains from better IEQ. in Proceedings ofHealthyBuildings2000.2000.Helsinki,Finland:SIYIndoorAirInformation.

3. Singh,A., et al.,Effectsofgreenbuildingsonemployeehealthandproductivity.AmericanJournalofPublicHealth,2010.100(9):p.1665‐1668.

4. Carlopio,J.R.,Constructvalidityofaphysicalworkenvironmentsatisfactionquestionnaire.JournalofOccupationalHealthPsychology,1996.1(3):p.330‐344.

5. Veitch, J.A., et al., Amodel of satisfactionwith open‐plan office conditions: COPE fieldfindings.JournalofEnvironmentalPsychology,2007.27(3):p.177‐189.

6. Harter, J.K., F.L. Schmidt, and T.L. Hayes, Business‐unit‐level relationship betweenemployee satisfaction, employee engagement, and business outcomes: A meta‐analysis.JournalofAppliedPsychology,2002.87(2):p.268‐279.

7. De Croon, E.M., et al., Effect of office concepts on worker health and performance.Ergonomics,2005.48(2):p.119‐134.

8. Veitch, J.A., et al. Satisfactionwith lighting in open‐plan offices:COPE field findings. inProceedingsofLUXEUROPA2005.2005.Berlin,Germany.

9. Newsham, G.R., et al., Linking indoorenvironmentconditions to job satisfaction:A fieldstudy.BuildingResearchandInformation,2009.37(2):p.129‐147.

10. Karcher, A., et al.,LightPerspectives. between culture and technology. 2009, Germany:TimHenrikMaack,KayPawlik.

11. Lewy, A.J., et al., Light suppresses melatonin secretion in humans. Science, 1980.210(4475):p.1267‐1269.

12. Cajochen, C., et al.,Highsensitivityofhumanmelatonin,alertness,thermoregulation,andheart rate to shortwavelength light. Journal of Clinical Endocrinology & Metabolism2005.90(3):p.1311‐1316.

13. Lucas, R.J., R.H. Douglas, and R.G. Foster, Characterization of an ocular photopigmentcapableofdrivingpupillaryconstrictioninmice.NatureNeuroscience,2001.4(6):p.621‐626.

14. Foster,R.G.,Sheddinglightonthebiologicalclock.Neuron,1998.20(5):p.829‐832.15. Czeisler,C.A.,etal.,Brightlightresetsthehumancircadianpacemakerindependentofthe

timingofthesleep‐wakecycle.Science,1986.233(4764):p.667‐671.16. VanBommel,W.J.M.Dynamiclightingatwork–bothinlevelandcolour. inProceedingof

the 2nd CIE Expert Symposium on Lighting and Health, 2006. 2006. Braunschweig,Germany.

17. Boyce, P.R., et al., Lighting quality and officework: Two field simulation experiments.LightingResearchandTechnology,2006.38(3):p.191‐223.

18. Veitch, J.A., et al., Lightingappraisal,well‐beingandperformance inopen‐planoffices:Alinkedmechanismsapproach. Lighting Research and Technology, 2008. 40(2): p. 133‐151.

19. Bean, A.R. and T.M. McFadder, New developments in CSP (Comfort, Satisfaction andPerformance).LightingResearchandTechnology,1998.30(3):p.126‐132.

References

136

20. Bean, A.R. and R.I. Bell, TheCSP index:Apracticalmeasureofoffice lightingqualityasperceivedbytheofficeworker. LightingResearch andTechnology, 1992.24(4): p. 215‐225.

21. Boyce,P.R.,HumanFactorsinLighting.2003,London:Taylor&Francis.22. Rea,M.S.,IESNALightingHandbook:Reference&Applicationthe9thed.2000,NewYork:

IlluminatingEngineeringSocietyofNorthAmerica23. Hopkinson,R.G.,Glarefromdaylightinginbuildings. AppliedErgonomics 1972.3(4): p.

206‐215.24. Borisuit, A., et al.Effectsoftwodifferentofficelightingconditionsonsubjectivealertness

andvisualcomfortinyoungvolunteers. inDesign&HealthWorldCongress2011.Boston,US.

25. Zhang, L., M. Helander, and C.G. Drury, Identifying factorsofcomfortanddiscomfort inseating.HumanFactors,1996.38(3):p.377‐389.

26. Roulet, C., Santéetqualitéde l'environnement intérieurdans lesbâtiments. 2nd Edition.2008:PPURpressespolytechniques.

27. Slater,K.,HumanComfort.1985,Springfield,IL:ChalesC.Thomas.28. De Looze,M.P., L.F.M.Kuijt‐Evers, and J. VanDieën,Sittingcomfortanddiscomfortand

therelationshipwithobjectivemeasures.Ergonomics,2003.2003(46):p.10.29. Roufs, J.A.J. and M.C. Boschman, Visualcomfortandperformance, in TheMan‐Machine

Interface,J.A.J.Roufs,Editor.1991,MacMillanPress:London.30. Balder, J.J.,ErwünschteLeuchtdichten inBüroräumen. Lichttechnik, 1957.9(9): p. 455‐

461.31. Muck, E. and H.W. Bodmann, Die Bedeutung Des Beleuchtungsniveau Bei Praktischer

Sehtatigkeit.Lichttechnik,1961.13:p.502‐508.32. Pellegrino,A.,Assessmentofartificiallightingparametersinavisualcomfortperspective.

LightingResearchandTechnology,1999.31(3):p.107‐115.33. Veitch,J.A.,Psychologicalprocessesinfluencinglightingquality.JournaloftheIlluminating

EngineeringSocietyofNorthAmerica,2001.30(1):p.124‐140.34. Kim, S.‐Y.K., J‐J, Influenceoflightfluctuationonoccupantvisualperception. Buildingand

Environment,2007.42(8):p.2888‐2899.35. Osterhaus, W. Large area glare sources and their effect on discomfort and visual

performanceatcomputerworkstations. in1992IEEEIndustryApplicationsSocietyAnnualMeeting.1992.Houston,TX.

36. Veitch, J.A. and G.R. Newsham, Preferred luminous conditions in open‐plan offices:Researchandpracticerecommendations.LightingResearchandTechnology,2000.32(4):p.199‐212.

37. Veitch,J.A.andG.R.Newsham,Exercisedcontrol,lightingchoices,andenergyuse:anofficesimulationexperiment.JournalofEnvironmentalPsychology,2000.20(3):p.219‐237.

38. Rea, M.S., Toward a model of visual performance: foundations and data. Journal ofIlluminatingEngineeringSociety,1986.15(2):p.41‐57.

39. Velds, M., Assessment of lighting quality in office roomwith daylighting systems 2000,DelftUniversityoftechnology.p.209.

40. La Toison, M., Perceived contrasts and visual comfort in interior lighting. LightingResearchandTechnology,1997.29(2):p.80‐82.

41. Bodmann,H.W.,P.Haubner,andA.M.Marsden.Aunifiedrelationshipbetweenbrightnessandluminance.inCIE19thSession.1979.Kyoto,Japan.

42. CommissionInternationaledel’Eclairage(CIE),InternationalLightingVocabulary.1970,CIEPublication.

43. Commission Internationale de l’Eclairage (CIE), Method ofMeasuring and SpecifyingColourRenderingProperties of Light Sources, in CIE013.3. 1995, CIE Central Bureau:Vienna.

44. Winzen, J., F. Albers, and C. Marggraf‐Micheel, The influence of coloured light in theaircraftcabinonpassenger thermalcomfort. Lighting Research and Technology, 2013.online(June13,2013):p.1‐11.

References

137

45. Chauvel , P., et al.,Glarefromwindows:currentviewsoftheproblem. LightingResearchandTechnology,1982.14(1):p.31‐46.

46. Aries, M.B.C., J.A. Veitch, and G.R. Newsham,Windows, view,andoffice characteristicspredict physical and psychological discomfort. Journal of Environmental Psychology,2010.30(4):p.533‐541.

47. Markus, T.A.,Thefunctionofwindows‐areappraisal. Building Science, 1967.2: p. 97 ‐121.

48. Roche, L., E. Dewey, and P. Littlefair,Occupantreactionstodaylight inoffices. LightingResearchandTechnology,2000.32(3):p.119‐126.

49. Tuaycharoen, N. and P.R. Tregenza, Viewanddiscomfortglare fromwindows. LightingResearchandTechnology,2007.39(2):p.185‐200.

50. Tuaycharoen, N. and P.R. Tregenza, Discomfortglare from interesting images. LightingResearchandTechnology,2005.4(37):p.329‐338.

51. CommissionInternationaledel’Eclairage(CIE),CIECollectiononglare.2002,CIE.52. European committee for standardization (CEN), EN12665:Lightand lighting in Basic

termsandcriteriaforspecifyinglightingrequirements.2002:Brussels.53. Begemann,S.H.A.,G.J.VandenBeld,andA.D.Tenner,Daylight,artificiallightandpeople

inanofficeenvironment,overviewofvisualandbiologicalresponses.InternationalJournalofIndustrialErgonomics,1997.20(3):p.231‐239.

54. Junlén, H.T., M.C.H.M. Wouters, and A.D. Tenner, Preferred task‐lighting levels in anindustrialworkareawithoutdaylight.LightingResearchandTechnology,2005.37(3):p.219‐233.

55. Moore, T., D.J. Carter, and A.I. Slater, Long‐termpatternsofuseofoccupantcontrolledofficelighting.LightingResearchandTechnology,2003.35(1):p.43‐59.

56. Escuyer,S.andM.Fontoynont,Lightingcontrols:afieldstudyofofficeworkers'reactions.LightingResearchandTechnology,2001.33(2):p.77‐96.

57. European Committee for Standardization (CEN), EN 12464‐1: Light and lighting., inLightingofworkplaces.2003:Brussel.

58. Fotios, S.A. and C. Cheal, Stimulus range bias explains the outcome of preferred‐illuminanceadjustments.LightingResearchandTechnology,2010.42(4):p.433‐447.

59. Scholz, A., et al.,Minimumandoptimum lightoutputofMacintosh size3 laryngoscopyblades:amanikinstudy.Anaesthesia,2007.62(2):p.163‐168.

60. Boyce,P.R.,N.H.Eklund,andS.N.Simpson,Individuallightingcontrol:Taskperformance,moodandilluminance. Journal of the Illuminating Engineering Society, 2000.29(1): p.131‐142.

61. Boyce, P.R., et al., Occupant use of switching and dimming controls in offices. LightingResearchandTechnology,2006.38(4):p.358‐376.

62. Begemann,S.H.A.,G.J.VandenBeld,andA.D.Tenner.Daylight,artificiallightandpeople,Part2.inProceedingsof23rdSessiontheCIE.1995.NewDelhi.

63. Logadóttir,A.,J.Christoffersen,andS.A.Fotios,Investigatingtheuseofanadjustmenttaskto set the preferred illuminance in a workplace environment. Lighting Research andTechnology,2011.43(4):p.403‐422.

64. Kruithof, A.A., Tubular luminescence lamps for general illumination. Philips TechnicalReview,1941.6(2):p.65‐73.

65. Davis, R.G. and D.N. Ginthner, Correlated color temperature, illuminance level and theKruithofcurve.JournaloftheIlluminatingEngineeringSociety,1990.19(1):p.27‐38.

66. Boyce, P.R. and C. Cuttle, Effect of correlated colour temperature on the perception ofinteriors and colour discrimination performance. Lighting Research and Technology,1990.22(1):p.19‐36.

67. Veitch, J.A. and S.L. Mc Coll, A critical examination of perceptual and cognitive effectsattributedtofull‐spectrumfluorescentlighting.Ergonomics,2001.44(3):p.255‐279.

68. McNelis, J.F.H., J G, G.E. Dore, andW.B. DeLaney, Subjectiveappraisalofcoloredscenesundervariousfluorescentlampcolors.LightingDesign+Application,1985.15(6):p.25‐29.

References

138

69. Fotios, S.A. and C. Cheal, Predicting lamp spectrum effects at mesopic levels. Part2:Preferredappearanceandvisualacuity.LightingResearchandTechnology,2011.43(2):p.159‐172.

70. Cuttle,C.Peopleandwindowsinworkplaces.inPeopleandPhysicalEnvironmentResearchConference.1983.Wellington,NewZealand.

71. Heerwagen,J.H.andD.R.Heerwagen,Lightingandpsychologicalcomfort.LightingDesignandApplication,1986.16(4):p.47‐51.

72. Veitch, J.A., D.W. Hine, and R. Gifford, End‐users'knowledge,beliefsandpreferences forlighting.JournalofInteriorDesign,1993.19(2):p.15‐26.

73. Veitch, J.A. andR.Gifford,Assessingbeliefsaboutlightingeffectsonhealth,performance,mood,andsocialbehavior.EnvironmentandBehavior,1996.28(4):p.446‐470.

74. Cuttle,C.Peopleandwindowsinworkplaces.inPeopleandPhysicalEnvironmentResearchConference.1983.Wellington,NewZealand.

75. Christoffersen, J. and K. Johnsen.Windowsanddayligh.Apost‐occupancyevaluationofDanishoffices.inLighting2000CIBSE/ILEJointconference.2000.UniversityofYork,UK.

76. Galasiu, A.D. and J.A. Veitch, Occupant preferences and satisfactionwith the luminousenvironment and control systems in daylit offices: a literature review. Energy andBuildings,2006.38:p.728‐742.

77. Duffy, J.F. and C.A. Czeisler, Age‐related change in the relationship between circadianperiod,circadianphase,anddiurnalpreference inhumans. Neuroscience Letters, 2002.318(3):p.117‐120.

78. Newsham, G.R., et al., Individual control of electric lighting in a daylit space. LightingResearchandTechnology,2008.40(1):p.25‐41.

79. DeKort, Y.A.W.S.,KCH J,Effectsofdynamiclightingonofficeworkers:Firstresultsofafield studywithmonthlyalternating settings. Lighting Research and Technology, 2010.42(3):p.345‐360.

80. Akashi, Y. and P.R. Boyce,A fieldstudyof illuminancereduction. Energy and Buildings,2006.38(6):p.588‐599.

81. Fotios, S.A., Lamp colour properties and apparent brightness. Lighting Research andTechnology,2001.33(3):p.163‐181.

82. Flynn, J.E., et al.The influenceofspatial lightonhuman judgment. in ProceedingofCIE18thSession.1975.London,UK.

83. Tiller, D.K. and J.A. Veitch, Percieved room brightness: Pilot study on the effect ofluminancedistribution.LightingResearchandTechnology,1995.27(2):p.93‐101.

84. Berman, S.M., et al., Photopic luminance does not always predict perceived roombrightness.LightingResearchandTechnology,1990.22(1):p.37‐41.

85. Boyce, P.R., et al., The impact of spectral power distribution on theperformance ofanachromaticvisualtask.LightingResearchandTechnology,2003.35(2):p.141‐161.

86. Boyce, P.R., Investigationsofthesubjectivebalancebetweenilluminanceandlampcolourproperties.LightingResearchandTechnology,1977.9(1):p.11‐24.

87. Boyce, P.R. and C. Cuttle, Effect of correlated colour temperature on the perception ofinteriorsandcolourdiscrimination. Lighting Research and Technology, 1990.22(1): p.19‐36.

88. Viénot, F., M.L. Durand, and E. Mahler, Kruithof'srulerevisitedusingLED illumination.JournalofModernOptics,2009.56(13):p.1433‐1446.

89. Commission Internationale de l’Eclairage (CIE), Recommended System for VisualPerformanceBasedMesopicPhotometry,inCIEPublication191.2010,CIE:Vienna.

90. Rea, M.S., L.C. Radetsky, and J.D. Bullough, Toward amodel of outdoor lighting scenebrightness.LightingResearchandTechnology,2011.43(1):p.7‐30.

91. Bullough,J.D.,L.C.Radetsky,andM.S.Rea,Testingaprovisionalmodelofscenebrightnesswithandwithoutobjectsofdifferentcolours. Lighting Research and Technology, 2011.43(2):p.173‐184.

92. Hurden,A.,etal.,Visualperformanceatmesopiclightlevels:Anempiricalmodel.LightingResearchandTechnology,1999.31(3):p.127‐131.

References

139

93. Rea, M.S., Visual performance with realisticmethods of changing contrast. Journal ofIlluminatingEngineeringSociety,1981.10(3):p.164‐177.

94. Rea, M.S., Toward a model of visual performance: Foundations and data. Journal ofIlluminatingEngineeringSociety,1986.15(2):p.41‐58.

95. Rea,M.S. andM.J.Ouellette,Visualperformanceusingreactiontimes. LightingResearchandTechnology,1988.20(4):p.139‐153.

96. Rea, M.S. and M.J. Ouellette, Relativevisualperformance:Abasis forapplication. 1991.23(3):p.135‐144.

97. Eklund,N.H.,P.R.Boyce,andS.N.Simpson,Lightingandsustainedperformance:Moelingdata‐entry task performance. Journal of the Illuminating Engineering Society, 2001.30(2):p.126‐141.

98. Bailey, I.L.,R.Clear,andS.M.Berman,Sizeasadeterminantofreadingspeed. JournaloftheIlluminatingEngineeringSociety,1993.22(2):p.102‐117.

99. Veitch, J.A. and S.L.Mc Coll,Modulationoffluorescentlight:flickerrateandlightsourceeffects on visual performance and visual comfort. Lighting Research and Technology,1995.27(4):p.231‐241.

100. Rowlands, E., et al. Visual performance in illuminance of different spectral quality. inProceedingsoftheCIE17thSession.1971.Barcelona,Spain.

101. Smith, S.W. and M.S. Rea. Relationshipsbetweenoffice taskperformanceandratingsoffeelingsandtaskevaluationsunderdifferentlightsourcesandlevel. inProceedingof19thCIESession.1979.Kyoto,Japan.

102. Berman, S.M., et al., Luminance‐controlledpupilsizeaffectsLandoltCtaskperformance.Journalof the IlluminatingEngineeringSocietyofNorthAmerica,1993.22(2):p.150‐165.

103. Berman, S.M., et al., Landolt‐C recognition in elderly subjects is affected by scotopicintensityofsurroundilluminants.JournaloftheIlluminatingEngineeringSocietyofNorthAmerica,1994.23(2):p.123‐130.

104. Berman, S.M., Energy efficiency consequences of scotopic sensitivity. Journal of theIlluminatingEngineeringSocietyofNorthAmerica,1992.21(1):p.3‐14.

105. Berman, S.M., et al.,Despitedifferentwallcolours,verticalscotopic illuminancepredictspupilsize.JournaloftheIlluminatingEngineeringSocietyofNorthAmerica,1997.26(2):p.59‐68.

106. Rea, M.S.,Disscussionafter"Modulationof fluorescent light: flickerrateand lightsourceeffectsonvisualperformanceandvisual comfort"byVeitch, JAandMcColl,SL. LightingResearchandTechnology,1995.27(4):p.254‐255.

107. Kawasaki,A.andR.H.Kardon,Intrinsicallyphotosensitiveretinalganglioncells.JournalofNeuro‐Ophthalmology,2007.27(3):p.195‐204.

108. Gamlin, P.D.R., et al., Human and macaque pupil responses driven by melanopsin‐containingretinalganglioncells.VisionResearch,2007.47(7):p.946‐954.

109. Boyce,P.R.,C.M.Hunter,andO.Howlett,Thebenefitsofdaylightthroughwindows.2003,Lightingresearchcenter,RensselaerPolytechnicInstitute:Troy,NewYork.

110. Figueiro, M., et al. Daylightandproductivity:apossible link to circadian regulation. inProceedingsof theFifth InternationalLROLightingResearchSymposium. 2002. ElectricPowerResearchInstitute:PaloAlto,CA.

111. Czeisler, C.A., et al., Stability,precisionandnear‐24‐hourperiodofthehumancircadianpacemaker.Science,1999.284(5423):p.2177‐2181.

112. Goodman, T.M.,Measurementandspecificationoflighting:A lookatthefuture. LightingResearchandTechnology,2009.41:p.229‐243.

113. CommissionInternationaledel’Eclairage(CIE),Lightasatruevisualquantity:Principlesofmeasurement,inCIE041‐1978.1978,Commissioninternationaldel'éclairage:Vienna.

114. Eloholma,M., et al.,Mesopicmodels—frombrightnessmatchingtovisualperformanceinnight‐timedriving:Areview.LightingResearchandTechnology,2005.37(2):p.155‐175.

115. Gall,D.,CircadianeLichtgrössenundderenmesstechnischeErmittlung.Licht,2002.11‐12:p.1292‐1297.

References

140

116. Berson,D.M.,F.A.Dunn,andM.Takao,Phototransductionbyretinalganglioncellsthatsetthecircadianclock.Science,2002.295(5557):p.1070‐1073.

117. Ruby, N.F., T.J. Brennan, and X. Xie, Roleofmelanopsin incircadian responses to light.Science,2002.298(5601):p.2211‐2213.

118. Brainard, G.C., et al.,Actionspectrumformelatoninregulationinhumans:evidenceforanovelcircadianphotoreceptor.JournalofNeuroscience,2001.21(16):p.6405‐6412.

119. Thapan, K., J. Arendt, and D.J. Skene, An action spectrum formelatonin suppression:evidence foranovelnon‐rod,non‐conephotoreceptor system inhumans. The Journal ofPhysiology,2001.535(1):p.261‐267.

120. Rea, M.S., Phototransduction for human melatonin suppression. Journal of PinealResearch,2002.32(4):p.209‐213.

121. Bellia, L. and F. Bisegna, From radiometry to circadian photometry: A theoreticalapproach.BuildingandEnvironment,2013.62:p.63‐68.

122. Gall, D. Die Messung circadianer Strahlungsgrößen. in Proceeding of the Licht undGesundheit2004.Berlin.

123. Gall,D.andV.Lapuente,BeleuchtungsrelevanteAspektebeiderAuswahleinesförderlichenLampenspektrums.Licht,2002.54(7):p.860‐871.

124. Rea, M.S., et al., Amodel of phototransduction by the human circadian system. BrainResearchReviews,2005.50(2):p.213‐228.

125. Rea,M.S.,etal.,Circadianlight.JournalofCircadianRhythms,2010.8(2):p.1‐10.126. Kozakov, R., S. Franke, andH. Schöpp,Approachtoaneffectivebiologicalspectrumofa

light source. LEUKOS ‐ the Journal of the Illuminating Engineering Society of NorthAmerica,2008.4(4):p.255‐263.

127. al Enezi, J., et al., A “melanopic” spectral efficiency function predicts the sensitivity ofmelanopsinphotoreceptorstopolychromaticlights. JournalofBiologicalRhythms,2011.26(4):p.314‐323.

128. Brown, T., et al., The melanopic sensitivity function accounts for melanopsin‐drivenresponsesinmiceunderdiverselightingconditions.PLoSOne,2013.8(1):p.e53583.

129. DIN, BiologischwirksameBeleuchtung ‐Planungsempfehlungen, in BiologicallyEffectiveIllumination‐DesignGuidelines.2013,DIN:Germany.

130. Bierman, A., T.R. Klein, and M.S. Rea, TheDaysimeter:A device formeasuring opticalradiation as a stimulus for the human circadian system. Measurement Science andTechnology,2005.16(11):p.2292–2299.

131. Figueiro,M., et al.,Comparisonsofthreepracticalfielddevicesusedtomeasurepersonallight exposuresandactivity levels. Lighting Research and Technology, 2013. 45(4): p.421‐434.

132. Rea,M.S., et al.,Anewapproachtounderstandingthe impactofcircadiandisruptiononhumanhealth.JournalofCircadianRhythms,2008.6:p.7.

133. Hubalek, S., D. Zöschg, and C. Schierz, Ambulant recordingof light forvisionandnon‐visualbiologicaleffects.LightingResearchandTechnology,2006.38(4):p.314‐324.

134. Hubalek, S., M. Brink, and C. Schierz, Office workers' daily exposure to light and itsinfluenceonsleepqualityandmood. LightingResearchandTechnology, 2010.42(2):p.33‐50.

135. Philips Healthcare. Actiwatch,http://www.healthcare.philips.com/main/homehealth/sleep/actiwatch/default.wpd.2013 [cited 2013, June]; Available from:http://www.healthcare.philips.com/main/homehealth/sleep/actiwatch/default.wpd.

136. Thorne, H.C., et al., Daily and seasonal variation in the spectral composition of lightexposureinhumans.ChronobiologyInternational.26(5):p.854‐866.

137. Price,L.L.A.,M.Khazova,andJ.O'Hagan,Performanceassessmentofcommercialcircadianpersonalexposuredevices.LightingResearchandTechnology,2012.44(1):p.17‐26.

138. Commission Internationale de l’Eclairage (CIE), CIE Technical support, in Methods ofcharacterizing illuminancemeters and luminancemeters ‐ Performance, characteristicsandspecifications.1987,CIECentralbureau:Vienna.

References

141

139. Manav, B.,Anexperimentalstudyontheappraisalofthevisualenvironmentatoffices inrelationtocolourtemperatureandilluminance.BuildingandEnvironment,2007.42(2):p.979‐983.

140. Inanici,M.,Evaluationofhighdynamicrangephotographyasaluminancedataacquisitionsystem.LightingResearchandTechnology,2006.38(2):p.123‐134.

141. Beltrán, L.O. and B.M. Mogo. Assessment of luminance distribution using HDRphotography.inISESSolarWorldCongress.2005.Orlando,FL,USA.

142. Anaokar , S. and M. Moeck Validation ofHighDynamic Range Imaging to LuminanceMeasurement. LEUKOS ‐ the Journal of the Illuminating Engineering Society of NorthAmerica,2005.2(2):p.133‐144.

143. Borisuit, A., et al.Comparisonofobjectiveandsubjectivevisualcomfortandassociationswithnon‐visual functions inyoungsubjects. in ProceedingsofCISBAT2011InternationalScientificConference.2011.Lausanne,Switzerland.

144. Hopkinson, R.G. and J.B. Collins, The Ergonomics of Lighting. 1970, London, UK:MacDonald.

145. Nazzal,A.A.,Anewevaluationmethodfordaylightdiscomfortglare. InternationalJournalofIndustrialErgonomics,2005.35(4):p.295‐306.

146. Fisekis, K., et al., Prediction ofdiscomfort glare fromwindows. Lighting Research andTechnology,2003.35(4):p.360‐369.

147. Osterhaus,W.K.E.,Discomfortglareassessmentandpreventionfordaylightapplicationsinofficeenvironments.SolarEnergy,2005.79(2):p.140‐158.

148. Wienold, J. and J. Christoffersen, Evaluationmethods and development of a new glarepredictionmodel for daylight environmentswith the use of CCD cameras. Energy andBuildings,2006.38(7):p.743‐757.

149. Jakubiec, J.A. and C.F. Reinhart,The 'adaptivezone' ‐Aconcept forassessingdiscomfortglarethroughoutdaylitspaces. LightingResearchandTechnology,2012.44(2):p. 149‐170.

150. Bellia,L.,M.Marilena, andS.Gennaro, IlluminancemeasurementsthroughHDRimagingphotometryinscholasticenvironment.Energyandbuildings,2011.43(10):p.2843‐2849.

151. Borisuit, A., J.‐L. Scartezzini, and A. Thanachareonkit, Visual comfort and glare riskassessmentbyHDRimagingtechnique.ArchitecturalScienceReview,2010.53(4):p.359‐373.

152. Suk, J. andM. Schiler, InvestigationofEvalglaresoftware,daylightglareprobabilityandhigh dynamic range imaging for daylight glare analysis. Lighting Research andTechnology,2013.45(4):p.450‐463.

153. Commission Internationalede l’Eclairage (CIE),Discomfortglareintheinteriorlighting,inCIETechnicalReport117,C.I.d.l'Eclairage,Editor.1995,CIE.

154. Larson, G.W. Radiance. 1982 [cited 2008; Available from:http://radsite.lbl.gov/radiance.

155. Boff, K.R. and J.E. Lincoln, Engineering Data Compendium. Human Perception andPerformance:Volume2.1988,Ohio:Wright‐PattersonAirForceBase.

156. Reinhard, E., et al., Highdynamic range imaging,acquisition,display,and image‐basedlighting.2006,SanFrancisco,CA:Elsevier.

157. Debevec, P.E. and J. Malik. Recovering high dynamic range radiance maps fromphotographs. inSIGGRAPH:ConferenceonComputerGraphicsandInteractiveTechniques.1997.LosAngeles.

158. Larson, G.W. [Radiance‐general] High dynamic range image builder.http://www.jaloxa.eu/mirrors/radiance_mailinglists/general/2003‐October/001114.html2003[cited2013,April].

159. Larson, G.W. and R.A. Shakespeare, Renderingwith Radiance: the Art and Science ofLightingVisualization.1998,SanFrancisco,CA:MorganKaufmannPublishers.

160. Larson,G.W.AnyhereSoftware.Photospherev1.8.7U2011[cited2011,May];Availablefrom:http://www.anyhere.com.

References

142

161. Jacobs, A.WebHDR. http://www.jaloxa.eu/webhdr/software.shtml 2012 [cited 2012,February].

162. Frauhofer, I.S.E. Evalglare v1.04. http://www.ise.fraunhofer.de/areas‐of‐business‐and‐market‐areas/applied‐optics‐and‐functional‐surfaces/lighting‐technology/lighting‐simulations/radiance/radiance?set_language=en&cl=en 2012 2012 [cited 2012February]; Available from: http://www.ise.fraunhofer.de/areas‐of‐business‐and‐market‐areas/applied‐optics‐and‐functional‐surfaces/lighting‐technology/lighting‐simulations/radiance/radiance?set_language=en&cl=en.

163. Arm, C., et al., Icycam ‐ aHighDynamic Range Vision System, in CSEM Scientific andTechnicalReport.2008,CSEM.p.20.

164. Rüedi, P.F., et al. An SoC combining a 132dB QVGA pixel array and a 32b DSP/MCUprocessorforvisionapplication.in2009IEEEInternationalSolid‐StateCircuitsConference.2009.SanFrancisco,USA

165. Pangaud, S., Description de la télémétrie de l’application embarquée « viewer v3.5b »,DEVISE‐IcyCAMVSLOG.2011,CentreSuissed'électroniqueetdemicrotechnique(CSEM):Neuchâtel.

166. Fujinon,FE185C057HA‐1,inFish‐eyeforsecurity,Fujifilm,Editor.2010.167. Borisuit, A., et al.Anewdevicefordynamic luminancemappingandglareassessmentin

buildings.inSPIEOptics+PhotonicsConference2012.2012.SanDiego.168. Andersen, M., Innovativebidirectionalvideo‐goniophotometer foradvanced fenestration

systems, in Solar Energy and Building Physics Laboratory. 2004, Ecole PolytechniqueFédéraledeLausanne(EPFL):Lausanne.

169. CommissionInternationaledel’Eclairage(CIE),CommissionInternationaledel’ÉclairageProceedings1932,Cambridge:CambridgeUniversityPress.

170. Deutsches Institut für Normung, DIN 5032, Part 7: Class B, in LichtmessungKlasseneinteilung von Beleuchtusstärke‐ und Leuchtdichtemessgeräten. [Photometry:classificationofilluminancemetersandluminancemeters].1985.

171. Borisuit,A.,etal.Assessmentofcircadianweightedradiancedistributionusingacamera‐likelightsensor. inProceedingsofCISBAT2013InternationalScientificConference.2013.Lausanne,Switzerland.

172. Hattar, S., et al.,Melanopsin‐containing retinal ganglion cells: architecture, projections,andintrinsicphotosensitivity.Science,2002.295(5557):p.1065‐70.

173. Provencio, I., H.M. Cooper, and R.G. Foster, Retinal projections inmicewith inheritedretinal degeneration: Implications for circadian photoentrainment. Journal ofComparativeNeurology,1998.395(4):p.417‐439.

174. Zaidi, F.H., et al., Short‐wavelength light sensitivity of circadian, pupillary, and visualawareness inhumans lackinganouterretina. Current Biology, 2007. 17(24): p. 2122‐2128.

175. Wehr, T.A., Thedurationofhumanmelatoninsecretionandsleeprespond tochanges indaylength(photoperiod).JournalofClinicalEndocrinology&Metabolism,1991.73(6):p.1276‐1280.

176. VanBommel,W.J.M.,Non‐visualbiologicaleffectoflightingandthepracticalmeaningforlightingforwork.AppliedErgonomics,2006.37:p.461‐466.

177. Altun,A. andB.Ugur‐Altun,Melatonin:therapeuticandclinicalutilization. InternationalJournalofClinicalPractice,2007.61(5):p.835‐845.

178. Aschoff, J., Handbook of Behavioral Neurobiology, Biological Rhythms. Vol. Biologicalrhythms.1981,NewYork:PlenumPress.

179. Wetterberg,L.,LightandBiologicalRhythmsinMan.1993,Stockholm:PergamonPress.180. Wurtman, R.J., M.J. Baum, and J.T.J. Potts, TheMedical andBiologicalEffects of Light.

1985,NewYork:TheNewYorkAcademyofSciences.181. Honma,K.,S.Honma,andT.Wada,Phase‐dependentshiftoffree‐runninghumancircadian

rhythmsinresponsetoasinglebrightpulse.Experientia,1987.43(11‐12):p.1205‐1207.182. Lewy,A.J.,etal.,Antidepressantandcircadianphase‐shiftingeffectsoflightScience,1987.

235(4786):p.352‐354.

References

143

183. Khalsa,S.B.S.,etal.,Aphaseresponsecurvetosinglebrightlightpulsesinhumansubjects.TheJournalofPhysiology,2003.549(3):p.945‐952.

184. Minors, D., et al., Ahumanphase‐response curve to light. Neuroscience Letters, 1991.133(1):p.36‐40.

185. Koller,M., et al.,Differentpatternsoflightexposureinrelationtomelatoninandcortisolrhythmsandsleepofnightworkers.JournalofPinealResearch,1994.16(3):p.127‐135.

186. Arendt, J., et al., Efficacy ofmelatonin treatment in jet lag, shiftwork, and blindness.JournalofBiologicalRhythms,1997.12(6):p.604‐617.

187. Zipapel,N.,Circadianrhythmsleepdisorders.CNSDrugs,2001.15(4):p.311‐328.188. Commission Internationale de l’Eclairage (CIE), Ocular lighting effects on human

physiologyandbehaviour, in CIETechnicalSupport158, C.I.d. l'Eclairage, Editor. 2004,CIEVienna.

189. Cajochen,C.,Alertingeffectsoflight.SleepMedicineReviews,2007.11(6):p.453‐464.190. Rüger, M., et al., Time‐of‐day‐dependent effects of bright light exposure on human

psychophysiology: comparison of daytime and nighttime exposure. American Journal ofPhysiologyRegulatory,IntegrativeandComparativePhysiology,2006.290(5):p.R1413‐R1420.

191. Campbell,S.S.andD.Dawson,Enhancementofnighttimealertnessandperformancewithbrightambientlight.Physiology&Behavior,1990.48(2):p.317‐320.

192. Badia, P., et al., Bright light effects on body temperature, alertness, EEG and behavior.Physiology&Behavior,1991.50(3):p.583‐588.

193. Cajochen, C., et al., Dose‐response relationship for light intensity and ocular andelectroencephalographic correlates of human alertness. Behavioural Brain Research,2000.115(1):p.75‐83.

194. Daurat, A., et al., Bright light during nighttime: effects on the circadian regulation ofalertnessandperformance.BiologicalSignalandReceptors,2000.9(6):p.309‐318.

195. Rautkylä, E., M. Puolakka, and L. Halonen,Alertingeffectsofdaytime lightexposure ‐aproposed link between light exposure and brain mechanisms. Lighting Research andTechnology,2012.44(2):p.238‐252.

196. Phipps‐Nelson, J., et al., Daytime exposure to bright light, as compared to dim Light,decreasessleepinessandimprovespsychomotorvigilanceperformance.Sleep,2003.26(6):p.695‐700.

197. Vandewalle, G., et al., Daytime light exposure dynamically enhances brain responses.CurrentBiology,2006.16(16):p.1616‐1621.

198. Perrin, F., et al., Nonvisual responses to light exposure in the human brain during thecircadiannight.CurrentBiology,2004.14(20):p.1842‐1846.

199. Haack, M. and J. Mullington, Sustained sleep restriction reducesemotionalandphysicalwell‐being.Pain,2005.119(1):p.56‐64.

200. Kupermann,M.,etal.,Sleepproblemsandtheircorrelatesinaworkingpopulation.1995.201. Doi,Y.,M.Minowa,andT.Tango, ImpactandcorrelatesofpoorsleepqualityinJapanese

white‐collaremployees.Sleep,2003.26(4):p.467‐471.202. Viola, A.U., et al., Blue‐enriched white light in the workplace improves self‐reported

alertness,performanceandsleepquality.Scandinavian Journalof Work,Environment&Health,2008.34(4):p.297‐306.

203. Lu, B.S. and P. Zee, Circadian rhythm sleepdisorder. CHEST Journal, 2006. 130(6): p.1915‐1923.

204. Koch, S., et al., Effectiveness of sleepmanagement strategies for residents ofaged carefacilities: findingsofa systematic review. Journal of Clinical Nursing, 2006. 15(10): p.1267‐1275.

205. Cajochen, C., et al., Evening exposure toa light‐emittingdiodes (LED)‐backlit computerscreen affects circadian physiology and cognitive performance. Journal of AppliedPhysiology,2011.110(5):p.1432‐1438.

206. Gaina, A., et al., Epidemiologicalaspectsof self‐reported sleeponset latency in Japanesejuniorhighschoolchildren.JournalofSleepResearch,2006.15(3):p.266‐275.

References

144

207. Golder, S.A. and M.W. Macy, Diurnal and seasonal mood vary with work, sleep, anddaylengthacrossdiversecultures.Science,2011.333(6051):p.1878‐1881.

208. Boivin, D.B., et al., Complex interaction of the sleep‐wake cycle and circadian phasemodulatesmoodinhealthysubjects.ArchivesofGeneralPsychiatry,1997.54(2):p.145‐152.

209. Bauer, M.S., et al., Mood and behavioral effects of four‐week light treatment inwinterdepressivesandcontrols.JournalofPsychiatricResearch,1994.28(2):p.135‐145.

210. Cunningham, M.R., Weather, mood and helping behavior: Quasi experiments with thesunshinesamaritan.JournalofPersonalityandSocialPsychology,1979.37(11):p.1947‐1956.

211. Daurat, A., et al., Bright light affects alertness and performane rhythms during a 24‐hconstantroutine.Physiology&Behavior,1993.53(5):p.929‐936.

212. Partonen, T. and J. Lönnqvist, Bright light improves vitality and alleviates distress inhealthypeople.JournalofAffectiveDisorders,2000.57(1):p.55‐61.

213. Avery,D.H., et al.,Brightlighttherapyofsubsyndromalseasonalaffectivedisorderintheworkplace:morningvs.afternoonexposure.ActaPsychiatricaScandinavica,2001.103(4):p.267‐274.

214. McColl, S.L. and J.A. Veitch, Full‐spectrum fluorescent lighting:areviewof itseffectsonphysiologyandhealth.PsychologicalMedicine,2001.31(6):p.949‐964.

215. Boyce, P.R., et al., Lighting the graveyard‐shift: the influence of a daylight simulatingskylightonthetaskperformanceandmoodofnight‐shiftworkers. LightingResearch andTechnology,1997.29(3):p.105‐134.

216. Leppämäki, S., et al., Timed bright‐light exposure and complaints related to shiftworkamongwomen.ScandinavianJournalofWork,Environment&Health,2003.29(1):p.22‐26.

217. Iwata, N., S. Ichii, and K. Egashira,Effectsofbrightartificial lightonsubjectivemoodofshiftworknurses.IndustrialHealth,1997.35(1):p.41‐47.

218. Eastman, C.I., L. Liu, andL.F. Fogg,Circadianrhythmadaptationtosimulatednightshiftwork:effectofnocturnalbright‐lightduration.Sleep,1995.18(6):p.399‐407.

219. Aries, M.B.C., Human lighting demands, healthy lighting in an office environment, inDepartment of the Built Environment. 2005, University of Technology Eindhoven:Eindhoven,TheNetherlands.

220. Fostervold,K.I. and J.Nersveen,Proportionsofdirectandindirectindoorlighting—Theeffectonhealth,well‐beingandcognitiveperformanceofofficeworkers.LightingResearchandTechnology,2008.40(3):p.175‐200.

221. Mills, P.R., S.C. Tomkins, and L.J.M. Schlangen, The effect of high correlated colourtemperature office lighting on employee wellbeing and work performance. Journal ofCircadianRhythms,2007.5(2):p.2‐10.

222. Wilhelm, B., et al., Increased illuminanceat theworkplace:Does ithaveadvantages fordaytimeshifts?LightingResearchandTechnology,2011.43(2):p.185‐199.

223. Rosen, L.N., et al., Prevalenceofseasonalaffectivedisorderat four latitudes. PsychiatryResearch,1989.31:p.131‐144.

224. Terman, M. and J. Terman, Circadian rhythm phase advance with dawn simulationtreatmentforwinterdepression.Journalofbiologyrhythms,2010.25(4):p.298‐301.

225. Roenneberg, T., A. Wirz‐Justice, and M. Merrow, Life between clocks: daily temporalpatternsofhumanchronotypes.JournalofBiologicalRhythms,2003.18(1):p.80‐90.

226. Horne, J.A. and O. Ostberg, A self‐assessment questionnaire to determinemorningness‐eveningnessinhumancircadianrhythms. International Journal of Chronobiology, 1976.4(2):p.97‐110.

227. Roenneberg,T.,etal.,Amarkerfortheendofadolescence.CurrentBiology,2004.14(24):p.R1038‐R1039.

228. Herljevic, M., et al., Light‐induced melatonin suppression: age‐related reduction inresponsetoshortwavelengthlight.ExperimentalGerontology,2005.40(3):p.237‐242.

References

145

229. Boyce, P. and D.J. Kennaway, Effects of light on melatonin production. Biologicalpsychiatry,1987.22(4):p.473‐478.

230. Nathan, P.J., et al., The effect of gender on themelatonin suppression by light: A doseresponse relationship. Journal of Neural Transmission/General Section JNT, 2000.107(3):p.271‐279.

231. Wright,K.P.,etal.,Caffeineandlighteffectsonnighttimemelatoninandtemperaturelevelsinsleep‐deprivedhumans.BrainResearch,2000.747(1):p.78‐84.

232. Monteleone, P., et al., Doesbright light suppressnocturnalmelatonin secretionmore inwomenthanmen?JournalofNeuralTransmission/GeneralSectionJNT,1995.102(1):p.75‐80.

233. Brainard, G.C., et al., Dose‐response relationship between light irradiance and thesuppressionofplasmamelatonin inhumanvolunteers. Brain Research 1988.454(1): p.212‐218.

234. Vandewalle,G.,P.Maquet,andD.J.Dijk,Lightasamodulatorofcognitivebrainfunction.TrendsinCognitiveSciences,2009.13(10):p.429‐438.

235. Smolders,K.C.H.J.andY.A.W.DeKort,Ahigherilluminanceinducesalertnessevenduringofficehours:Findingsonsubjectivemeasures,taskperformanceandheartratemeasures.Physiology&Behavior,2012.107(1):p.7‐16.

236. Kaida, K., et al., Indoor exposure to natural bright light prevents afternoon sleepiness.Sleep,2006.29(4):p.462‐469.

237. Brainard, G.C. and J.P. Hanifin,Photons,Clocks,andConsciousness. Journal of BiologicalRhythms,2005.20(4):p.314‐325.

238. Gall,D.andK.Bieske.Definitionandmeasurementofcircadianradiometricquantities. inCIE Symposium 2004 on Light and Health: Non‐Visual Effects. 2004. Vienna, Austria:Commissioninternationaledel’éclairage.

239. Münch, M., et al., Wavelength‐dependent effects of evening light exposure on sleeparchitecture and sleep EEG power density in men. American Journal of PhysiologyRegulatory,IntegrativeandComparativePhysiology,2006.290(5):p.421‐428.

240. Lockley, S.W., Circadian Rhythms: Influence of Light in Humans. Encyclopedia ofNeuroscience2009:p.971‐988.

241. Küller, R. and L. Wetterberg, Melatonin, cortisol, EEG, ECG and subjective comfort inhealthy humans: impact of two fluorescent lamp types at two light intensities. LightingResearchandTechnology,1993.25(2):p.71‐81.

242. Chellappa, S.L., et al., Non‐visual effects of light onmelatonin, alertness and cognitiveperformance:Canblue‐enrichedlightkeepusalert?PLoSOne,2011.6(1):p.e16429.

243. Münch,M.andV.Bromundt,Lightandchronobiology:implicationsforhealthanddisease.DialoguesinClinicalNeuroscience,2012.14(4):p.448‐453.

244. Benedetti, F., et al., Morning sunlight reduces length of hospitalization in bipolardepression.JournalofAffectiveDisorders,2001.62(3):p.221‐223.

245. Pechacek, C.S., M. Andersen, and S.W. Lockley, Preliminary Method for ProspectiveAnalysis of the Circadian Efficacy of (Day)Light with Applications to HealthcareArchitecture. LEUKOS ‐ the Journal of the Illuminating Engineering Society of NorthAmerica,2008.5(1):p.1‐26.

246. Wirz‐Justice, A. and C. Fournier, Light, Health and Wellbeing: Implications fromchronobiologyforarchitecturaldesign.WorldHealthDesign,2010.8:p.44‐49.

247. Webb,A.R.,Considerationsforlightinginthebuiltenvironment:non‐visualeffectsoflight.EnergyandBuildings,2006.38(7):p.721‐727.

248. Laurentin,C.,V.Berrutto,andM.Fontoynont,Effectofthermalconditionsandlightsourcetypeonvisualcomfortappraisal.LightingResearchandTechnology,2000.32(4):p.223‐233.

249. Wienold, J.,Daylightglareinoffices, inFraunhoferInstituteforSolarEnergySystemsISE.2009,UniversitätKarlsruhe:Freibourg.p.132.

References

146

250. Linhart, F.,Energetic,visualandnon‐visualaspectsofofficelighting, inSolarEnergyandBuildingPhysicsLaboratory. 2010, Ecole Polytechnique Fédérale de Lausanne (EPFL):Lausanne.

251. Münch,M., et al.,Effectsofpriorlightexposureonearlyeveningperformance,subjectivesleepiness,andhormonalsecretion.BehavioralNeuroscience,2012.126(1):p.196‐203.

252. Altherr, R. and J.B. Gay, A low environmental impact anidolic façade. Building andEnvironment,2002.37(12):p.1409‐1419.

253. Wewers, M.E. and N.K. Lowe, A critical review of visual analogue scales in themeasurementofclinicalphenomena.ResearchinNursing&Health,1990.13(4):p.227‐236.

254. Eklund,N.H.andP.R.Boyce,Thedevelopmentofareliable,validandsimpleofficelightingsurvey.JournaloftheIlluminatingEngineeringSocietyofNorthAmerica,1996.25(2):p.25‐40.

255. Curran‐Everett,D.,Multiplecomparisons:philosophiesandillustrations.AmericanJournalofPhysiologyRegulatory,IntegrativeandComparativePhysiology,2000.279(1):p.R1‐R8.

256. Markus, T.A. The significance of sunshine and view for office workers. in Sunlight inBuilding,BoewcentrumInternational.1967.Rotterdam,TheNetherlands

257. Wells, B.W.P., Subjectiveresponsestothelightinginstallationinamodernofficebuildingandtheirdesignimplications.BuildingScience,1965.1(1):p.57‐68.

258. Osterhaus, W.K.E. Discomfortglare fromdaylight in computeroffices:howmuchdowereallyknow?inProceedingsofLUXEUROPA2001.2001.Reykjavik,Iceland.

259. Goulet,G.,etal.,Dailylightexposureinmorning‐typeandevening‐typeindividuals.JournalofBiologicalRhythms,2007.22(2):p.151‐158.

260. Hébert, M., M. Dumont, and J. Paquet, Seasonal and diurnal patterns of humanilluminationundernaturalconditions. Chronobiology International, 1998.15(1): p. 59‐70.

261. Roenneberg, T., S. Daan, and M. Merrow, Theartofentrainment. Journal of BiologicalRhythms,2003.18(3):p.183‐194.

262. Vollmer, C., U. Michel, and C. Randler, Outdoor lightatnight (LAN) is correlatedwitheveningnessinadolescents.ChronobiologyInternational,2012.29(4):p.502‐508.

263. Borisuit, A., et al. Light preference and mood in extreme chronotypes in response todifferentoffice lightingconditions:preliminaryresults. inProceedingsExperiencingLight2012.2012.Eindhoven,theNetherlands.

264. Borisuit,A.,etal.Alteredsubjectivealertnessandvisualcomfortinextremechronotypesinresponsetodifferentofficelightingconditions:preliminaryresultsin SLTBR2012AnnualMeeting.2012.Geneva,Switzerland.

265. Maierova, L., et al. The impactofdifferent lighting conditionson extreme chronotypes:Effectsonphysicalcomfortandalertness. inProceedingsExperiencingLight2012. 2012.Eindhoven,theNetherlands.

266. Maierova,L.,etal.Indoorlightingconditionsandinter‐individuallightpreferences:effectson subjectivealertness,physicalwellbeingandelectricalenergy consumption. in CLIMA.2013.Prague,CzechRepublic.

267. AssociationSuissedel'Eclairage(SLG/ASE),Eclairageintérieurparlalumièredujour,inRecommandations de l'Association Suisse de l'Eclairage (SLG). 1989, SchweizerischeNormen‐Vereinigung:Zürich.

268. Meteosuisse.WebsiteofMeteosuisse.http://www.meteosuisse.ch 2012;Availablefrom:http://www.meteosuisse.ch.

269. Hunt, D.G.R., The use of artificial lighting in relation to daylight levels and occupancy.BuildingandEnvironment,1979.14(1):p.21‐33.

270. Love,J.A.,Manualswitchingpatternsinprivateoffices.LightingResearchandTechnology,1998.30(1):p.45‐50.

271. Reinhart, C.F. and K. Voss, Monitoringmanual control of electric lighting and blinds.LightingResearchandTechnology,2003.35(3):p.243‐260.

References

147

272. Moore, T., D.J. Carter, and A.I. Slater, Userattitudes toward occupant controlledofficelighting.LightingResearchandTechnology,2002.34(3):p.207‐219.

273. Commission Internationalede l’Eclairage (CIE),DiscomfortglareintheinteriorworkingenvironmentinCIE055‐1983.1983,Commissioninternationaldel'éclairage:Vienna.

274. Einhorn, H.D.,Anewmethod forassessmentofdiscomfortglare. Lighting Research andTechnology,1969.1(4):p.235‐247.

275. Boyce, P.R. and J.W. Beckstead. Influencesondiscomfortglareratings. inProceedingsofthe1stInternationalSymposiumonGlare.1991.Orlando,Florida,USA.

276. CIE,DiscomfortGlareintheInteriorLighting.,C.I.d.L'Eclairage,Editor.1995:Vienna.277. Scartezzini, J.‐L.andG.Courret,Anidolicdaylightingsystems.SolarEnergy,2002.73(2):

p.123‐135.278. Courret,G.,Systèmesanidoliquesd'éclairagenaturel,inSolarEnergyandBuildingPhysics

Laboratory.1999,SwissFederalInstituteofTechnologyinLausanne:Lausanne.279. Waters, C.E., R.G. Mistrick, and C.A. Bernecker, Discomfort glare from sources of

nonuniformluminance. Journal of the IlluminatingEngineering Society, 1995.24(2): p.73‐85.

148

AppendixA

AppendixA

149

I. QuestionnairesforExperimentI.

a) Visualcomfortvariables

Subject Code :Liper_XX_XX XX.XX.20XX

Test Number:1

Echelle de confort visuel Voici quelques questions concernant l’environnement lumineux de cette salle. Veuillez marquer votre consentement avec chaque déclaration sur la ligne correspondante. (1) J’aime la lumière dans cette salle.

OUI NON

(2) Globalement, l’éclairage de cette salle est agréable.

OUI NON (3) Cette salle me semble trop lumineuse.

OUI NON (4) Cette salle me semble trop sombre.

OUI NON (5) Il n’y a pas assez de lumière pour travailler / lire correctement.

OUI NON (6) Il y a trop de lumière pour travailler / lire correctement.

OUI NON (7) Comment ressentez-vous l’éblouissement dans cette salle.

Juste Juste Juste Juste Perceptible acceptable inconfortable intolérable

Imperceptible Perceptible Acceptable Inconfortable Intolérable

AppendixA

150

b) Non‐ImageForming(NIF)functions

Subject Code : Liper_XX_XX XX.XX.20XX

Test Number: 1

Indiquez sur l’échelle suivante avec un trait, comment vous vous sentez en ce moment.

Extrêmement Extrêmement relaxé tendu Physiquement Physiquement

à l’aise pas du tout à l’aise Extrêmement Extrêmement éveillé fatigué Rassasié Affamé De mauvaise De très humeur bonne humeur

Extrêmement ___________________________________________________ Extrêmement chaud froid

Remarques :

AppendixA

151

II. QuestionnairesforExperimentII.

a) Visualcomfortvariables

Subject Code : Chroli_xx_xx XX.XX.20XX

Test Number:1

Echelle de confort visuel (part 1) Voici quelques questions concernant l’environnement lumineux de cette salle. Veuillez marquer votre consentement avec chaque déclaration sur la ligne correspondante. (1) J’aime la lumière dans cette salle.

OUI NON (2) Globalement, l’éclairage de cette salle est confortable.

OUI NON (3) Il n’y a pas assez de lumière pour travailler / lire correctement.

OUI NON (4) Il y a trop de lumière pour travailler / lire correctement.

OUI NON (5) La lumière est mal distribuée dans cette salle.

OUI NON (6) Comment ressentez-vous l’éblouissement dans cette salle.

Juste Juste Juste Juste Perceptible acceptable inconfortable intolérable

Imperceptible Perceptible Acceptable Inconfortable Intolérable

AppendixA

152

Subject Code : Chroli_XX_XX XX. XX. 20XX

Test Number:1

Echelle de confort visuel (part 2) Voici quelques questions concernant l’environnement lumineux de cette salle. Veuillez marquer votre consentement avec chaque déclaration sur la ligne correspondante. (7) La couleur de ma peau apparait peu naturelle sous cet éclairage.

OUI NON (8) La lumière dans cette salle est trop « chaude » pour un lieu de travail.

OUI NON (9) La lumière dans cette salle est trop « froide » pour un lieu de travail.

OUI NON (10) Si je compare la situation lumineuse de cette salle avec d’autres dans

lesquels j’ai travaillé auparavant, je dirais que la situation lumineuse ici est ...

meilleure

plutôt meilleure

plutôt pareille

plutôt pire

pire

(11) Dans le cadre d’une journée de travail, je pourrais bien m’imaginer travailler

dans cet environnement lumineux pendant …

moins de 2 heures

2 à 4 heures

4 à 6

plus que 6 heures

AppendixA

153

b) Non‐ImageForming(NIF)functions

CHROLI_XX__XX XX.XX.20XX

Test Number: 1

Indiquez sur l’échelle suivante avec un trait, comment vous vous sentez en ce moment.

Extrêmement Extrêmement relaxé tendu Physiquement Physiquement

à l’aise pas du tout à l’aise Extrêmement Extrêmement éveillé fatigué Rassasié Affamé De mauvaise De très humeur bonne humeur

Extrêmement ___________________________________________________ Extrêmement chaud froid Remarques :

154

CurriculumVitaeContactinformation:[email protected]

Education

2009‐Present PhD DoctoralPrograminEnvironment(EDEN)

EcolePolytechniqueFédéraledeLausanne(EPFL)

2003‐2003 GraduateSchool

MasterofAppliedArts

DESSArtAppliqués:Colourinenvironmentprojects(MentionBien)

UniversitédeToulouseIILeMirail

1994‐1999 University FactultyofArchitecture,ChulalongkornUniversity

Employmenthistory

2009‐Present Research&TeachingAssistant

SolarEnergyandBuildingPhysicsLaboratory


2009 VisitingScholar SolarEnergyResearchInstituteofSingapore

NationalUniversityofSingapore

2007–2008 ResearchAssistant SolarEnergyandBuildingPhysicsLaboratory


2003‐2007 ProjectLightingDesigner VisionDesignStudioinBangkok,Thailand

2003 InternLightingDesigner LightCibles&LouisClairinParis,France

1999–2001 Architect TerraArchitectinBangkok,Thailand

CurriculumVitae

155

Professionalqualifications

Certification Diplômed'étudesenlanguefrançaise,1erdegré

UniversitéStendhal–GrenobleIII,France

Accreditation Licenseno.Arch.5527(Thailand)

ComputerSkills General AutoCAD,AdobePhotoshop/Illustrator,MSOffice

Specific Evalglare,Radiance,Dialux,Relux,Photosphere,LESOSAI

Languages Thai MotherTongue

English Fluent(ModuleC1)

French Fluent(ModuleB2)

Awards

2012 PhDMobilityAwards DoctoralPrograminCivilandEnvironmentalEngineering


2013 WinnerofScienceSlam ThaiStudentAcademicConference,Göttingen,Germany

Publications

Articlesinpeerreviewedjournals

Münch M, Linhart F, Borisuit A, Jaeggi SM, Scartezzini JL. Effects of prior light exposure oncognitiveperformance,subjectivesleepiness,andhormonalsecretionintheevening.BehavioralNeuroscience,2012,126(1):196‐203.Epub2011Dec26.

BorisuitA,ScartezziniJ‐L,ThanachareonkitA.VisualDiscomfortandGlareRatingAssessmentofIntegrated Daylighting and Electric Lighting Systems using HDR Imaging Technique.ArchitecturalScienceReview,53(4),359‐373

Peerreviewedconferenceproceedings

BorisuitA,DeschampsL,Kämpf J,Scartezzini J‐L,MünchM.Assessmentofcircadianweightedradiancedistributionusingacamera‐likelightsensor.ProceedingsofCISBAT2013InternationalScientificConference;Lausanne,Switzerland,September4‐6,2013.

CurriculumVitae

156

Borisuit A, Münch M, Deschamps, L, Kämpf J, Scartezzini J‐L. A new device for dynamicluminance mapping and glare risk assessment in buildings. Proceedings of SPIE2012InternationalSymposiumonOpticalEngineering;SanDiego,USA,August13‐16,2012.

BorisuitA,LinhartF.,KämpfJ,ScartezziniJ‐L,MünchM.Comparisonofobjectiveandsubjectivevisual comfort and associations with non‐visual functions in young subjects. Proceedings ofCISBAT2011InternationalScientificConference;2011Lausanne, Switzerland, September14‐16,2013.

BasurtoC,BorisuitA.,KämpfJ,MünchM.,ScartezziniJ‐L.DaylightOptimizationofbuildingsandapplication of advanced daylighting systems in central Mexico. Proceedings of CISBAT2011InternationalScientificConference;Lausanne,Switzerland,September14‐16,2013.

Professionalmembership

MemberofAssociationofSiameseArchitectsundertheRoyalPatronage(ASA),Thailand

Others

2011‐2012 PresidentoftheAssociationofThaiStudentsinSwitzerland(ATSS)

the impact of light including non-image forming effects on ... · prof. j.-l. scartezzini, dr m....

Documents