The Camera Verses the Human Eye

7/26/2019 The Camera Verses the Human Eye

1/10

The

Camera

Versus

the

Human

Eye

Nov17,2012 RogerCicala

Thisarticlewasoriginallypublishedasablog.PermissionwasgrantedbyRogerCicalatore

publishthearticleontheCTIwebsite.Itisanexcellentarticleforthosepolicedepartments

consideringtheuseofcameras.

ThisarticlestartedafterIfollowedanonlinediscussionaboutwhethera35mmora50mmlens

onafullframecameragivestheequivalentfieldofviewtonormalhumanvision.Thisparticular

discussionimmediatelydelvedintotheopticalphysicsoftheeyeasacameraandlensan

understandablecomparisonsincetheeyeconsistsofafrontelement(thecornea),anaperture

ring(theirisandpupil),alens,andasensor(theretina).

Despite

all

the

impressive

mathematics

thrown

back

and

forth

regarding

the

optical

physics

of

theeyeball,thediscussiondidntquiteseemtomakesenselogically,soIdidalotofreadingof

myownonthetopic.

Therewontbeanydirectbenefitfromthisarticlethatwillletyourunoutandtakebetter

photographs,butyoumightfinditinteresting.Youmayalsofinditincrediblyboring,soIllgive

youmyconclusionfirst,intheformoftwoquotesfromGarryWinogrand:


2/10

Aphotographistheillusionofaliteraldescriptionofhowthecamerasawapieceoftimeand

space.

Photographyisnotaboutthethingphotographed.Itisabouthowthatthinglooks

photographed.

Basicallyindoingallthisresearchabouthowthehumaneyeislikeacamera,whatIreally

learnedishowhumanvisionisnotlikeaphotograph.Inaway,itexplainedtomewhyIsooften

findaphotographmuchmorebeautifulandinterestingthanIfoundtheactualsceneitself.

TheEyeasaCameraSystem

Superficially,itsprettylogicaltocomparetheeyetoacamera.Wecanmeasurethefrontto

backlengthoftheeye(about25mmfromthecorneatotheretina),andthediameterofthe

pupil(2mmcontracted,7to8mmdilated)andcalculatelenslikenumbersfromthose

measurements.

Youllfindsomedifferentnumbersquotedforthefocallengthoftheeye,though.Someare

fromphysicalmeasurementsoftheanatomicstructuresoftheeye,othersfromoptometric

calculations,sometakeintoaccountthatthelensoftheeyeandeyesizeitselfchangewiththe

contractionsofvariousmuscles.

Tosummarize,though,onecommonlyquotedfocallengthoftheeyeis17mm(thisis

calculatedfromtheOptometricdioptervalue).Themorecommonlyacceptedvalue,however,


3/10

is22mmto24mm(calculatedfromphysicalrefractionintheeye).Incertainsituations,the

focallengthmayactuallybelonger.

Sinceweknowtheapproximatefocallengthandthediameterofthepupil,itsrelativelyeasyto

calculatetheaperture(fstop)oftheeye.Givena17mmfocallengthandan8mmpupilthe

eyeballshouldfunctionasanf/2.1lens.Ifweusethe24mmfocallengthand8mmpupil,it

shouldbef/3.5.Therehaveactuallybeenanumberofstudiesdoneinastronomytoactually

measurethefstopofthehumaneye,andthemeasurednumbercomesouttobef/3.2tof/3.5

(Middleton,1958).

Atthispoint,bothofyouwhoreadthisfarprobablyhavewonderedIfthefocallengthofthe

eyeis17or24mm,whyiseveryonearguingaboutwhether35mmor50mmlensesarethe

samefieldofviewasthehumaneye?

Thereasonisthatthemeasuredfocallengthoftheeyeisntwhatdeterminestheangleofview

ofhumanvision.Illgetintothisinmoredetailbelow,butthemainpointisthatonlypartofthe

retinaprocessesthemainimagewesee.(Theareaofmainvisioniscalledtheconeofvisual

attention,therestofwhatweseeisperipheralvision).

Studieshavemeasuredtheconeofvisualattentionandfoundittobeabout55degreeswide.

Ona35mmfullframecamera,a43mmlensprovidesanangleofviewof55degrees,sothat

focallengthprovidesexactlythesameangleofviewthatwehumanshave.Damnifthatisnt

halfwaybetween35mmand50mm.Sotheoriginalargumentisended,theactualnormallens

ona35mmSLRisneither35mmnor50mm,itshalfwayinbetween.


4/10

TheEyeisNotaCameraSystem

Havinggottentheanswertotheoriginaldiscussion,Icouldhaveleftthingsaloneandwalked

awaywithyetanotherbitoffairlyuselesstriviafiledawaytoamazemyonlinefriendswith.But

NOOoooo.WhenIhaveabunchofworkthatneedsdoing,IfindIllalmostalwayschooseto

spendanothercoupleofhoursreadingmorearticlesabouthumanvision.

Youmayhavenoticedtheabovesectionleftoutsomeoftheeyetocameraanalogies,because

onceyougetpastthesimplemeasurementsofapertureandlens,therestofthecomparisons

dontfitsowell.

Considertheeyessensor,theretina.Theretinaisalmostthesamesize(32mmindiameter)as

thesensoronafullframecamera(35mmindiameter).Afterthat,though,almosteverythingis

different.

The

retina

of

a

human

eye

Thefirstdifferencebetweentheretinaandyourcamerassensorisratherobvious:theretinais

curvedalongthebacksurfaceoftheeyeball,notflatlikethesiliconsensorinthecamera.The

curvaturehasanobviousadvantage:theedgesoftheretinaareaboutthesamedistancefrom

thelensasthecenter.Onaflatsensortheedgesarefurtherawayfromthelens,andthecenter

closer.Advantageretinaitshouldhavebettercornersharpness.


5/10

Thehumaneyealsohasalotmorepixelsthanyourcamera,about130millionpixels(you24

megapixelcameraownersfeelinghumblenow?).However,onlyabout6millionoftheeyes

pixelsarecones(whichseecolor),theremaining124millionjustseeblackandwhite.But

advantageretinaagain.Bigtime.

Butifwelookfurtherthedifferencesbecomeevenmorepronounced

Onacamerasensoreachpixelissetoutinaregulargridpattern.Everysquaremillimeterofthe

sensorhasexactlythesamenumberandpatternofpixels.Ontheretinatheresasmallcentral

area,about6mmacross(themacula)thatcontainsthedensestconcentrationofphoto

receptorsintheeye.Thecentralportionofthemacula(thefovea)isdenselypackedwithonly

cone(colorsensing)cells.Therestofthemaculaaroundthiscentralcoloronlyareacontains

bothrodsandcones.

Themaculacontainsabout150,000pixelsineach1mmsquare(comparethatto24,000,000

pixelsspreadovera35mmx24mmsensorina5DMkIIorD3x)andprovidesourcentralvision

(the55degreeconeofvisualattentionmentionedabove).Anyway,thecentralpartofour

visualfieldhasfarmoreresolvingabilitythaneventhebestcamera.

Therestoftheretinahasfarfewerpixels,mostofwhichareblackandwhitesensingonly.It

provideswhatweusuallyconsiderperipheralvision,thethingsweseeinthecornerofour


6/10

eye.Thispartsensesmovingobjectsverywell,butdoesntprovideenoughresolutiontoread

abook,forexample.

Thetotalfieldofview(theareainwhichwecanseemovement)ofthehumaneyeis160

degrees,butoutsideoftheconeofvisualattentionwecantreallyrecognizedetail,onlybroad

shapesandmovement.

Theadvantagesofthehumaneyecomparedtothecameragetreducedabitasweleavethe

retinaandtravelbacktowardthebrain.Thecamerasendseverypixelsdatafromthesensorto

acomputerchipforprocessingintoanimage.Theeyehas130millionsensorsintheretina,but

theopticnervethatcarriesthosesensorssignalstothebrainhasonly1.2millionfibers,soless

than10%oftheretinasdataispassedontothebrainatanygiveninstant.(Partlythisis

becausethechemicallightsensorsintheretinatakeawhiletorechargeafterbeing

stimulated.Partlybecausethebraincouldntprocessthatmuchinformationanyway.)

Andofcoursethebrainprocessesthesignalsalotdifferentlythanaphotographycamera.

Unliketheintermittentshutterclicksofacamera,theeyeissendingthebrainaconstantfeed

videowhichisbeingprocessedintowhatwesee.Asubconsciouspartofthebrain(thelateral

geniculatenucleusifyoumustknow)comparesthesignalsfrombotheyes,assemblesthemost

importantpartsinto3Dimages,andsendsthemontotheconsciouspartofthebrainfor

imagerecognitionandfurtherprocessing.


7/10

Thesubconsciousbrainalsosendssignalstotheeye,movingtheeyeballslightlyinascanning

patternsothatthesharpvisionofthemaculamovesacrossanobjectofinterest.Overafew

splitsecondstheeyeactuallysendsmultipleimages,andthebrainprocessesthemintoamore

completeanddetailedimage.

Thesubconsciousbrainalsorejectsalotoftheincomingbandwidth,sendingonlyasmall

fractionofitsdataontotheconsciousbrain.Youcancontrolthistosomeextent:forexample,

rightnowyourconsciousbrainistellingthelateralgeniculatenucleussendmeinformation

fromthecentralvisiononly,focusonthosetypedwordsinthecenterofthefieldofvision,

movefromlefttorightsoIcanreadthem.Stopreadingforasecondandwithoutmovingyour

eyestrytoseewhatsinyourperipheralfieldofview.Asecondagoyoudidntseethatobject

totherightorleftofthecomputermonitorbecausetheperipheralvisionwasntgettingpassed

ontotheconsciousbrain.

Ifyouconcentrate,evenwithoutmovingyoureyes,youcanatleasttelltheobjectisthere.If

youwanttoseeitclearly,though,youllhavetosendanotherbrainsignaltotheeye,shifting

theconeofvisualattentionovertothatobject.Noticealsothatyoucantbothreadthetext

andseetheperipheralobjectsthebraincantprocessthatmuchdata.

Thebrainisntdonewhentheimagehasreachedtheconsciouspart(calledthevisualcortex).

Thisareaconnectsstronglywiththememoryportionsofthebrain,allowingyoutorecognize

objectsintheimage.Weveallexperiencedthatmomentwhenweseesomething,butdont

recognizewhatitisforasecondortwo.Afterweverecognizedit,wewonderwhyintheworld

itwasntobviousimmediately.Itsbecauseittookthebrainasplitsecondtoaccessthe

memoryfilesforimagerecognition.(Ifyouhaventexperiencedthisyet,justwaitafewyears.

Youwill.)

Inreality(andthisisveryobvious)humanvisionisvideo,notphotography.Evenwhenstaring

ataphotograph,thebrainistakingmultiplesnapshotsasitmovesthecenteroffocusoverthe


8/10

picture,stackingandassemblingthemintothefinalimageweperceive.Lookataphotograph

forafewminutesandyoullrealizethatsubconsciouslyyoureyehasdriftedoverthepicture,

gettinganoverviewoftheimage,focusinginondetailshereandthereand,afterafew

seconds,realizingsomethingsaboutitthatwerentobviousatfirstglance.

So

whats

the

Point?

Well,Ihavesomeobservations,althoughtheyrefarawayfromwhichlenshasthefieldof

viewmostsimilartohumanvision?.Thisinformationgotmethinkingaboutwhatmakesmeso

fascinatedbysomephotographs,andnotsomuchbyothers.Idontknowthatanyofthese

observationsaretrue,buttheyreinterestingthoughts(tomeatleast).Allofthemarebased

ononefact:whenIreallylikeaphotograph,Ispendaminuteortwolookingatit,lettingmy

humanvisionscanit,grabbingthedetailfromitorperhapswonderingaboutthedetailthats

notvisible.

Photographstakenatanormalangleofview(35mmto50mm)seemtoretaintheirappeal

whatevertheirsize.Evenwebsizedimagesshotatthisfocallengthkeeptheessenceofthe

shot.Theshotbelow(takenat35mm)hasalotmoredetailwhenseeninalargeimage,butthe

essenceisobviousevenwhensmall.Perhapsthebrainsprocessingismorecomfortable

recognizinganimageitseesatitsnormalfieldofview.Perhapsitsbecausewephotographers

tendtosubconsciouslyemphasizecompositionandsubjectsinanormalangleofview

photograph.

ThephotoabovedemonstratessomethingelseIvealwayswonderedabout:doesour

fascinationandloveforblackandwhitephotographyoccurbecauseitsoneofthefewways

thedensecone(coloronly)receptorsinourmaculaareforcedtosendagrayscaleimagetoour

brain?

Perhapsourbrainlikeslookingatjusttoneandtexture,withoutcolordatacloggingupthat

narrowbandwidthbetweeneyeballandbrain.


9/10

Likenormalangleshots,telephotoandmacroshotsoftenlookgreatinsmallprintsorweb

sizedJPGs.Ihavean810ofanelephantseyeandasimilarsizedmacroprintofaspideron

myofficewallthatevenfromacrosstheroomlookgreat.(Atleasttheylookgreattome,but

youllnoticethattheyrehanginginmyoffice.Ivehungtheminacoupleofotherplacesinthe

houseandhavebeentactfullytoldthattheyreallydontgowiththelivingroomfurniture,so

maybetheydontlooksogreattoeveryone.)

Theresnogreatcompositionorotherfactorstomakethosephotosattractivetome,butIfind

themfascinatinganyway.Perhapsbecauseevenatasmallsize,myhumanvisioncansee

detailsinthephotographthatInevercouldseelookingatanelephantorspiderwiththenaked

eye.


10/10

Ontheotherhand,whenIgetagoodwideangleorscenicshotIhardlyevenbothertoposta

websizedgraphicormakeasmallprint(andImnotgoingtostartforthisarticle).Iwantit

printedBIG.Ithinkperhapssothatmyhumanvisioncanscanthroughtheimagepickingout

thelittledetailsthatarecompletelylostwhenitsdownsized.AndeverytimeIdomakeabig

print,evenofasceneIvebeentoadozentimes,InoticethingsinthephotographIvenever

seenwhenIwasthereinperson.

Perhapsthevideomybrainismakingwhilescanningtheprintprovidesmuchmoredetailand

Ifinditmorepleasingthanthecompositionofthephotowouldgivewhenitsprintedsmall(or

whichIsawwhenIwasactuallyatthescene).

Andperhapsthesubconsciousscanningthatmyvisionmakesacrossaphotographaccounts

forwhythingsliketheruleofthirdsandselectivefocuspullsmyeyetocertainpartsofthe

photograph.Maybewephotographerssimplyfiguredouthowthebrainprocessesimagesand

tookadvantageofitthroughpracticalexperience,withoutknowingallthescienceinvolved.

ButIguessmyonlyrealconclusionisthis:aphotographisNOTexactlywhatmyeyeandbrain

sawatthescene.WhenIgetagoodshot,itssomethingdifferentandsomethingbetter,like

whatWinograndsaidinthetwoquotesabove,andinthisquotetoo:

Youseesomethinghappeningandyoubangawayatit.Eitheryougetwhatyousaworyouget

somethingelseandwhicheverisbetteryouprint.

About

the

author:RogerCicalaisthefounderofLensRentals.Thisarticlewasoriginally

publishedhere.

Image

credits:myeyeupclosebymachinecodeblue,Nikhseyethroughcameraseyefrommy

eyesforyoureyes:)byslalit,SchematicoftheHumanEyebyentirelysubjective,Mylefteye

retinabyRichardMasoner/Cyclelicious,Chromaticaberration(sortof)bymoppet65535

The Camera Verses the Human Eye

Documents