Field Guide to

Illumination

Angelo V. Arecchi
Tahar Messadi
R. John Koshel

SPIE Field Guides

Volume FG11

John E. Greivenkamp, Series Editor

Bellingham, Washington USA

 Introduction to the Series

Welcome to the SPIE Field Guides—a series of

publications written directly for the practicing engineer or
scientist. Many textbooks and professional reference
books cover optical principles and techniques in depth.
The aim of the SPIE Field Guides is to distill this
information, providing readers with a handy desk or
briefcase reference that provides basic, essential
information about optical principles, techniques, or
phenomena, including definitions and descriptions, key
equations, illustrations, application examples, design
considerations, and additional resources. A significant
effort will be made to provide a consistent notation and
style between volumes in the series.

Each SPIE Field Guide addresses a major field of optical

science and technology. The concept of these Field Guides
is a format-intensive presentation based on figures and
equations supplemented by concise explanations. In most
cases, this modular approach places a single topic on a
page, and provides full coverage of that topic on that page.
Highlights, insights, and rules of thumb are displayed in
sidebars to the main text. The appendices at the end of
each Field Guide provide additional information such as
related material outside the main scope of the volume,
key mathematical relationships, and alternative methods.
While complete in their coverage, the concise presentation
may not be appropriate for those new to the field.

The SPIE Field Guides are intended to be living

documents. The modular page-based presentation format
allows them to be easily updated and expanded. We are
interested in your suggestions for new Field Guide topics
as well as what material should be added to an individual
volume to make these Field Guides more useful to you.

John E. Greivenkamp, Series Editor

College of Optical Sciences
The University of Arizona
 The Field Guide Series

Field Guide to Geometrical Optics, John E. Greivenkamp

(FG01)

Field Guide to Atmospheric Optics, Larry C. Andrews

(FG02)

Field Guide to Adaptive Optics, Robert K. Tyson &

Benjamin W. Frazier (FG03)

Field Guide to Visual and Ophthalmic Optics, Jim

Schwiegerling (FG04)

Field Guide to Polarization, Edward Collett (FG05)

Field Guide to Optical Lithography, Chris A. Mack (FG06)

Field Guide to Optical Thin Films, Ronald R. Willey

(FG07)

Field Guide to Spectroscopy, David W. Ball (FG08)

Field Guide to Infrared Systems, Arnold Daniels (FG09)

Field Guide to Interferometric Optical Testing, Eric P.

Goodwin & James C. Wyant (FG10)
 Field Guide to Illumination

In writing this Field Guide to Illumination, the first task

was to decide what topics to include. Illumination tends to
mean different things to different people. Certainly any
subject matter under the purview of the CIE, Commission
Internationale de l’Eclairage (the International
Commission on Illumination) or the Illuminating
Engineering Society of North America (IESNA) must be
considered. Some particular areas pertaining to imaging
systems and nonimaging optics are potentially
overlooked. Thus, we chose to address a number of topics
that fall under the following three categories: imaging
system illumination, nonimaging optics for illumination,
and architectural illumination, which all call upon
principles of radiometry and photometry. Although this is
not a guide to radiometry, enough information on the
subject is included to make this manual a self-contained
document. Additionally, those optical properties of
materials that are pertinent to illumination, such as
surface color, scattering, and retroreflection are described.
The content in this Field Guide starts with traditional
illumination in imaging systems, followed by the recent
advances in computer-aided design of high efficiency
nonimaging illumination optics, along with the modern
source models that support these techniques. Sections on
the illumination of visual displays are included.
There was not enough room for a complete treatment of
architectural illumination, but some important topics are
included at the end of this Field Guide such as indoor and
outdoor architectural illumination.
The notation and terminology are consistent throughout
this Guide, but we do not lose sight of the fact that they
may not be consistent in the field. Examples of alternate
notation and terminology are presented.
Angelo Arecchi Tahar Messadi R. John Koshel
Hebron, NH Fayetteville, AR Tucson, AZ
 Table of Contents

Glossary ix

Basic Quantities in Illumination 1

Flux and Irradiance 1
Solid Angle 2
Intensity, Radiance, Projected Solid Angle 3
Solid Angle and Projected Solid Angle 4
Spectroradiometric and Radiometric Quantities 5
Photometric Quantities 6
Matrix of Basic Quantities 7
Photopic and Scotopic Vision 8
Luminous Efficacy 9
Typical Values of Illumination Quantities 10
Averaged LED Intensity 11

Color 12
Light Source Color 12
Chromaticity Diagram 13
Color Temperature and CCT 14
Dominant Wavelength and Purity 15
Surface Color 16
Color of Fluorescent Surfaces 17
Color Rendering and CRI 18
Calculating CRI and Problems with CRI 19

Sources for Illumination 20

Typical Source Parameters 20
Tungsten Lamps 21
Tungsten and Sunlight 22
Fluorescent Lamps 23
H.P. Sodium and Metal Halide 24
Xenon and White LED 25
Light Emitting Diodes (LEDs) 26

Illumination Properties of Materials 27

Transmittance, Reflectance, and Absorptance 27
Reflectance Factor and BRDF 28
Harvey / ABg Scatter Model 29
Directional Properties of Materials 30
vii
 Table of Contents (cont.)

Retroreflectors—Geometry 31
Retroreflectors—Radiometry 32

Illumination Transfer 33
Lambertian and Isotropic Models 33
Known Intensity 34
Known Flux and Known Radiance 35
Form Factor and Average Projected Solid Angle 36
Configuration Factor 37
Useful Configuration Factor 38
Useful Form Factor 39
Irradiance from a Uniform Lambertian Disk 40
Cosine Fourth and Increase Factor 41
Known Irradiance 42
ω, Ω, NA, and f/# for a Circular Cone 43
Invariance of Radiance 44

Illumination in Imaging Systems 45

Image Radiance 45
Limitations on Equivalent Radiance 46
Image Irradiance 47
f/#, Working f/#, T/#, NA, Ω 48
Flux and Étendue 49

Illumination in Nonimaging Systems 50

Generalized Étendue 50
Concentration 51
Skew Invariant 52

Fibers, Lightpipes, and Lightguides 53

Fibers—Basic Description 53
Numerical Aperture and Étendue 54
Fiber Bundles 55
Tapered Fibers and Bundles 56

Classical Illumination Designs 57

Spherical Reflector 57
Abbe Illumination 58
Köhler Illumination 59

viii
 Table of Contents (cont.)

Ellipsoidal and Paraboloidal Mirrors 60

Spectral Control and Heat Management 61
Illumination in Visual Afocal Systems 62

Uniform Illumination 63
Searchlight 63
Source at a Distance 64
Mixing Rod 65
Bent Lightpipes 66
Integrating Sphere 67
Lenslet Arrays 68
Small Reflectors, Lenslet Arrays, and Facets 69

Source Models 70
Source Modeling Overview 70
Source Modeling Methods 71
LED Modeling 72
Incandescent Lamp Modeling 73
Arc and Fluorescent Lamp Modeling 74

Nonimaging Compound Concentrators 75

Nonimaging Compound Concentrators 75
Concentrators as Luminaires 76
Compound Parabolic Concentrators 77
Compound Elliptical and Hyperbolic Concentrators 78
Tailored-Edge-Ray Design 79
Faceted Reflector Design 80
Advanced Nonimaging Optic Design 81

Displays 82
Displays—Overview 82
Backlit Display Components 83
Backlit Display: Source and Injector 84
Backlit Display: Lightguides, Features, Reflectors 85
Backlit Display: Polarizers, LC, and BEF 86
Projection Displays 87

Characterizing Illumination Systems 88

Mapping Flat-Fielding Sources 88

ix
 Table of Contents (cont.)

Goniophotometers 89
Types A, B, C Goniometer Coordinate Systems 90
“Snapshot” Goniophotometers 91

Software Modeling 92
Software Modeling Discussion 92

Architectural Illumination 93
Role of Light in Architecture 93

Light and Visual Performance 94

Eye Adaptation and Visual Fields 94
Apparent Brightness 95

Lighting Design 96
Lighting Design––Layering of Light 96

Luminaire for Open-Plan Office 97

Photometric Report and VCP 97
Spacing Criteria and Coefficient of Utilization 98

Daylight Compensation 99
Daylight Factor 99
Daylight Strategies 100

Exterior Lighting 101

Nighttime Visibility Criteria 101
Recommended Illuminance for Façades 102
Façade Floodlighting for Uniform Illumination 103
Illumination of Outdoor Areas 104
Special Considerations for Outdoor Fixtures 105

Parking 106
Outdoor Luminaire—Transverse Light
Distribution 106
Outdoor Luminaire—Lateral Light Distribution 107

Roadway Lighting 108

Criteria for Roadway Lighting 108

x
 Table of Contents (cont.)

Small Target Visibility 109

Recommended Roadway Luminaires 110
Recommended Lamps for Roadway Luminaires 111

Appendix 112
Equation Summary 112
CIE Illuminants A and D65 119
x , y , z , V(λ), and V′(λ) 122
Archaic and Arcane Units of Illumination 125

Bibliography 127

Index 133

xi
 Glossary

α Absorptance
α Observation angle (in retroreflection)
ai Input area to a compound concentrator
ao Output area of a compound concentrator
A Absorbance
Ai Illuminated area
Ar Radiating area
ax Area of plane x
aΩ Throughput, étendue
C Concentration ratio
CCT Correlated color temperature
CIE International Commission on Illumination
CRI Color rendering index
di Diameter of input aperture to a CPC
do Diameter of output aperture of a CPC
ds Diameter of small aperture of a CPC
D(µ,λ) Donaldson matrix
E Irradiance
E⊥ Illuminance normal to the illumination
E0 Axial irradiance
Ee Edge irradiance
Ei Image irradiance
Ei0 Axial image irradiance
Eλ Spectral irradiance
f/# F-number
f/#w Working F-number
F′ Increase factor
Fa to b Form factor from a to b
fskew Skew invariant
I Intensity
ILED A Averaged LED intensity, CIE condition A
ILED B Averaged LED intensity, CIE condition B
Iλ Spectral intensity
L Radiance
L*,a*,b* CIE 1976 (L*a*b*) color space; CIELAB
L*,u*,v* CIE 1976 (L*u*v*) color space; CIELUV
Li Image radiance
Lo Object radiance
Lλ Spectral radiance
xii
 Glossary (cont.)

M Integrating sphere multiplier

m Lateral image magnification
n Index of refraction
NA Numerical aperture
OD Optical density
pe Purity
pf Packing fraction
psa Projected solid angle
R Reflectance factor
RA Coefficient of retroreflection
Ra General color rendering index
RI Coefficient of retroreflected luminous intensity
RL Coefficient of retroreflected luminance
Sλ(λ) Spectral density of a light source
T Transmittance factor
T/# T-number
u, v CIE 1960 UCS chromaticity coordinates
u′, v′ CIE 1976 UCS chromaticity coordinates
V(λ) Photopic luminous efficiency
V′(λ) Scotopic luminous efficiency
W*U*V* CIE 1964 uniform space coordinates
x, y CIE 1931 chromaticity coordinates
X, Y, Z CIE tristimulus values
x̄, ȳ, z̄ CIE color matching functions
β Entrance angle (in retroreflection)
ξ Generalized étendue
λ Wavelength, emission wavelength
λ0.5m Center wavelength (for LED)
λc Centroid wavelength (for LED)
λd Dominant wavelength
λp Peak wavelength (for LED)
µ Excitation wavelength
ρ Reflectance
ρ̄ Average reflectance
τ Transmittance
υ Viewing angle (in retroreflection)
Φ Flux
Φλ Spectral flux
Ω Projected solid angle (psa)

xiii
 Glossary (cont.)

ω Solid angle
Ω
⎯ a to b Average projected solid angle from a to b
Ωi Input psa to a compound concentrator
Ωo Output psa from a compound concentrator
Ωx Projected solid angle viewed from plane x
θi Input half-angle of compound concentrator
θo Output half-angle from compound concentrator
θmax Maximum output half-angle from CPC

xiv
Basic Quantities in Illumination 1

Flux and Irradiance

In examining terminology for illumination, it is useful to

separate the spatial considerations from the spectral
concerns. In many cases, the spatial and spectral issues
are independent and can be separated without losing any
generality. In other cases, the spatial and spectral issues
cannot be separated physically, but it is useful to separate
them conceptually. The commonly used spatial quantities
are flux, irradiance, intensity, and radiance.

Flux, Φ, is the optical power or rate of flow of radiant

energy.

Irradiance, E, is the flux per unit area striking a surface.

Occasionally, the flux per unit area leaving a surface,
called exitance, M, is important. However, the geometry
is the same as for irradiance, so it will not be treated
separately here. Furthermore, when exitance is used, it is
often the flux leaving a nonphysical surface such as the
exit port of an integrating sphere or the real image in an
imaging system, where it is identical to the irradiance
onto the surface.

The irradiance quantity itself says absolutely nothing

about the directionality of the flux. For example, if the
three cases in the figure below all have the same flux per
unit area striking the surface, then they all have the
same irradiance. Because of this ambiguity, specifications
for illumination systems often qualify the irradiance
quantity with an added description of the desired
directional properties.

normal oblique diffuse

collimated collimated illumination
illumination illumination
2 Illumination

Solid Angle

The definition of intensity involves the concept of a solid

angle. A solid angle is a 3D angular volume that is
defined analogously to the definition of a plane angle in
two dimensions.

A plane angle, θ, made up of the lines from two points

meeting at a vertex, is defined by the arc length of a circle
subtended by the lines and by the radius of that circle, as
shown below. The dimensionless unit of plane angle is the
radian, with 2π radians in a full circle.

A A
B l
B

θ θ θ =l/r
r (radians)
O O
2π radians in
a full circle

A solid angle, ω, made up of all the lines from a closed

curve meeting at a vertex, is defined by the surface area
of a sphere subtended by the lines and by the radius of
that sphere, as shown below. The dimensionless unit of
solid angle is the steradian, with 4π steradians in a full
sphere.

Closed curve area, a,

on
surface of
sphere

r ω=a/r2 (steradians)
ω

ω 4π steradians in
a full sphere
Basic Quantities in Illumination 3

Intensity, Radiance, and Projected Solid Angle

Intensity, I, is the flux per unit solid angle. It is the

amount of flux from a point source contained in a small
angular volume. A source can be considered a point source
for this application if the irradiance falls off as the
inverse square of the distance from the source.
Intensity, for a given source, can vary with direction.

The term “intensity” is used in many disciplines, some

even closely related to optics, to mean things other
than flux per unit solid angle. Use caution and rely on
context to determine the meaning of the word in a
particular situation.

Radiance, L, applies to extended sources and surfaces. It

is the flux per unit solid angle per unit projected area of
the source or surface. The projected area is the projection
of the area onto a surface normal to the direction of view
and is equal to the actual area times the cosine of the
angle between the surface normal and the direction of
view. Radiance can vary with position on a surface, and
like intensity, it can vary with direction. A source or
surface with constant radiance in all directions is called
Lambertian. A Lambertian source or surface has
intensity that varies with the cosine of the angle with the
surface normal.

In many cases, the angle of view changes over the extent

of the receiver. These cases require an alternate definition
of radiance: radiance is the flux per unit area per unit
projected solid angle. (In fact, this is the more general
definition and covers the simpler case where the entire
surface of the extended source is at essentially the same
angle as the direction of view.)
4 Illumination

Solid Angle and Projected Solid Angle

The relationship between solid angle and projected solid

angle can be confusing. Projected solid angle has meaning
primarily for a small Lambertian source, which has
intensity that varies as the cosine of the angle with the
surface normal. The projected solid angle, Ω, is the
solid angle, ω, weighted by the cosine of the angle with
the surface normal.
ω

Ω = ω cos ϕ
ϕ

When the solid angle is large enough so that the angle

with the surface normal is not the same over the entire
solid angle, the total projected solid angle must be
computed by integrating the incremental projected solid
angles. See the reference by Bartell for a more detailed
explanation.

For some special cases, the integration results in simple

expressions, such as for a large circular cone that is
normal to a surface and subtends a half angle, θ.

Ω = π sin2 θ

A hemisphere has 2π steradians (solid angle) but π

projected steradians (projected solid angle).
Basic Quantities in Illumination 5

Spectroradiometric and Radiometric Quantities

In the spectral dimension of illumination, the most

general view looks at the spectral density—the amount of
radiation per unit wavelength interval. In terms of the
four spatial quantities already considered, the spectral
quantities are spectral flux, Φλ; spectral irradiance,
Eλ; spectral intensity, Iλ; and spectral radiance, Lλ.
These quantities, usually written with a subscript to
indicate that they are integrable, must be integrated to
determine the amount of radiation in a particular spectral
band. For example, the total radiant flux, Φ (in units of
watts), in the band between wavelength λ1 and
wavelength λ2 is
λ2
Φ ( λ1, λ2) = ∫Φ
λ1
λ ( λ ) ⋅ dλ.

Similar expressions can be written for the total

irradiance, E (watts/m2); total radiant intensity, I
(watts/ sr); and total radiance, L (watts/m2·sr).

Photometry measures the response of the human eye to

light. Although not everyone has exactly the same
response, the standardized CIE 1924 luminous
efficiency function works very well for most people.
(The CIE is the International Commission on
Illumination.) This function, shown on the following page,
is designated V(λ). The values for this function, in 5-nm
increments, are given in the Appendix. Not coincidentally,
this function is identical to the CIE color matching
function, ȳ. The unit of luminous (photopic) flux is the
lumen. The luminous flux is found from the spectral flux
and the V(λ) function from the following relationship:

luminous flux = 683∫ Φ λ ( λ ) ⋅ V ( λ ) ⋅ dλ.

The factor of 683 in this equation comes directly from the

definition of the fundamental unit of luminous intensity,
the candela.
6 Illumination

Photometric Quantities

CIE 1924 Luminous Efficiency

(Vλ)

1.0
Luminous Efficiency

0.8

0.6

0.4

0.2

0.0
350 400 450 500 550 600 650 700 750 800
wavelength (nm)

Notes on notation:
• The photopic quantities of flux, irradiance,
intensity, and radiance are called luminous flux,
illuminance, luminous intensity, and
luminance, respectively.
• These quantities are sometimes notated with a
subscript “v” (for visual), as Φv, Ev, Iv, and Lv. But
often the subscript is omitted since the meaning is
usually clear from the context, and it could be
confused with the subscript notation often
reserved for integrable quantities.
• The designations Φ, E, I, and L are common but
not universally standard. Another set of symbols
sometimes used is P, H, J, and N, respectively, for
radiometric quantities; Pλ, H λ, J λ, and N λ for
spectral quantities; and F, E, I, and B for the
corresponding photometric quantities.
• Solid angle and projected solid angle are not
always distinguished by ω and Ω, respectively.
Basic Quantities in Illumination 7

Matrix of Basic Quantities

SPECTRAL
Radio-
metric Spectral Photopic
Power Power/
wavelength Luminous flux
Flux interval
watts (W) watts/nm lumens (lm)
Illuminance
Irradiance Spectral
Flux/ irradiance lm/m2
S or
area
P W/m2 W/m2·nm lux
A (Luminous)
T (Radiant) Spectral intensity
I Flux/ intensity intensity
A solid lm/sr
L angle or
W/sr W/sr·nm candela (cd)
Radiance Spectral Luminance
radiance
Flux/ lm/m2·sr
area· or
solid cd/m2
angle or
W/m2·sr W/m2·sr·nm nit
The table above shows the four spatial quantities and the
three spectral categories that are discussed in the
preceding pages. These create 12 distinct cells that cover
the vast majority of specifications for illumination
systems.
With two exceptions, both used mainly in the United
States, work in illumination is almost always done in SI
units. The two exceptions (both deprecated) are:
Illuminance
1 footcandle (lm/ft2) = 10.764 lux (lm/m2)
Luminance
1 footlambert (candela/πft2) = 3.426 nit (candela/m2)
8 Illumination

Photopic and Scotopic Vision

The human visual system responds to light over a wide

dynamic range, in excess of 6 orders of magnitude. To
achieve this dynamic range, the mechanisms for high-
light-level vision and low-light-level vision are different.
The high-level region, called the photopic region, is
active at luminance levels above about 3 cd/m2. The low-
level region, called the scotopic region, is active below
approximately 0.01 cd/m2. The region between pure
photopic and pure scotopic is called the mesopic region,
where the visual response is a mixture of the two. The
photopic efficiency, usually designated V(λ), peaks at
555 nm, while the scotopic efficiency, usually
designated V′(λ), peaks at 507 nm.

Scotopic and Photopic

Luminous Efficiency

1.0
Luminous Efficiency

0.8
0.6
0.4
0.2
0.0
350 400 450 500 550 600 650 700 750 800
wavelength (nm)

The values for photopic efficiency and scotopic efficiency,

both in 5-nm increments, are given in the Appendix.
Basic Quantities in Illumination 9

Luminous Efficacy

Luminous efficacy, quantified in lumens per watt, is a

measure of the ability of a light source to produce a visual
response from its power. In the photopic region, luminous
efficacy peaks at 683 lumens per watt at 555 nm. In fact,
the lumen is defined in terms of the power at 555 nm
(frequency of 540 × 1012 Hz). Specifically, the definition
(adopted in 1979) is in terms of the candela (lumen per
steradian).

The candela is the luminous intensity, in a given

direction, of a source that emits monochromatic
radiation at a frequency of 540 × 1012 Hz and that has
a radiant intensity in that direction of 1/683 watt per
steradian.

Luminous Efficacy
(photopic)

700
Luminous Efficacy (lm/W)

600
500
400
300
200
100
0
350 400 450 500 550 600 650 700 750 800
wavelength (nm)

It is usually clear from the context whether the power is

the radiated power (as in the discussion above) or, often
for lamps, the “wall-plug” power.
10 Illumination

Typical Values of Illumination Quantities

Irradiance and Illuminance

Direct sunlight 1000 W/m2 (250–2500
nm)
Direct sunlight 100,000 lux
Shade 10,000 lux
Overcast day 1,000 lux
Office space 300–600 lux
Full moon 0.2 lux
Quarter moon 0.01 lux
Moonless clear night 0.001 lux
Luminous Intensity
Automobile headlight 5,000–20,000 cd
Household flashlight 100–1,000 cd
100-W tungsten lightbulb 100 cd
LED traffic signal 250–700 cd
Single LED 1 mcd–25 cd
Radiance and Luminance
Sun 2 × 107 W/m2·sr
(250–2500nm)
Sun 2 × 109 nit
Frosted lightbulb 100,000 nit
Fluorescent lamp 5,000 nit
Computer screen 100 nit

Wavelength Ranges for Illumination

UV-C* 250 to 280 nm
UV-B 280 to 315 nm
UV-A 315 to 400 nm
Visible ~360–400 to ~760–800 nm
Near-infrared (NIR)† 760 nm to 1.1 μm
* Actual definition of UV-C is 100 to 280 nm. However, the
range from 100 to 250 nm is not of interest for
illumination systems.
† Actual definition of NIR is to 1.4 μm. However, 1.1 μm is

the upper limit for silicon-based detectors.

Basic Quantities in Illumination 11

Averaged LED Intensity

In 1997 the CIE established a special quantity for light-

emitting diodes (LEDs) called the averaged LED
intensity. This was introduced because, as stated in CIE
Publication 127:2007, Measurement of LEDs, “There are
significant differences between LEDs and other light
sources which made it necessary for the CIE to introduce
a new quantity for their characterization with precisely
defined measurement conditions.”

To obtain averaged LED intensity, the LED is

measured on its mechanical axis (in line with the
package) by a circular detector of area 100 mm2 at a
prescribed distance from the front tip of the LED package.
Two distances are used: 316 mm (condition A) and 100
mm (condition B), with the solid angles defined as 0.001
sr and 0.01 sr, respectively. The measurements made are
notated as ILED A and ILED B in units of intensity (candela
or W/sr). Since the entire measurement geometry is
completely defined, the measurements should be
repeatable.

Circular aperture
detector
LED of area, A = 100
mm2
Mechanical axis

316 mm – condition A (0.001 sr)

100 mm – condition B (0.01 sr)
12 Illumination

Light Source Color

The perceived color of a light source is quantified by its

chromaticity. Chromaticity is calculated from the
spectral density of the light source, Sλ, and the CIE color
matching functions, x , y , and z as follows:

X
X = ∫ Sλ ( λ ) ⋅ x ( λ ) ⋅ d λ , x= ,
X +Y + Z
Y = ∫ Sλ ( λ ) ⋅ y ( λ ) ⋅ dλ ,
Y
Z = ∫ Sλ ( λ ) ⋅ z ( λ ) ⋅ dλ , y= ,
X +Y + Z

where X, Y, and Z are called tristimulus values, and x

and y are the CIE 1931 chromaticity coordinates. The
integrals above are usually calculated as block
summations from 360 to 830 nm, generally at 1-nm or 5-
nm intervals. A table of x , y , and z , in 5-nm intervals can
be found in the Appendix.

The 1931 chromaticity coordinates (x, y) are common

coordinates for light-source colors and are represented
graphically by the familiar “horseshoe” graph. All of the
possible colors of light are contained inside the horseshoe
shape, with the pure monochromatic spectral colors
around the curved perimeter, the purples along the
straight line at the bottom, and less-saturated colors in
the interior. The various shades of white, which are of the
most interest in illumination systems, occupy the central
region.

Those white lights that have near-blackbody spectra (such

as tungsten incandescent lamps) lie along the Planckian
locus. The lower blackbody temperatures lie toward the
red, and the higher temperatures toward the blue.
Color 13

Chromaticity Diagram

1931 x,y chromaticity diagram

0.8

GREEN

0.6 YELLOW

y ORANGE

0.4 WHITE

RED
Planckian locus

0.2 BLUE

PURPLE

0
0 0.2 0.4 0.6 0.8
x

Two other coordinate systems are used to describe the

chromaticity of light sources: the CIE 1960 UCS
coordinate system (u, v), and the CIE 1976 UCS
coordinate system (u′, v′). Both attempt to portray equal
perceived color differences by equal distances.

u = 4 X ( X + 15Y + 3Z ) = 4 x ( −2x + 12 y + 3),

v = 6Y ( X + 15Y + 3Z ) = 6 y ( −2x + 12 y + 3),
u′ = 4 X ( X + 15Y + 3Z ) = 4 x ( −2x + 12 y + 3),
v′ = 9Y ( X + 15Y + 3Z ) = 9 y ( −2x + 12 y + 3) .

The CIE 1960 UCS coordinate system is obsolete except

for calculating correlated color temperature.
14 Illumination

Color Temperature and CCT

Any light source whose chromaticity coordinates fall

directly on the Planckian locus has a color temperature
equal to the blackbody temperature of the Planckian
radiator with those coordinates. Color temperature is
usually expressed in Kelvins (K). The concept of color
temperature is especially useful for incandescent lamps,
which very closely approximate a blackbody spectrum
throughout the visible region. For these lamps, the color
temperature also defines the spectrum in this region.

For white lights that don’t have chromaticity coordinates

that fall exactly on the Planckian locus but do lie near it,
the correlated color temperature (CCT) is used. The
CCT of a light source, also expressed in Kelvins, is defined
as the temperature of the blackbody source that is closest
to the chromaticity of the source in the CIE 1960 UCS
(u, v) system. CCT is an essential metric in the general
lighting industry to specify the perceived color of
fluorescent lights and other nonincandescent white-light
sources such as LEDs and high intensity discharge HID
lamps.

The difference in perceived color is closely related to the

reciprocal of CCT. The reciprocal is expressed in
reciprocal megakelvin (MK)-1, with one (MK)-1
approximately equal to a just-noticeable color difference:

( MK )−1 = 106 CCT .

There are limitless different spectra, all with the same

CCT, that may have little or no resemblance to the
blackbody curve for that temperature or to each other.

There is no approved method for computing CCT nor is

there a simple and accurate closed-form expression. One
simple and accurate method is to use a program such as
Excel with solver to find the blackbody temperature that
minimizes the distance between its (u, v) coordinates and
those of the light in question.
Color 15

Dominant Wavelength and Purity

Colored light sources can be modeled as a mixture of a

monochromatic source and a white light. The wavelength
of this theoretical monochromatic source is called the
dominant wavelength, λd, and is the perceived color of
the light. The percent of the total power provided by the
monochromatic source is the purity, pe.

1931 x, y chromaticity diagram

a y − yw x − xw
pe = = =
a + b yd − yw x d − x w
0.8

GREEN

0.6 YELLOW

Dominant
Wavelength
y ORANGE Point (xd,yd)

• •
0.4 b
White point a
(xw, yw) • light
(x,y)
RED

0.2 BLUE

PURPLE

0
0 0.2 0.4 0.6 0.8
x

The choice of the “white point” is arbitrary. Very often,

the default choice is an “equal energy white” (x = 0.3333,
y = 0.3333).
16 Illumination

Surface Color

The color of a surface, like that of a light source, can be

quantified. The reflectance factor, R(λ), of the surface is
combined with the spectral density of the illumination, Sλ,
and the CIE color matching functions, x , y , and z , in the
calculation of tristimulus values:

X = k ∫ Sλ ( λ ) ⋅ R( λ ) ⋅ x ( λ ) ⋅ dλ,
Y = k ∫ Sλ ( λ ) ⋅ R( λ ) ⋅ y( λ ) ⋅ dλ,
Z = k ∫ Sλ ( λ ) ⋅ R( λ ) ⋅ z ( λ ) ⋅ dλ,
100
k= .
∫ S ( λ ) ⋅ y ( λ ) ⋅ dλ
λ

Or, in summation form:

780 780
Y =k ∑
λ= 380
Sλ ( λ ) R( λ ) y( λ ), X =k ∑ S (λ ) R(λ ) x ( λ ),
λ = 380
λ

780 780
Z =k ∑ S (λ ) R( λ ) z (λ ) ,
λ = 380
λ k = 100 ∑ S ( λ ) y( λ ).
λ= 380
λ

The chromaticity of light sources is 2D, with the

photometric value of luminous flux, illuminance,
luminous intensity, or luminance playing the role of the
third “dimension.” However, the chromaticity of surfaces
is 3D, with a “lightness” dimension included. Common 3D
surface-color spaces are derived from the tristimulus
values.

x, y, Y – CIE 1931 chromaticity plus Y tristimulus value

L*,u*,v* – CIE 1976 (L*u*v*) color space; CIELUV

L*,a*,b* – CIE 1976 (L*a*b*) color space; CIELAB

W*U*V*– CIE 1964 uniform space coordinates (obsolete

except for the calculation of color rendering index)
Color 17

Color of Fluorescent Surfaces

The phenomenon of fluorescence is characterized by the

absorption of light at one wavelength and the nearly
instantaneous emission at a longer wavelength. Various
surfaces of interest in illumination systems, such as road
signs designed to convert the blue-rich skylight at twilight
to more visible yellow, also exhibit intentional fluorescent
properties to make them brighter or more detectable.

The calculation of the chromaticity of fluorescent surfaces

has a degree of complexity that is not present for
nonfluorescent surfaces. For fluorescent surfaces, the
reflectance factor R(λ) is replaced by the Donaldson
matrix, D(μ,λ), where μ is the absorbed or excitation
wavelength, and λ is the emission wavelength:

780 780
X =k ∑ ∑ S (μ) D(μ, λ ) x (λ ),
λ = 380 μ= 300
μ

780 780
Y =k ∑ ∑ S (μ) D(μ, λ ) y( λ ),
λ = 380 μ= 300
μ

780 780
Z =k ∑ ∑ S (μ) D(μ, λ ) z (λ ).
λ= 380 μ= 300
μ

Note that for nonfluorescent

surfaces, the Donaldson The normalization
matrix collapses to the constant, k, is the
reflectance factor for the same as for non-
diagonal elements (μ = λ), fluorescent surfaces.
and is zero everywhere else.
In fact, the reflectance factor can be thought of as a
special case of the Donaldson matrix.

The details of computing the chromaticity of fluorescent

surfaces can be found in ASTM Standard E2152-01,
Standard Practice for Computing the Colors of Fluorescent
Objects from Bispectral Photometric Data.
18 Illumination

Color Rendering and CRI

The color of an object depends on both the object and the

illuminating source. Some sources are better than others
at rendering the colors of objects. This ability is quantified
in a figure of merit called the color rendering index
(CRI) for the source. The CRI is calculated by comparing
the chromaticity coordinates of eight prescribed
nonfluorescent test-color samples (specified by their
spectral radiance factors) under the source being
evaluated with the chromaticity coordinates of the same
eight samples under a reference illuminant. The
reference illuminant is a blackbody radiator (with the
same CCT as the source) for sources with a CCT of less
than 5000 K, and a phase of daylight (with the same CCT
as the source) for sources with a CCT of 5000 K or higher.
The distance between the two chromaticity coordinates for
a particular test-color sample represents the color
difference between the sample illuminated by the source
being evaluated and the same sample illuminated by the
reference illuminant. The general color rendering
index, designated Ra, is the average distance (in W*, U*,
V* space) between the eight pairs of chromaticity
coordinates (a pair for each test-color sample). It is
normalized so a source that is identical to its reference
illuminant has a CRI of 100, and a “warm white”
fluorescent lamp has a CRI of about 50. Here are
examples of typical CRI values for several illuminating
sources:
Sunlight (CIE D65) 100
Tungsten lamp (CIE A) 100
Xenon 97
White light LED (blue + YAG) 83
Compact fluorescent lamp 80
Daylight fluorescent 75
Metal halide lamp 61
Warm white fluorescent 52
High-pressure sodium lamp 20
White light from RGB LED combination 20 to 65
White light from four LED combination up to 90
Color 19

Calculating CRI and Problems with CRI

Complete details for calculating CRI are beyond the scope

of this guide. The information is available in two CIE
publications (both are necessary):

• CIE 13.3-1995, Method of Measuring and Specifying

Color Rendering Properties of Light Sources
• CIE Publication 15:2004, Colorimetry, 3rd Edition

With these publications comes a software disk for

calculating CRI from a light source spectral density. The
publications and disk can be purchased from the CIE at
www.cie.co.at/cie/index.html.

Work on the CRI began in the 1940s when widespread use

of fluorescent tubes began for general lighting. The figure
of merit has come under criticism lately for several
reasons. Among them:

• The eight test-color samples are all moderate in

saturation. CRI does not produce numbers that
correspond well with observations on highly saturated
colors.
• The CRI is considered to be less accurate when the
test illuminant and the reference illuminant differ by
more than 0.0054 (in u, v space). Many real
illuminants are farther away than this from their
respective reference illuminants.
• Many different illuminants, such as sunlight and
tungsten lamps, have “perfect” CRIs near 100, yet
they render colors quite differently.
• New white light sources, such as combinations of
LEDs, seem to perceptually render colors far better
than is predicted by their rather low CRIs.

The CIE is expected to issue a new recommendation for

quantifying color rendering within the next several years.
20 Illumination

Typical Source Parameters

Spectral densities for several common light sources are

displayed on the following pages. The table below lists
some of the relevant parameters for these sources.

Typical values for common illumination sources

Approx. Luminous
CRI
Source CCT Efficacy
(Ra)
(K) (lm/W)
2800 12
Tungsten† to 100 to
3200 20
Sunlight 6500 100 100‡
Daylight
6360 75 55
fluorescent
Warm white
3000 52 60
fluorescent
High-pressure
2030 20 115
sodium
Metal halide 4020 61 90
5000
Xenon to 97 10 to 20
6000
† Tungsten lamps allow a great deal of control over these

parameters. Guidelines for controlling tungsten lamps

are detailed on the next page.
‡ Lumens per radiated watt. All others are “wall-plug.”

LEDs are becoming common illumination sources. They

produce white light by using a blue LED and one or more
phosphors. Depending on the phosphor used, they can
have CCTs from less than 3000 K to in excess of 9000 K.
The luminous efficacies also vary over a wide range, from
40 or 50 lm/W for high-flux devices and up to and above
100 lm/W for low-flux devices, with CRIs up to about 80
(Ra). White light is also produced from LEDs by mixing
the light from several single-color LEDs. LED lighting
technology is changing extremely rapidly as newer devices
are developed.
Sources for Illumination 21

Tungsten Lamps

Tungsten filament incandescent lamps, particularly

tungsten halogen lamps, are often used in illumination
systems. Paired with stable, current-controlled power
supplies, they provide extremely stable sources. If the
current is altered, the filament temperature changes,
with a resulting change in filament resistance. Therefore,
the voltage change is not linearly related to the current
change. Similar nonlinear relationships hold for other
characteristics of the lamp, such as power consumed,
luminous flux output, efficacy (lumens/Watt), color
temperature, and lifetime.
Tungsten halogen lamps should be operated between 95%
and 105% of their rated current. Within this operating
envelope, the relationships between the operating
parameters are approximately exponential. Modifying the
current to a value different from the rated current
changes the other parameters according to the exponents
shown in the table below:

Parameter Exponent
Color temperature (K) 0.80
Lifetime (hours) –25
Luminous efficacy 3.6
(lumens/Watt)
Luminous flux (lumens) 6.5
Power (Watts) 2.9
Voltage (volts) 1.9

An example of a lamp operated at 5% over its rated

current is shown below:
Parameter Rated Operated
Current 6.02 A 6.02 × 1.05 = 6.32
Color temperature 2900 K 2900 × 1.050.8 = 3015
Lifetime 1000 hrs 1000 × 1.05-25 = 295
Luminous efficacy 16 lm/W 16 × 1.053.6 = 19
Luminous flux 1040 lm 1040 × 1.056.5 = 1428
Power 65 W 65 × 1.052.9 = 75
Voltage 10.8 V 10.8 × 1.051.9 = 11.8
22 Illumination

Tungsten and Sunlight

This page and those that follow show typical spectra of

several common illumination sources.

Tungsten Lamp (CIE A)

350 400 450 500 550 600 650 700 750

wavelength (nm)

Sunlight (CIE D65)

350 400 450 500 550 600 650 700 750

wavelength (nm)
Sources for Illumination 23

Fluorescent Lamps

Warm White Fluorescent

350 400 450 500 550 600 650 700 750

wavelength (nm)

Daylight Fluorescent

350 400 450 500 550 600 650 700 750

wavelength (nm)
24 Illumination

H.P. Sodium and Metal Halide

High Pressure Sodium

350 400 450 500 550 600 650 700 750

wavelength (nm)

Metal Halide

350 400 450 500 550 600 650 700 750

wavelength (nm)
Sources for Illumination 25

Xenon and White LEDs

Xenon

350 400 450 500 550 600 650 700 750

wavelength (nm)

White Light LED (Blue + YAG)

350 400 450 500 550 600 650 700 750

wavelength (nm)
26 Illumination

Light Emitting Diodes (LEDs)

LEDs are moderately narrowband emitters with an

approximately Gaussian spectral shape. The spectrum of
an LED is often expressed by a single wavelength, with
four different single-wavelength descriptions in general
use. The most common spectrum-based description is the
peak wavelength, λp, which is the wavelength of the
peak of the spectral density curve. Less common is the
center wavelength, λ0.5m, which is the wavelength
halfway between the two points with a spectral density of
50% of the peak. For a symmetrical spectrum, the peak
and center wavelengths are identical. However, many
LEDs have slightly asymmetrical spectra. Least common
is the centroid wavelength, λc, which is the mean
wavelength. The peak, center, and centroid wavelengths
are all derived from a plot of Sλ(λ) versus λ. The fourth
description, the dominant wavelength, λd, is a
colorimetric quantity that is described in the section on
color. It is the most important description in visual
illumination systems because it describes the perceived
color of the LED.

Spatially, LEDs, especially those in lens-end packages,

are often described by their viewing angle, which is the
full angle between points at 50% of the peak intensity.

LEDs

350 400 450 500 550 600 650 700 750

wavelength (nm)
Illumination Properties of Materials 27

Transmittance, Reflectance, and Absorptance

Several alternative methods describe the response of

materials to illumination. One common approach is the
ratio of the light that is transmitted, reflected, or
absorbed to the incident light. This method describes a
material by its transmittance, τ, its reflectance, ρ, or
its absorptance, α. Do not confuse absorptance with
absorbance, A, which is equivalent to optical density
(OD) and is a conversion of transmittance or reflectance
to a log scale. For example, 10% transmittance can be
described as 1A, 1% as 2A, etc.

Transmitted
Incident light, A light, B

Reflected
light, C Transmittance = B/A
Reflectance = C/A
Absorptance = (A-B-C)/A

A material that produces intensity proportional to the

cosine of the angle with the surface normal is called
Lambertian. The radiance of a Lambertian surface is
constant with viewing direction (since the projected area
of a viewed surface is also proportional to the cosine of the
angle with the surface normal). Furthermore, the
directional distribution of scattered light is independent
of the directional distribution of the incident illumination.
It is impossible to tell, by looking at a Lambertian
surface, where the incident light comes from. Perfectly
Lambertian surfaces don’t really exist, but many
materials, such as matte paper, flat paint, and
sandblasted metal (in reflection), as well as opal glass and
sandblasted quartz (in transmission), are good
Lambertian approximations over a wide range of
incidence and view angles.
28 Illumination

Reflectance Factor and BRDF

A quantity sometimes confused with reflectance is the

reflectance factor, R. The reflectance factor is defined
in terms of a hypothetical perfectly reflecting diffuser
(PRD), a surface that is perfectly Lambertian and has a
100% reflectance. The reflectance factor is the ratio of the
amount of light reflected from the material to the amount
of light that would be reflected from a PRD if similarly
illuminated and similarly viewed.
A A
B C

PRD

R = B/C
100% reflectance
Lambertian

Notes on reflectance (ρ) and the reflectance factor (R):

• For a Lambertian surface, ρ and R are identical.
• Reflectance must be between 0 and 1. The
reflectance factor is not similarly bound. A highly
polished mirror, for example, has near-zero R for any
nonspecular incident and viewing angles, and a very
high R (>1.0) for any specular incident and viewing
angles.
• The reflectance factor is more closely related to the
bidirectional reflectance distribution function
(BRDF) than to reflectance. The BRDF is defined as
the radiance of a surface divided by its irradiance:
BRDF = L/E.
• The reflectance factor measures per hemisphere
(there are π projected steradians in a hemisphere)
what the BRDF measures per projected steradian:
R = BRDF·π.
Illumination Properties of Materials 29

Harvey / ABg Method

The Harvey or ABg method is used to parameterize

scatter from a weakly scattering surface, which is typical
for optical surfaces such as lenses and mirrors. It also can
be used to model Lambertian surfaces and anisotropic
(i.e., asymmetric) scatter. An example for a three-axis
polished surface is provided here, which has a total
integrated scatter (TIS or TS) of about 1.6%. The
vertical scale represents the BSDF, for which an R
(reflection) or T (transmission) can be substituted for the
S (surface). The horizontal scale represents the absolute
difference between β0 = sinθ0, or the specular direction,
and β = sinθ, or any direction away from specular. Note
that both axes are plotted in log space such that the roll-
off slope is linear. The ABg parameters are:
• g is the slope of the roll-off as shown in the figure
whose value of 0 defines a Lambertian surface.
• B is the roll-off parameter defined as

g
B = βrolloff .

• A is the amplitude factor and can be found from

A
BSDF = g
.
B + β − β0

Rolloff point:
100
Fit Beta: 0.00773
BSDF: 45.5
1

B Data ABg:
S 0.01 A: 3.14e-006
D Rolloff B: 6.89e-008
F 0.0001 g: 3.39
1E-006
βrolloff Integrated scatter:
0.01596
1E-005 0.0001 0.001 0.01 1 100
Data
|β – βo| Signature
Synthetic
30 Illumination

Directional Properties of Materials

The reflectance of a material can depend on the direction

of the incident light. This dependence is often indicated by
a number or letter.
• ρ(0°): reflectance for normal incidence.
• ρ(45°): reflectance for a 45-deg oblique incidence.
• ρ(d) or ρ(h): reflectance for diffuse illumination.
The reflectance factor of a material can depend on both
the direction of illumination and the viewing geometry.
This is usually indicated by two letters or numbers, the
first indicating the incident geometry and the second the
viewing geometry.
• R(0°/45°): the reflectance factor for normal incidence
and a 45-deg oblique viewing (a common geometry for
measuring the color of a surface).
• R(0°/d): the reflectance factor for normal incidence
and diffuse (everything except the specular) viewing
only.
• R(8°/h): the reflectance factor for near-normal
incidence and hemispherical (everything, including
the specular) viewing.
• R(45°/h): the reflectance factor for hemispherical illu-
mination and a 45-deg oblique viewing.
The same notation used for reflecting materials can
be applied to transmitting materials, where trans-
mittance τ can be dependent on incident geometry,
and the transmittance factor, T, on both the incident
and transmitting geometries. The use of the
transmittance factor is not as common as
transmittance, reflectance, and the reflectance
factor.
Some materials have reflecting properties that are not the
same for every azimuthal angle, even for the same
elevation angle, e.g., the specular geometry of mirrorlike
surfaces has vastly different reflecting properties than
any geometry with the same incident and reflecting
elevation angles that are not both in the same plane with
the surface normal.
Illumination Properties of Materials 31

Retroreflectors—Geometry

Retroreflectors reflect incident light back toward the

direction of the light source, operating over a wide range
of angles of incidence. Typically they are constructed in
one of two different forms, 90-deg corner cubes or high
index-of-refraction transparent spheres with a reflective
backing. Retroreflectors are used in transportation
systems as unlighted night-time roadway and waterway
markers, as well as in numerous optical systems,
including lunar ranging. Some are made of relatively
inexpensive plastic pieces or flexible plastic sheeting, and
some are made of high-priced precision optics.

The performance of retroreflectors is characterized within

a geometrical coordinate system, usually with three
angles for the incident and viewing geometries and a
fourth orientation angle for prismatic designs like corner
cubes, which are not rotationally isotropic in their
performance. All the geometric variations are described in
detail in ASTM E808-01, Standard Practice for Describing
Retroreflection, along with expressions for converting from
one geometric system to another.

Two angles commonly used to specify the performance of

retroreflectors are the entrance angle, β, and the
observation angle, α. The entrance angle is the angle
between the illumination direction and the normal to the
retroreflector surface. High-quality retroreflectors work
over fairly wide entrance angles, up to 45-deg or more (up
to 90 deg for pavement marking). The observation angle,
the angle between the illumination direction and the
viewing direction, is generally very small, often one
degree or less.

Another useful angle for interpreting the performance of

retroreflectors is the viewing angle,υ, the angle between
the viewing direction and the normal to the retroreflector
surface.
32 Illumination

Retroreflectors––Radiometry

The performance of retroreflectors is quantified by several

coefficients. These are the most common:

RI, coefficient of retroreflected luminous intensity,

I
RI = ,
E⊥
where E⊥ is the illuminance on a plane normal to the
direction of illumination, and I is the intensity of the
illuminated retroreflector.

RA, coefficient of retroreflection,

RI I A
RA = = ,
A E⊥
where A is the area of the retroreflector.

RL, coefficient of retroreflected luminance,

RA L
RL ==
cos υ E⊥
is the ratio of the luminance in the direction of
observation to E⊥.

RΦ, coefficient of retroreflected luminous flux:

RA
RΦ = .
cos β
RF, retroreflectance factor
π ⋅ RI π ⋅ RA π ⋅ RL
RF = = = .
A ⋅ cos β ⋅ cos υ cos β ⋅ cos υ cos β

It is the retroreflectance factor, RF that is numerically

equivalent to the reflectance factor, R.
Retroreflectors are often specified by the coefficient of
retroreflection, RA, for various observation angles and
entrance angles.
Values for RA of several hundred (cd/m2)/lux are not
uncommon, corresponding to reflectance factors up to and
over 1000.
Illumination Transfer 33

Lambertian and Isotropic Models

There are no direct “conversion factors” between the four

basic quantities in illumination: flux, Φ; irradiance, E;
intensity, I; and radiance, L. But for many situations,
knowledge of one factor allows the calculation of the
others. Making this calculation usually requires
knowledge of the directional properties of the illuminating
source, or at least a fair model of these directional
properties. The two most common models are isotropic
and Lambertian.

An isotropic source is defined here as having intensity

independent of direction. For a Lambertian source, the
radiance is independent of direction and the intensity is
therefore proportional to the cosine of the angle with the
surface normal. A few nearly isotropic sources exist, such
as a round, frosted light bulb, a frosted ball-end on a fiber,
and a line filament (in one plane, anyway). However, most
flat radiators, diffusely reflecting surfaces, and exit pupils
of illuminating optical systems are more nearly
Lambertian than isotropic. Reasonable predictions can be
made by modeling them as Lambertian.

The model of directional illumination properties need only

apply, of course, over the range of angles applicable to
your particular situation. In many cases, the mutually
contradictory models of an isotropic and a Lambertian
source are used simultaneously. This is valid over small
angular ranges where the cosine of the angle with the
surface normal doesn’t change much. This assumption is
not all that restricting. For example, for a small
Lambertian source illuminating an on-axis circular area,
the error in flux caused by using an isotropic model is less
than 1% for a subtended full angle of 22 deg [NA = 0.19,
f/2.6], less than 5% for a full angle of 50 deg [NA = 0.42,
f/1.2], and less than 10% for 70 deg [NA = 0.57, f/0.9].
However, for a full angle of 180 deg (a full hemisphere),
the error is 100%!
34 Illumination

Known Intensity

Consider a small source at a distance. For a known

intensity that is essentially constant over all relevant
directions, i.e., toward the illuminated area:

ω ξ
Ar I E
L Φi Ai
θ
d

where I is the intensity of the radiating area in the

direction of the illuminated area;
Ar is the radiating area;
θ is the angle between the normal to the radiating area
and the direction of illumination;
Ar cosθ is the projected radiating area as viewed from the
illuminated area;
Ai is the illuminated area;
ξ (xi) is the angle between the normal to the illuminated
area and the direction of illumination (assumed
constant over this small angular range);
d is the distance between the two areas (assumed to be
constant);
ω is the solid angle formed by the illuminated area when
viewed from the radiating area (assumed to be small);
Ω = ω cosθ is the corresponding projected solid angle (for
small solid angles);
E is the irradiance at the illuminated area;
Φi is the total flux irradiating the illuminated area; and
L is the radiance of the radiating area.

I cos ξ I
E= , Φ i = I ω, L= .
d2 Ar cos θ
Illumination Transfer 35

Known Flux and Known Radiance

If, in the same situation, the flux within the solid angle
is known, then the intensity is
I = Φi ω ,
the irradiance is
E = Φi Ai ,
and the radiance is
Φi Φi
L= = .
ω Ar cos θ Ω Ar
Consider the same situation, but not necessarily with a
small radiating area or small illuminated area:

L ω,Ω
ξ
Ar E
Φi Ai
θ
I d
If the radiance is known and the radiating area is small,
then
I = L Ar cos θ .
If cosθ is essentially constant from all points on the
radiating area to all points on the illuminated area, then
Φi = L Ar ω cos θ = L Ar Ω .
If cosθ varies substantially over the illuminated area,
then the second form of this equation, using the projected
solid angle, should be used.
Since there are π projected steradians in a hemisphere,
the total flux radiated (for a Lambertian radiator) is
Φ r = L Ar π .
The irradiance at the illuminated area (E) is
Φ i L Ar Ω L Ar cos θ cos ξ
E= = = = L Ωi ,
Ai Ai d2
where Ωi is the projected solid angle of the radiating area
when viewed from the illuminated spot.
36 Illumination

Form Factor and Average Projected Solid Angle

Here the approximations of constant cosines cannot be

used.
The angles between
L Φi the normal to the
A radiating surface
Ai
and the directions
Φr to points on the
illuminating sur-
face vary not only with the locations of the points on the
illuminated surface, but also with the locations of points
on the radiating surface. The concept of projected solid
angle takes the former into account, but not the latter.
What is needed is an average projected solid angle,
Ω⎯ r to i, which is the projected solid angle subtended by the
illuminated area and averaged over all points on the
radiating area. Then the illuminating flux, Φi, from a
Lambertian radiator is
Φ
Φi = L Ar Ωr to i = r Ωr to i .
π
In practice, the average projected solid angle is not used.
However, its geometrical equivalent, called the form
factor, Fa to b, is used. The only difference between the
form factor and the average projected solid angle is a
multiplier of π:
Fa to b = Ωa to b π .
The form factor measures in hemispheres what the
average projected solid angle measures in projected
steradians. The form factor also can be interpreted as the
portion of the flux leaving a Lambertian radiator, a, that
illuminates a surface, b:
Φi = Φ r Fr to i .
Note that the form factor is directional, as are the solid
and the projected solid angles. Fa to b is not in general
equal to Fb to a. However, the product of the area and the
form factor is constant:
Aa Fa to b = Ab Fb to a .
Illumination Transfer 37

Configuration Factor

The form factor and the average projected solid

angle both link two extended areas. The form factor
measures in hemispheres what the average projected
solid angle measures in projected steradians. Another
term, the configuration factor, C, is similarly related to
the projected solid angle, linking a small area with an
extended area. Like the form factor, the configuration
factor measures in hemispheres what the projected solid
angle measures in projected steradians:

C = Ω π.

Tables of configuration factors and form factors for

myriad geometries can be found in handbooks on
illumination, in books on radiative heat transfer (where
the issues are identical to illumination by Lambertian
radiators), and on the Internet. Three cases with
applicability to many optical situations are listed here:

Case 1: Small area to an extended circular area; both

areas parallel and with axial symmetry.
radius, r
θ

r2
C= = sin 2 θ
r + d2
2

and

π r2
Ω= = π sin 2 θ.
r2 + d2
38 Illumination

Useful Configuration Factor

Case 2: Small area to an extended circular area; both

areas parallel, but without axial symmetry.

radius, r
offset, x

d
θ δ

⎛ ⎞
⎜ ⎟
( ) ( )
2 2

1⎜ 1+ d − r ⎟
C = ⎜1 − x x ⎟
1
2⎜
{ }
⎡ 2
⎤ 2 ⎟
( ) ( ) ( )
2 2 2
⎜ ⎢ 1 + d + r −4 r ⎥ ⎟
⎜ x x x ⎟
⎝ ⎣⎢ ⎦⎥ ⎠

or, equivalently:

⎛ ⎞
1⎜ 1 + tan 2 δ − tan 2 θ ⎟
C = ⎜1 − 1 ⎟
2⎜
⎝ ⎣ ( )( )
⎡tan 4 δ + 2 tan 2 δ 1 − tan 2 θ + sec 4 θ ⎤ 2
⎦
⎟
⎠

and

Ω = π ⋅ C.

These expressions degenerate to the expressions for case 1

above when x, or equivalently, δ, is equal to zero.
Illumination Transfer 39

Useful Form Factor

Case 3: An extended circular area illuminating another

extended circular area; both areas parallel and centered
on the same axis.
radius, rr radius, ri

radiating illuminated
area, Ar area, Ai

⎧ 1
⎫
⎪ 2 ⎡⎛ 2
⎞
2
⎤2 ⎪
⎛ ri ⎞ ⎢ ⎛ ri ⎞
⎪ 1+⎜ ⎟ ⎜ 1+⎜ ⎟ ⎟ 2⎥ ⎪

Fr to i
1⎪
= ⎨1 + ⎝ d ⎠ − ⎢⎜1 + ⎝d⎠ ⎟ − 4 ⎛ ri ⎞ ⎥ ⎪ .
2 ⎢⎜ 2 ⎟ ⎜ ⎟ ⎥ ⎬
2⎪ ⎛ rr ⎞ ⎢⎜ ⎛ rr ⎞ ⎝ rr ⎠ ⎥ ⎪
⎜d⎟ ⎜d⎟ ⎟
⎪ ⎝ ⎠ ⎢⎜⎝ ⎝ ⎠
⎟
⎠ ⎥ ⎪
⎪ ⎣ ⎦ ⎪
⎩ ⎭

Some numerical values of Fr to i for this case are shown in

the table below for several sizes of radiating and
illuminated disks (each expressed as a multiple of the
distance between the two parallel circular areas that are
centered on the same axis).

Form Factor, F r to i
ri/d
0.03 0.10 0.30 1.00 3.00 10.0
0.03 .001 .010 .083 .500 .900 .990
0.10 .001 .010 .082 .499 .900 .990
0.30 .001 .009 .077 .489 .899 .990
rr/d
1.00 .000 .005 .044 .382 .890 .990
3.00 .000 .001 .009 .099 .718 .989
10.0 .000 .000 .001 .010 .089 .905
40 Illumination

Irradiance from a Uniform Lambertian Disk

Many illumination situations can be modeled as

illumination by a uniform circular Lambertian disk,
with the illuminated area parallel to the disk and at some
distance from it.

radius, r offset, x

Lr
Ei

d
δ θ
Radiating area Illuminated spot

The irradiance at the illuminated spot is equal to the

radiance of the radiating area times the projected solid
angle of the radiating area when viewed from the
illuminated spot:
Ei = Lr Ωi .
If the illuminated spot is on axis (x = 0, δ = 0), then
r2
Ei = π Lr sin 2 θ = π Lr 2 .
r + d2
If the spot is offset from the axis, it is necessary to use the
projected solid angle or the configuration factor discussed
previously for case 2:

⎧ ⎫
π⎪ 1 + tan 2 δ − tan 2 θ ⎪
Ωi = ⎨1 − 1 ⎬
.
2⎪ ⎡
⎩ ⎣
4
( 2
)( 2
)4
⎤
tan δ + 2tan δ 1 − tan θ + sec θ ⎦
2
⎪
⎭

Note: The configuration factor, form factor, and

projected solid angle are useful mainly when the
radiation pattern is Lambertian or nearly Lambertian.
Illumination Transfer 41

Cosine Fourth and Increase Factor

Consider the previous case of illumination by a uniform

circular Lambertian disk, with the illuminated area
parallel to the disk and at some distance from it. For
many values of aperture size (θ) and field angle (δ), the
irradiance falls off very nearly at cos4δ, a phenomenon
often referred to as the cosine-fourth law.

Two of the cosine terms in the cos4 law are due to the
fact that, off axis, the distance increases with the
cosine of δ and the inverse square law applies. The
third cosine factor comes from the Lambertian source,
and the fourth from the fact that the illuminated
surface is inclined to the direction of propagation.
In reality, the cos4 “law” is not exactly true, and is far
from true for large values of θ and δ. The table below
displays values of the increase factor, F′, which is the
multiplier that must be applied to the irradiance
calculated by using the axial irradiance and cos4 falloff. F′
compensates for the inaccuracy in the “cosine-fourth”
assumption:
Ei = π Lr sin 2 θ ⋅ cos4 δ ⋅ F ′
Increase Factor, F′
θ
(deg) 1.8 3.6 7.2 14.5 30.0 45.0
NA 0.03 0.06 0.13 0.25 0.50 0.71
f/# 16 8 4 2 1 0.71
0 1.00 1.00 1.00 1.00 1.00 1.00
10 1.00 1.00 1.00 1.01 1.03 1.05
20 1.00 1.00 1.01 1.03 1.11 1.20
δ(deg)
30 1.00 1.00 1.01 1.05 1.23 1.49
40 1.00 1.00 1.02 1.08 1.37 1.94
60 1.00 1.01 1.02 1.09 1.48 2.69
The cos4 approximation is valid within a few percent up to
very large apertures and field angles.
42 Illumination

Known Irradiance

.I
ξ
E
Φi Ai
E
d

If a surface is illuminated by a source of uniform intensity

at a distance d and the irradiance on the surface is
known, then the intensity of the source is

Ai E
E ⋅ d2
I= .
cos ξ
E For any surface that is illuminated by
uniform irradiance, the total flux
illuminating the surface is

Φ = E ⋅ Ai .
The radiance of the surface, caused by the light reflecting
from the surface, depends on the reflecting properties of
the surface.
If the surface is Lambertian over all angles of reflection
(for this incident geometry), then

ρ⋅E
L= ,
π
where ρ is the reflectance of the surface for the relevant
incident geometry.
If the surface is not Lambertian over all angles but is
Lambertian over the direction of concern, then
R⋅E
L= ,
π
where R is the reflectance factor of the surface for the
relevant incident geometry and for the direction of
concern.
Illumination Transfer 43

ω, Ω, NA, and f/# for a Circular Cone

The case of a circular disk subtending a known half-angle,

θ, shows up often in illumination situations.

There are at least four common ways of describing the

cone: solid angle (ω), projected solid angle (Ω),
numerical aperture (NA), and f-number (f/#):

ω = 2π(1 − cos θ) Ω = π sin 2 θ

NA = n ⋅ sin θ f /# = 1 2sin θ ,

where n is the index of refraction.

Cone subtended by a circular disk

θ(deg) ω Ω NA/n f/#
1.8 0.003 0.003 0.03 16.00
3.6 0.012 0.012 0.06 8.00
7.2 0.049 0.049 0.13 4.00
12.7 0.154 0.152 0.22 2.27
14.5 0.200 0.196 0.25 2.00
20.0 0.379 0.367 0.34 1.46
25.0 0.589 0.561 0.42 1.18
30.0 0.842 0.785 0.50 1.00
35.0 1.14 1.03 0.57 0.87
40.0 1.47 1.30 0.64 0.78
45.0 1.84 1.57 0.71 0.71
50.0 2.24 1.84 0.77 0.65
60.0 3.14 2.36 0.87 0.58
70.0 4.13 2.77 0.94 0.53
80.0 5.19 3.05 0.98 0.51
90.0 6.28 3.14 1.00 0.50
44 Illumination

Invariance of Radiance

Unlike intensity, which is associated with a specific point,

and irradiance, which is associated with a specific surface,
radiance is associated with the propagating light rays
themselves. This distinction is not trivial and implies that
the radiance of a surface can be considered separate from
the actual physical emitter or reflector that produces the
radiance.

Consider a uniform Lambertian radiating source, Ar, with

radiance, Lr, illuminating an area, Ai, through a limiting
aperture that limits the solid angle of the source to ω:

Lr
Ai
ω

Ar

The physical location of the radiating source is irrelevant.

Only the solid angle matters. In fact, the physical location
(and shape) can be assumed to be anywhere (and any
shape) as long as the solid angle is the same. All of the
following descriptions of the radiating area, A1, A2, and A3,
are equivalent to Ar from an illumination point of view:

Lr
Ai
ω

Ar
A2 A3
A1
Lr Lr
Lr
Illumination in Imaging Systems 45

Image Radiance

In an imaging system with no vignetting or significant

aberrations, for Lambertian objects, point-by-point, the
radiance of an image is equal to the radiance of the object
except for losses due to reflection, absorption, and
scattering. These losses are usually combined into a
single value of transmittance, τ. This equivalence of
radiance is true for virtual as well as real images, and for
reflective or refractive imaging systems.

Real Image

Imaging
System
τ
Real
Object
Image
Lo
Li = τ·Lo
Virtual Image

Imaging
System
τ
Virtual Object
Image Lo
Li = τ·Lo

Viewed from any point on a real image, the entire exit

pupil of the optical system is also the radiance of the
corresponding object point but reduced by τ.

Imaging
System
Object Lexit pupil = τ·Lo
Lo exit pupil
46 Illumination

Limitations on Equivalent Radiance

In all cases, the image radiance only exists when the

image is viewed through the exit pupil of the imaging
system. When viewed in a direction that doesn’t include
the pupil, the radiance is zero.

Imaging
Li = τ·Lo
System

Li = zero
exit pupil
If the object is not Lambertian, then the angular
distribution of radiance of the image is also not
Lambertian. The relationship between the angular
distributions of object and image radiances is not
straightforward and must be determined by ray tracing
on the specific system. However, in many practical cases,
the entrance pupil of the imaging system subtends a
small angle from the object, and the source is essentially
Lambertian over this small angle.

If the object and the image are in media of different

refractive indices, no for the object and ni for the image,
then the expression for equivalent radiance is

Li L
= τ ⋅ 2o .
ni2 no

The point-by-point equivalence of radiance from object to

image is only valid for well-corrected optical systems. For
systems that suffer from aberrations or are not in focus,
each small point in the object is mapped to a “blur spot” in
the image. Thus, the radiance of any small spot in the
image is related to the average of the radiances of the
corresponding spot in the object and its surrounding area.
Illumination in Imaging Systems 47

Image Irradiance

Since the exit pupil, when viewed from the image, has the
radiance of the object, then the irradiance at the image is
the same as the irradiance from a source of the same size
as the exit pupil and the same radiance as the object
(reduced by τ). In most imaging systems, the exit pupil is
round and the irradiance is the same as the irradiance
from a uniform Lambertian disk:

δ θ Image
Exit pupil, L = τ·Lo

Ei = π τLo sin 2 θ ⋅ cos4 δ ⋅ F ′ .

A table of values for the increase factor, F', is presented

in the section on illumination transfer. F' is very close to
1.0 except for a combination of large field angle (δ) and
large aperture (θ), which is not a common combination in
imaging systems.

The cos4δ term contributes to substantial field darkening

in wide-angle imaging systems—for example, cos445 deg =
0.25.

If the physical aperture stop is not the limiting aperture

for all the rays converging to an off-axis image point, the
light is vignetted. The irradiance at image points where
there is vignetting will be lower than predicted.

On axis, cos4δ = 1.0 and F' = 1.0. The image irradiance on

axis, Ei0, is

Ei 0 = π τLo sin 2 θ .
48 Illumination

f/#, Working f/#, T/#, NA, Ω

For a camera working at infinite conjugates (distant

object, magnification, |m|<<1), the image irradiance can
be expressed in terms of the lens’ f-number, f/#:
π τ Lo
Ei 0 = .
4 ( f /#)2

This f/#, usually associated with a lens, is an “infinite

conjugates” quantity. When a lens is used at finite
conjugates, the working f-number, f/#w, describes the
cone angle illuminating the image:
f /#w = ( f /# ) ⋅ (1 − m),

where m is the lateral magnification of the image

(negative for real images), and the axial image irradiance
is:
π τ Lo
Ei 0 = .
4 ( f /#w )2

Note that f/#w degenerates to the conventional “infinite

conjugates” f/# when the lens is used at infinite
conjugates.
Occasionally, a lens will be designated with a T-number,
T/#, which combines the f/# and the transmittance into a
single quantity,
f /# π Lo
T /# = with axial image irradiance: Ei 0 = .
τ 4 (T /#)2

Another descriptor of the image illumination cone angle is

the numerical aperture, NA,
NA = sin θ with axial image irradiance: Ei 0 = π τLo NA 2 .

In all cases, even without circular symmetry, on or off

axis, the cone illuminating the image can be described by
its projected solid angle, Ω, with image irradiance:
Ei = τLo Ω .
Illumination in Imaging Systems 49

Flux and Étendue

The total flux reaching the image is the product of the

image irradiance and the area of the image. The image
irradiance is proportional to the projected solid angle of
the exit pupil when viewed from the image:

Φi = τLo ai Ωi ,

where Ωi is the projected solid angle of the exit pupil

viewed from the image, ai is the area of the image, Lo is
the [assumed uniform] radiance of the object, and τLo is
the radiance of the exit pupil.

The flux reaching the image also can be expressed in

terms of the radiance of the exit pupil, τLo, the area of the
exit pupil, ap, and the projected solid angle of the image
when viewed from the exit pupil, Ωp:

Φi = τLo a p Ω p .

The quantity aΩ, representing the area of a plane in the

optical system times the projected solid angle of another
plane when viewed from it, appears equivalently in both
expressions. This area-solid-angle-product is a
fundamental property of the optical system that
determines the amount of light that can get through the
system. It is called the throughput or étendue.

The radiance of an object is invariant and cannot be

increased by an optical system, and the étendue is a
fundamental property of an optical system. These two
concepts mean that, for a source of given radiance and
a given optical system, the maximum flux that can be
transmitted through the system is predetermined.

And, without “throwing away” light, the étendue cannot

be decreased, but area and solid angle can be traded off.
50 Illumination

Generalized Étendue

The terminology for illumination in nonimaging systems

is the same as that for imaging systems; however, the
range of validity is extended to include all angular space,
while that of imaging systems is limited to paraxial
systems. With this taken into account, étendue is often
called generalized étendue. In this domain the étendue
cannot be regarded as the simple product of the area and
solid angle; it must be integrated per the following
equation and figure:

E = n2 ∫∫ cos θdAs dω,

dAs
aperture Aperture

where n is the refractive

index, θ is the angle θ
from the normal, dAs is
the differential source
area, and dω is the
dω
differential solid angle.

The total flux through

the aperture is found by
integrating the radiance over the aperture:

Φ= ∫∫ L ( r, aˆ ) cos θdAs dω,

aperture

where r and â denote the positional and directional

aspects of source emission. Assuming that the source is
Lambertian so radiance is independent of angle, then

Ls E
Φ = Ls ∫∫
aperture
cos θdAs dω =
n2
.

Note that total flux is the product of the radiance and the
geometrical étendue factor. This also shows the
conservation of étendue that follows from the
conservations of radiance and energy.
Illumination in Nonimaging Systems 51

Concentration

Concentration (C) is a term associated with the

generalized étendue. It represents the ability to transfer
more light into a desired area by using the conservation of
étendue to alter the angle at the output of an optical
system. It is defined as the ratio of the input area (A) to
the output aperture area (A′) that transmits the
prescribed flux from area A. For this reason it is called
the concentration ratio:

C = A A′ .

This expression, a limit factor of the laws of

thermodynamics, is a forbearer of the invariance of
radiance and étendue.

In a 2D system, which is analogous to an extruded trough,

and a 3D system, which is analogous to a well, we find
that the respective concentrations are given by

2
a n′ sin θ′ A ⎛ n′ sin θ′ ⎞
C2D = = and C3D = = ,
a ′ n sin θ A ′ ⎜⎝ n sin θ ⎟⎠

where a and a′ are the aperture widths, A and A′ are the

aperture areas, n is the input index, n′ is the output
index, θ′ is the output angle, and θ is the input angle.
Optimal concentration is realized when the output
angle is π/2, giving

2
n′ ⎛ n′ ⎞
C2D,opt = and C3D,opt =⎜ ⎟ ,
n sin θa ⎝ n sin θa ⎠

where θa, the acceptance angle, is the prescribed upper

input angle over which conservation of étendue is
maintained.
52 Illumination

Skew Invariant

The skew invariant is another limiting factor in

nonimaging system design. Its definition is rather
esoteric:
dE ( s )
fskew ( s ) = , y
ds
x
where s = rminkt, and rmin is the
ray’s closest approach to the
optical axis (z, as shown), and
kt is the tangential com-
ponent of the ray’s prop- kt
agation direction.

A simpler way to think about rmin

the skew invariant is to z
recognize that in a rota-
tionally symmetric system (e.g., a lens), loss is introduced
from the input to the output if the two spatial
distributions are not the same shape. For example, if the
object shape is a uniform square but a uniform round
output is desired, then transfer losses will be produced.

To maximize transfer efficiency with different

distributions, the symmetry of the optical system must be
broken; or, in other words, there must be a “twist” in the
optical components to force rays out of their respective
sagittal planes. Many nonimaging optical systems take
advantage of this property by including faceted reflectors
(e.g., segmented headlights), segmented lenses (e.g.,
pillow optics for projection displays), or 3D edge-ray
concentrators that employ V-wedges near the source (i.e.,
solar concentrators).

For a rotationally symmetric system, the rotational

skewness of each ray is conserved or invariant.
This skew invariant is given by the first derivative
of the étendue.
Fibers, Lightpipes, and Lightguides 53

Fibers—Basic Description

Optical fibers, lightpipes, and lightguides are all

variations on the same theme. They each contain a
central transparent core, usually circular in cross-section,
surrounded by an annular cladding. The cladding has a
lower index of refraction than the core.

Core

Cladding

The core can transmit light for long distances with low
loss because of total internal reflection at the interface
between the core and the cladding. The primary purpose
of the cladding is to maintain the integrity of this
interface. Without it, total internal reflection would occur
at a core-air interface, but dust, nicks, abrasions, oils, and
other contamination on the interface would reduce the
transmission to unacceptably low levels.

Sometimes layers of buffering and/or jacketing are

placed outside the cladding for additional protection.

The core diameter can range from very small, on the order
of the wavelength of light, to a centimeter or more. The
very thin cores are essentially waveguides and not used
for illumination. Flexible glass and quartz fibers have
core diameters ranging from approximately 50 microns to
about 1 millimeter. If they are thicker than that, they are
rigid and called rods or light pipes. Plastic fibers are
flexible at thicker core diameters. Sometimes liquid cores
and plastic cladding are used to make flexible, high-
transmittance lightguides that are over a centimeter in
core diameter.
54 Illumination

Numerical Aperture and Étendue

The maximum angle that a fiber can accept and transmit

depends on the indices of refraction of the core and
cladding (as well as the index of the surrounding medium,
usually air, n0 = 1).

Index = n0 cladding, index = n2

θmax core, index = n1

1
sinθmax = n12 − n22
n0
and the NA is
NA = n0 sinθmax = n12 − n22 .
The fiber has a maximum acceptance projected solid
angle, Ω = πsin2θmax, and an acceptance area, the cross-
sectional area of the core. Together, they define a
throughput or étendue for the fiber in air:
π2 2
Étendue = d NA 2 ,
4
where d is the core diameter.
This étendue defines the maximum flux-carrying
capability of the fiber when presented with a source of
radiance.

Note: A fiber illuminated at less than its maximum

acceptance angle will, theoretically, preserve the
maximum illumination angle at its output. However,
bending and scattering at the core-cladding interface
broadens this angle toward the maximum allowable.
This effect is not important in illumination systems in
which it is desirable to utilize the maximum étendue of
low-throughput components such as fibers and fill the
full input NA.
Fibers, Lightpipes, and Lightguides 55

Fiber Bundles

To achieve high throughput with flexible glass or quartz

fibers, multiple fibers are often arranged in a bundle,
such as the 19-fiber tightly packed bundle shown below:

core

cladding

interfiber
space

The ratio of the light-carrying core area to the area of the

entire bundle is called the packing fraction (pf), and
can be as high as 85%. This packing fraction reduces the
effective area of the bundle and, correspondingly, its
étendue.

In addition to flexibility, fiber bundles have other possible

advantages in illumination systems:

• Shape Conversion: In some situations, such as

when illuminating a spectrometer, it can be useful to
convert a circular cross-section of fibers to a line cross-
section to align with, or actually become, the entrance
slit to the spectrometer.

• Splitting the Bundle: By feeding a large fiber

bundle with a single light source and splitting the
bundle into two or more branches, it is possible to
illuminate multiple locations, from multiple angles,
with one source.

• Mixed Bundle: When illuminating with light over a

wide spectral band, such as the full solar spectrum
(~250 to 2500 nm), a mixed bundle of high OH silica
fibers for good UV transmission and low OH silica
fibers for good IR transmission can compensate for the
lack of an adequate single-fiber type.
56 Illumination

Tapered Fibers and Bundles

By tapering a single fiber, it is possible to trade off

between area and solid angle while keeping the product
(étendue) approximately constant.

cladding
input output
area, ai area, ao
core
input output NA,
NA, NAi NAo

ai·NAi2 ≈ ao·NAo2

On the other hand, when a bundle of straight fibers is

tapered, the tradeoff is between the area and packing
fraction.

input output
area, ai area, ao

input output
packing packing
fraction, fraction,
pfi pfo

input NA, output

NAi NA, NAo

NAo = NAi = NAfiber

ao·pfo = ai·pfi
Classical Illumination Designs 57

Spherical Reflector

The light emitted from a source in the direction away

from the optical system can be redirected toward the
optical system by using a spherical mirror with the
source located at the center of curvature.

source
and
image
of
source

If the source is solid, it is necessary to place the source

slightly away from the center of curvature and the image
just above, below, or alongside the physical source.

Ignoring losses on
source
reflection, the image has
the same radiance as the
source, but the effective
image source area (source plus
image) is doubled.

Sometimes this technique is used to place the image of

a source in a location where the physical source itself
could not fit because of an obstruction such as a lamp
envelope or socket.
If the source is not solid, such as a coiled wire tungsten
filament, imaging the source almost directly onto itself
can help fill in the area between the coils.

source In this case, the effective area of

the source is not appreciably
image increased, but the apparent
radiance is nearly doubled.
58 Illumination

Abbe Illumination

Abbe illumination is characterized by imaging the

source (or imaging an image of the source) directly onto
the illuminated area. Since the uniformity of illumination
is directly related to the uniformity of source radiance,
Abbe illumination requires an extended source of uniform
radiance such as a well-controlled arc, a ribbon filament
lamp, the output of a clad rod, a frosted bulb, an
illuminated diffuser, or the output of an integrating
sphere.

The paraxial layout below shows Abbe illumination used

in a projection system. The source is imaged by a
condenser onto the film. The projection objective images
the film and the image of the source onto the screen. The
purple dotted lines show the marginal and chief rays from
the source. The black dotted lines show the marginal and
chief rays from the film (and the image of the source). The
marginal rays go through the on-axis points on the object
and image and on the edges of the pupils (which are the
lenses in this case). The chief rays go through the edges of
the object and image and the on-axis points of the pupils.

screen
source projection
image, lens
source
film

condenser
Classical Illumination Designs 59

Köhler Illumination

Köhler illumination is used when the source is not

uniform, such as a coiled tungsten filament. Köhler
illumination is characterized by imaging the source
through the film onto the projection lens. The film is
placed adjacent to the condenser, where the illumination
is quite uniform, provided the source has a relative
uniform angular distribution of intensity.

The paraxial layout below shows Köhler illumination used

in a projection system. The source is imaged by a
condenser onto the projection lens. The projection
objective images the film onto the screen. The purple
dotted lines show the marginal and chief rays from the
source. The black dotted lines show the marginal and
chief rays from the film.

source projection screen

source image lens

condenser,
film

With similar sources, similar condenser NAs,

source/condenser étendue as limiting étendue, and
similar screen sizes, the average screen irradiance
levels are the same for both Abbe and Köhler
illumination systems. The choice between the two
generally depends upon the type of source available.
60 Illumination

Ellipsoidal and Paraboloidal Mirrors

Very efficient collection of light from a source can be

achieved using an ellipsoidal mirror, placing the source
at one of the foci. The source is imaged at the other focus,
with light collected over more than a hemisphere.
ellipsoid

source

image

An alternative is to use a paraboloidal mirror to

collimate the light from a source and a lens to reimage it.
Again, the light from the source is collected over more
than a hemisphere.
paraboloid

source

lens

image

The forward light is usually ignored in both of these types

of designs.

In both cases, the image of the source may not be good

quality, but image quality may not be important in
illumination systems. Also, obstructions like lamp bases,
sockets, and mounting hardware can produce directional
anomalies in the radiance of the image.
If the quality of illumination is important, devices
such as lenslet arrays or faceted reflectors may be
used.
Classical Illumination Designs 61

Spectral Control and Heat Management

Specifications for illumination systems often contain

spectral requirements. Some of these requirements can be
partially met by the selection of lamp type, but usually
some sort of filtering is needed. Also, for visual systems,
especially those using tungsten lamps, unwanted heat
from infrared light may need to be removed. Again,
filtering is needed.

The simplest type of filter is the absorbing filter placed

in front of the light source. Filter glasses with a wide
range of spectral characteristics are available from glass
manufacturers. The primary concern with absorbing glass
filters is cracking from excessive absorbed heat.

Often a cracked filter will continue to work just fine.

Interference filters use multilayer thin-film coatings

that either transmit or reflect light at specific
wavelengths. Cracking is generally not a concern unless
the filter is made of an absorbing substrate. These filters
are available with a much wider variety of spectral
properties than absorbing filters, including narrow
bandwidth and sharp cut-off, and can be designed and
manufactured to achieve specific custom properties. They
are also available for different angles of illumination,
typically 0 deg and 45 deg.

Interference filters shift their spectral properties with

incident angle and therefore may not be suitable for
uncollimated light with a divergence of more than
about 10 deg from the axis.

Hot mirrors and cold mirrors are excellent ways to

manage heat that must be removed from a light source. A
hot mirror reflects infrared light and transmits visible
light. A cold mirror reflects visible light and transmits
infrared light. The reflector behind the light on a dentist’s
chair is a cold mirror.
62 Illumination

Illumination in Visual Afocal Systems

Afocal visual systems, such as binoculars, take the

collimated light from an extended distant object and
present collimated light to the eye, but with angular
magnification. Therefore, the object appears larger.
However, the apparent radiance (and therefore perceived
brightness) of the object is the same as that of the naked
eye, provided the size of the aperture stop is the same
with and without the binoculars. Without the binoculars,
the aperture stop is merely the eye pupil. With the
binoculars, the aperture stop is the smaller of:
• the pupil of the eye magnified by the angular
magnification, or
• the aperture of the objective lens.
In other words, if the collimated ray bundle entering the
eye from the binoculars is smaller than the eye pupil, the
apparent radiance of the object will be less with the
binoculars than with the naked eye. If the pupil of the eye
is the limiting aperture both with and without the
binoculars, the apparent radiance will be the same.
Binoculars are traditionally designated by two numbers,
the first being the angular magnification, the second the
diameter of the objective lens in mm. A light-adapted eye
pupil with a 2-mm diameter would remain the aperture
stop for all of the following common sizes of bird-watching
binoculars: 8 × 42, 8 × 32, 10 × 42, 6 × 25, and 10 × 25.
These binoculars are generally used during the day.
However, marine binoculars, which are used under all
lighting conditions, are typically 7 × 50 to accommodate a
7-mm-diameter dark-adapted eye pupil.

Note that a true point source, such as a star, will

have higher apparent intensity (and therefore appear
brighter) with binoculars than with the naked eye
because more light is collected with the binoculars,
but there is no angular magnification.
Uniform Illumination 63

Searchlight

A searchlight can provide uniform irradiance in three

dimensions that is extremely insensitive to the position of
the irradiated object.

A searchlight consists of a small circular Lambertian

source at the focal point of a collimating lens. Anywhere
inside the shaded area in the figure below, the source
appears as a circular disk at infinity, subtending a full
angle α. The entire extent of the source is visible, because
it does not completely fill the collimating lens. Since the
view of the source is the same anywhere inside this
region, the irradiance is the same.

Outside this region and beyond point P, the lens restricts

the area of the source that is visible. The lens itself
appears as a disk of the same radiance as the source. In
this region, the irradiance falls off as the square of the
distance from the lens.

source lens region of uniform

irradiance
P

angle α angle α

For real searchlights with small sources and large-

diameter lenses, the paraxial description above is not
exactly valid over the entire shaded region. However, over
a relatively small portion of this region, the irradiance is
extremely uniform. This region of irradiance uniformity
extends not only laterally, but longitudinally as well.

A searchlight provides a volume of uniform irradiance.

 64 Illumination

Source at a Distance

A small source at a distance from an object can provide

reasonably uniform irradiance across the object. It is
somewhat counterintuitive that a bare lamp filament, with
its obviously terrible radiance uniformity, can produce
excellent irradiance uniformity. For example, a small
(assumed Lambertian) lamp filament at 500 mm from a
flat object whose largest dimension is 150 mm will provide
irradiance uniformity across the object of better than 95%
(considering only cos4 falloff). The same lamp and object at
1.0-meter distance produces nearly 99% irradiance
uniformity.

Ee 1
5
500mm Eo 0
m
Irradiance uniformity ≡ Ee / Eo > 0.95 m

A source of uniform radiance can be created by

illuminating a transmission or reflection diffuser
with uniform irradiance.

A common calibration laboratory method used to realize a

standard of known radiance is to illuminate a reflection
diffuser, typically 50 mm in diameter, with a standard of
known irradiance, typically a calibrated 1000-W tungsten
halogen lamp (ANSI type FEL), at a 500-mm distance.
The irradiance uniformity across the diffuser is better
than 99.5%. If the reflectance factor of the diffuser is
uniform, the radiance uniformity of the standard is also
better than 99.5%.
Uniform Illumination 65

Mixing Rod

A mixing rod is a long piece of clear quartz, glass, or

plastic. Light entering one face of the rod undergoes
multiple total internal reflections emerging from the other
parallel face.

Due to the multiple reflections, the irradiance at the exit

face can be extremely uniform. In a well-designed and
illuminated rod, the radiance can be quite directionally
uniform as well. The directional uniformity of radiance
can be enhanced by placing a diffuser at the exit face of
the rod or simply frosting the rod-end itself.

Mixing rods can have any shape desired. The rods with
plane sides do a better mixing job in most cases.

Typically the rods have an aspect ratio (length to largest

transverse dimension) of about 10:1, and are usually
about 75- to 150-mm long. They can be clad like an optical
fiber, but generally are not. Unlike a fiber, the number of
reflections in a mixing rod is quite small, and losses are
not a serious problem.

Rather than using a rod with polished faces, it is

possible to achieve a similar effect using a mirrored
tube with a hollow center.

The combination of a rod and its illuminator are

sometimes designed by computer simulation. But the
degree of uniformity required doesn’t always demand this
level of complexity, so simple trial-and-error is often
sufficient.
66 Illumination

Bent Lightpipes

Complex lightpipes made from straight sections, bends,

and tapers are common in many industries. Bent
lightpipes are components used to mix or collect light
from different paths that bend around objects or provide
light output over an extended region. An example is
automotive dashboard illuminators that employ lightpipes
coupled to a small incandescent source or an LED. The
lightpipe allows the source light to be directed around
dials and knobs. The bends allow simple packaging and
lower costs at the expense of design complexity.
Any cross-sectional shape, bend angle, bend shape, and so
forth is possible, but the simplest is a single, right-angle
bend using common-center, circular bends and an
arbitrary cross section. A circular cross-section is shown
here. Two important slices are called principal sections:
the vertical (V), which shows the bend of the lightpipe,
and the horizontal (H), which shows the bend going into
the page. The vertical slice defines the transmission
properties of the lightpipe. For normally incident input
light coupled to the lightpipe, there are no propagation
losses except Fresnel losses if the bend ratio, R, is
R = r2 r1 = 1 + t r1 ≤ n,
r2
H V where r1 and r2 are the two
bend radii, t is the lightpipe
r1 t thickness in the vertical
section, and n is the
lightpipe index in air. As the
input angle increases, there
n are losses at the limit of this
equation, but the equation
is transcendental. By
decreasing the thickness of
the lightpipe, one can increase the acceptance angle such
that there is no loss.
More complex parameterization of lightpipes, including
uncommon bend centers, noncircular bends, and arbitrary
cross sections, can be found in the literature.
Uniform Illumination 67

Integrating Sphere

Integrating spheres produce illumination that has

extremely uniform radiance and irradiance. An
integrating sphere is a hollow spherical shell coated on
the inside with a highly reflecting diffuse coating. The
projected solid angle from any point on a sphere to any
element of area on the sphere is the same, regardless of
location. This fact combined with the diffuse coating and
the multiple reflections cause any light introduced into
the sphere to produce uniform irradiance on and radiance
of the wall of the sphere. A hole or “port” in the sphere
allows this uniform illumination to be used in an optical
system.

The radiance at the exit of an integrating sphere

extends to a full hemisphere (π projected steradians).
The irradiance at the wall of an integrating sphere is
incident from a full hemisphere.

The radiance, L, of the wall of an integrating sphere

generated by flux, Φ, introduced into the sphere is
Φ
L= ⋅ M,
π ⋅ As
where As is the area of the complete sphere wall, and M is
the “sphere multiplier,” which is equal to the average
numbers of reflections in the sphere. The multiplier, M, is
1
M = ,
1−ρ
where ρ is the average reflectance of the wall of the
sphere, counting the holes as areas of zero reflectance.

A good working model of an integrating sphere is to

consider the port to
port
∞ port be a hole in a wall,
and, at a totally
L = L arbitrary distance
? behind it, another
sphere -∞
wall of infinite extent
and radiance, L.
68 Illumination

Lenslet Arrays

Imaging illumination systems, whether single- or double-

lens systems, paraboloidal reflector and lens systems, or
single ellipsoidal reflector systems, all suffer from possible
nonuniformities in intensity (and consequently also in
irradiance). These are due, among other causes, to
possible nonuniformities in the source as well as
obstructions such as filament support wires, gas
discharge electrodes, and LED heat-sink structures.

These nonuniformities can be smoothed out by using a

lenslet array, an array (usually 2D) of small lenses.
Typically, the arrays are used in pairs. In the diagram
below, the dotted purple lines show the marginal rays for
one of the lenslets in the first array; the black dotted lines
show the marginal rays for the corresponding lenslet in
the second array.
1st array focusing
condensing 2nd array
lens
lens
illuminated
source target

In this configuration, the source is imaged by each lenslet

of the first array into the corresponding lenslet of the
second array. Each lenslet of the first array is imaged
onto the entire target. This overlaying creates uniform
illumination of the target. In effect, the lenslet arrays
create multiple Köhler illumination systems, all
superimposed on the target.
Lenslet arrays are generally designed using illumin-
ation design software.
Uniform Illumination 69

Small Reflectors, Lenslet Arrays, and Facets

Ellipsoidal and paraboloidal reflectors are often “small”

with respect to the lamp dimensions and the distances
between the lamp and the reflecting surfaces. In these
cases, in addition to the effects of lamp support
structures, the size and structure of the lamp itself can
produce nonuniformities in illumination.

One method of minimizing these nonuniformities is to

include a lenslet array in front of the detector. This
broadens the beam a little, depending on the f# of the
lenslets, but it can produce much more uniform
illumination than the reflector alone.
paraboloid lenslet array

Tandem lenslet arrays also can be used to minimize

the effects of small reflectors.

Another approach is to break the reflector into small flat

facets, either radially, circumferentially, or both.

Lenslet arrays and

faceted reflectors
are usually designed
with illumination
design software.
70 Illumination

Source Modeling Overview

A system software model, whether a simple paraxial

design or a detailed design of an illumination system, may
fail to agree with experimental results due to the lack of a
comprehensive source model. For the simplest case,
where the optics are far away from the source and collect
light over a small solid angle, a point source model or a
simple geometrical model of the source may be sufficient.
The directional distribution of light from these simple
models is usually assumed to be isotropic or Lambertian.
For more efficient designs with optics that are close to the
source and collect light over a large solid angle, a more
complete model of the source is required to obtain
meaningful results. These models must reflect the
physical size and shape of the source and should contain
directional distributions that account for factors such as
filament support wires and lamp envelopes.
Source models are made for all types of sources, including
LEDs, incandescent, fluorescent, metal vapor, and high-
pressure gas discharge sources. The modeling includes
spectral, radiance or luminance distributions, and lifetime
aspects. For example, accurate source models for the
following have been developed:
• The temperature distribution along an incandescent
filament varies from its ends to the center.
Additionally, the interior of the filament glows
“hotter” due to the re-incident radiation.
• Arc emission sources such as metal halide and HID
lamps change their radiance distribution and power
output over time due to ablation of the electrodes.
These lamps have a deposited material to capture this
ablation, called the “salt lake” in continuous sources
and the “getter” in a pulsed one.
There are essentially four ways of creating complex source
models. Three are described on the next page, while the
fourth, not presented here, is based on the physics of
emission. This method is outside the confines of this text.
Source Models 71

Source Modeling Methods

There are three source modeling methods, where the

accuracy of the model typically increases with number:
1. “Bottom-up” (geometrical model): the source
geometry starting with the electrodes, supports, and
envelope; finishes with the packaging. Emission is
assigned to the radiative components.
• Benefits: No complex measurements; handles re-
incident light; provides tolerancing capabilities.
• Limitations: Emission characteristics assumed;
approximate surface and material properties; can
include tedious CAD development.
2. “Top-down” (radiance model): the optical output of a
representative sample of the lamp. These mea-
surements are made with a goniometer, which
moves a detector around a lamp on two axes. A
camera measures the 2D radiance distribution of the
lamp from each of many goniometer positions. The
resulting 4D model represents a complete description
of the lamp that can be used in a computer optical
design program.
• Benefits: Emission is based on physical measure-
ments.
• Limitations: Does not handle reincident light; is
limited by the variance of the number of source
samples measured and aligned; and their complex
measurement.
3. “Bottom to top” (system model): Integrates the
bottom-up and top-down approaches to develop a more
thorough source model.
• Benefits: Complete geometrical and radiative
models.
• Limitations: Integration of two submethods.
There are many hybrid methods and methods based on
applying the physics of the emission process of a
prescribed source. Loosely, the first two methods show
agreement to within 25% of experimental results, while
the bottom-to-top method shows agreement within 10%.
In all cases, rays are assigned, typically in a Monte Carlo
approach, to the emission areas.
72 Illumination

LED Modeling

The components of an LED include the emitting die(s),

the lens, the reflecting dish, wire bond and pad, and
standoffs. Other components can include phosphors and
included detectors.
Lens
Geometrical modeling is useful to
develop LED sources; however, it Wire
is difficult to obtain or measure Bond
the shapes and sizes of the Dish
components within the lens.
Radiance modeling suffers Die
because of the large amount of
variance between LED samples of
Standoffs
one model. The primary issues
are the die position within the reflecting dish, the axial
position of the die and dish with respect to the lens
vertex, and the size and shape of the reflecting dish. Four
distinct methods are available for LED modeling:
1. Develop a flat object and assign rays to the surface
based upon the intensity distribution provided by the
manufacturer. This method ignores spatial variation
of the emission.
2. Develop a geometrical model of the LED and assign
rays to the emitting surfaces of the die. Optimize the
dish shape (typically a cone), size, and the axial offset
of the die-dish to the lens vertex. The lens shape must
be measured and the die and dish placed at the
transverse center of the lens. The model is complete
when the intensity pattern from the manufacturer
agrees with the ray-trace model.
3. Same as method #2, except develop the layer
structure within the die to generate Monte Carlo rays
within the active layer(s). This method is tedious for
ray tracing due to the index of refraction discontinuity
between the die (n = 2.5+) and the epoxy lens (n =
1.45+).
4. Radiance by itself or a system: integrated into #2 or
#3.
Source Models 73

Incandescent Lamp Modeling

The components of an incandescent lamp include the

base, filament(s), supports, and the envelope. Other
components can include
coatings and envelope faceting. Envelope
Note that the shapes and sizes
of components depend on the
application. Sources developed
for the automotive headlight Filament
industry provide the highest
level of tolerance from one Supports
sample to another. 0

Both geometrical and radiance

methods are useful for incandescent source modeling.
Radiance modeling is better suited to this source since a
goniometer can be focused on the filament source, while
the glass envelope supplies little effect on overall optical
ray paths. Only light rays that are re-incident
(approaching grazing incidence) on the envelope show
adverse effects. Geometric modeling involves breaking the
glass envelope to gain access to the internal components.
This process requires the use of calipers to measure the
coil spacing, the thicknesses and lengths of the
components, and the number of coils. Provided
parameters can help with this process:
• Maximum overall length (MOL): Overall distance
that includes the base and pins.
• Light center length (LCL): Distance between the
center of the emitter and a defined reference plane.
• Filament type: Designated by @-#, where @ is a
series of letters (e.g., C = coiled, CC = coiled coil, and
SR = straight ribbon), and # is a number providing an
arbitrary pattern for the filament supports.
• Bulb type: Designated by @-#, where @ is the bulb
shape (e.g., T = tubular), and # is the diameter in
eighths of an inch.
• Base type: Innumerable types that have no
shorthand notation to describe them. Examples
include screw, mogul, bipin, and prong.
74 Illumination

Arc and Fluorescent Lamp Modeling

The components of an arc lamp include the base(s),

electrodes, and envelope(s). Other components can include
coatings, salt lake (continuous)
or getter (pulsed), and ignition Arc Envelopes
Region
wire (flashlamp). The optical
radiation is represented by a
virtual object called the arc.
Note that the aspects of
components depend on the
application. Automotive head-
light arcs provide the highest Electrodes
level of accuracy from one
sample to another.
Radiance modeling is especially suited to this source since
geometrical modeling cannot effectively represent the arc.
The arc must be approximated with a cylinder, tube, or
some other geometric shape. Radiance modeling is also
suited to this source because a goniometer can be focused
on the arc. Due to the typical smaller sizes of these
sources compared to incandescent sources, the effect of re-
incident rays is more pronounced. Thus, methods to
integrate a simplified measurement of the radiance
distribution into the geometrical model have been
employed. One such method uses the Abel transform
based on a single image capture of the arc. The Abel
transform assumes symmetry of the arc shape and
revolves it around a localized centroid of the arc source.
Such system models are the most effective way to model
such sources.
Fluorescent lamps include the tube and base(s). These
are the simplest sources to model other than the complex
geometry of compact fluorescent lamps now available.
After the geometry is entered, the inner surface of the
tube acts as the emitter. Internally, mercury vapor is
excited, releasing UV radiation, which is then converted
into visible light upon being incident on the phosphor.
Geometrical modeling is better suited to this source due to
the large size and simplicity of the configurations.
Nonimaging Compound Concentrators 75

Nonimaging Compound Concentrators

Nonimaging compound concentrators were first

developed for solar energy collection to concentrate the
irradiance from the sun. In solar collection, depending
on the degree of sophistication of the sun tracking system,
the range of sun input angles can be fairly small. The
collectors (solar cells, water pipes, etc.) respond
essentially to irradiance and can be illuminated at any
angle. The compound concentrator trades off between
area and solid angle, presenting a large collection area to
the sun (collecting over a narrow solid angle) and
delivering the energy to a smaller area (and over a wider
solid angle). These devices come close to achieving the
theoretical maximum concentration (in three dimensions).

input projected output projected output

solid angle(Ωi) solid angle(Ωo) half-
input angle
half- Nonimaging (θo)
angle Compound
(θi) Concentrator
ai Ω i = a o Ω o

input area (ai) output area (ao)

If the output projected solid angle is the maximum, π,

(θo = 90 deg), the concentration is maximum:

ai 1
(max) = .
ao sin 2 θi

Nonimaging compound concentrators are designed using

the edge-ray principle, which directs all rays that are
at the maximum input angle (θi for θo = 90 deg in the
drawing above) to the edge of the output aperture. All
rays at input angles less than this maximum are directed
inside the output aperture with no concern for image
quality. Often this angle, θi, is called the acceptance
angle, θa.
76 Illumination

Concentrators as Luminaires

Nonimaging compound concentrators are used in

illumination as luminaires—devices used to direct the
light from a source for illumination. For illumination,
they are used in the reverse direction from their
configuration in solar collection; they collect light from as
large an angle as possible from a small source and direct
it over a smaller angle through a larger aperture. For
solar collection, they collect energy over a large area and a
small angle, delivering it to a small area.
small source, output projected
wide solid solid angle(Ωo) output
angle half-
angle
(θo)

output area (ao)

Nonimaging compound concentrators are efficient

because:
• They collect light from the source over a very large
solid angle.
• They are designed using the edge-ray principle,
keeping all the energy within the intended field.
Imaging systems are designed to be best on axis, with
the edges of the field “spilling over.” Nonimaging
compound concentrators are designed to be best at the
edges of the field, keeping all the energy inside the
design boundaries.

Nonimaging concentrators used as luminaires are usually

composed of an internal mirror surface with the figure of
a compound parabolic concentrator (CPC),
compound elliptical concentrator (CEC), or
compound hyperbolic concentrator (CHC). Dielectric
filled concentrators that employ total internal reflection
are also used.
Nonimaging Compound Concentrators 77

Compound Parabolic Concentrators

The compound parabolic concentrator (CPC) is a

common shape of nonimaging concentrator used for
illumination. A CPC is formed by a parabola with its
focus at one edge of the entrance (small) aperture, rotated
around an axis that is perpendicular to and through the
center of both apertures. CPCs can be quite long.
max.
field
A′ angle
(θmax)

A
axis of rotation

B′

B is the focus of parabola AA′.

A is the focus of parabola BB′.

The complete equation for the surface of a CPC can be

found in the equation summary in the Appendix.

The ratio of the diameter of the small and large apertures

is determined by maximum field angle
do 1
= ,
di sin θmax

where do and di are the diameters of the output (large)

and input (small) apertures, respectively.

The length of the concentrator is

do + di
Length = .
2tan θmax
78 Illumination

Compound Elliptical and Hyperbolic Concentrators

The compound elliptical concentrator (CEC) and

compound hyperbolic concentrator (CHC) work in
the domains of finite and diverging conjugates, while the
CPC worked at infinite conjugates. For CECs, the CPC-
development methods can be used. To visualize their
shapes, consider what is called the string method, where
the string acts as the edge ray:
Involute DA
• Choose two points, DA and DS.
• Select a length of string that
VA
allows the pen, P, to read points Absorber
VA and VS.
• While pulling the string taut
with P, sweep out the shape on
one side of the reflector. Source
• Flip the string to the other side VS
and repeat.
Involute DS
For a CPC, the absorber points DA
and VA are located at infinity such that a constant angle,
θa, is obtained. This method is adaptable to handle
nonplanar sources. If the source, as shown in the figure,
impedes on the string path, then a secondary region
called the involute is formed. The involute ensures that
rays with output angles less than θa are transferred by
the reflector. Such reflectors with nonplanar sources or
absorbers are better denoted as edge-ray reflectors. Note
that the terms “absorber” and “source” are swapped for
collector design.

The CHC is designed with the flow-line method, which

treats light rays as fluid flow. Due to the conservation of
étendue, we find vectors that define the geometrical flux
through the amplitude and their directions define the flow
line. The flow lines are hyperbolae. A reflector can be
placed along one of the rotationally symmetric flow lines,
and due to invariance there is no adverse effect on light
emission from a Lambertian source located at the flow
line origin.
Nonimaging Compound Concentrators 79

Tailored-Edge-Ray Design

The involute sections of edge-ray luminaires for

nonplanar sources or absorbers indicate that the
acceptance angle does not need to be constant over the
extent of the reflector. Designs with a functional
acceptance angle are called tailored-edge-ray
reflectors. Using the figure, the equation that governs
their shape for a point source y Point
design is Source
⎡φ ⎛ s − θ (s) ⎞ ⎤ x
r ( φ ) = r1 exp ⎢ ∫ tan ⎜⎜ ⎟⎟ ds ⎥ , O
⎢⎣ φ1 2 rφ
⎝ ⎠ ⎥⎦
where r is the distance from φ
the point source to the α α Ray
reflector, r1 is the distance for
the polar angle φ1, and θ is the A
desired output intensity from
the reflector. θ is the variable θ(φ)
Reflector
acceptance angle (note that
“acceptance” is a holdover from
solar concentrator design). For
uniformity at the target,
⎡ φ
⎤
θ ( φ ) = arctan ⎢ tan θ1 + ∫ I src (v ) dv ⎥ ,
⎣⎢ φ1 ⎦⎥
where Isrc is the intensity distribution emitted by the
point source.

To allow for finite-extent sources, the first equation can be

modified, and the reader is encouraged to consult the
literature.

Though the formalism presented here might appear

daunting, tailored-edge-ray design is a powerful tool to
design optimal optics around both the emission aspects of
the source and the desired irradiance distribution at the
target. The one caveat is that tolerances are quite
demanding unless one places sufficient leeway into the
source intensity distribution (Isrc).
80 Illumination

Faceted Reflector Design

Essentially there are two design procedures for faceted

reflectors: those based on the tailored-edge-ray
method and those that provide a stippled illumination
pattern. Stippling means that the target irradiance
distribution is created from the overlap of the light from
different segments of the
reflector. This washes out any
Stippled structure that could be imaged
Reflector from the sources, such as a
filament and its supports. Thus,
the designer builds a basic
reflector shape, such as
parabolic, and then replaces the
one smooth reflector with a
series of flat, areal segments.
This type of faceted reflector can
be found in LCD and overhead
projectors.
Tailored-edge-ray reflectors can
also use this effect with some
added benefits:
• Energy conservation re-
strictions mean the 0.4

LED Intensity

reflectors grow large,

0.3 Distribution
Intensity (W/sr)

but faceting allows the

0.2

shape to be “restarted”
0.1
Angle (Degrees)

to minimize the overall -80 -40 0 40 80

volume.
• Facets can individually
address different por-tions of the desired target
distribution.
• Tolerancing is improved since various allowances can
be incorporated as a function of segment position.
These reflectors are typical in the automotive headlight
industry and are increasingly used in other applications.
This example (LED) shows the utility of faceting. The
LEDs’ intensity distribution pattern, along with
uniformity at the target, gives the reflector shape shown
here.
Nonimaging Compound Concentrators 81

Advanced Nonimaging Optic Design

There are a number of advanced nonimaging design

algorithms, such as nonimaging Fresnel lens design,
nonedge-ray design, and simultaneous multiple
surfaces method (SMS). Nonimaging Fresnel lens
design is used in lighthouses, solar concentrators, traffic
lights, and automotive lamps. The in-expensive, small-
volume optics are thin dielectrics, plastic or glass, with
two types of Fresnel elements:
• Catadioptric: uses two
refractions and one TIR to bend
the light in the desired Catadioptric
direction.
• Dioptric: uses two refractions
to bend the light in the desired Dioptric
direction.
Dioptrics are used toward the lens center while
catadioptrics are used once Fresnel losses become large.
The TIR condition reduces the angles of incidence on the
two refractive surfaces.
Nonedge-ray design follows the equations of tailored-edge
ray design but adds two additional factors:
• System performance criteria drive optimization; and
• Multiple extended-size sources are allowed.
This design method trades between system performance
and transfer efficiency from the source to the absorber. It
is used in multiple small-source applications, such as
LED lighting and diode-laser pumping.
SMS provides for multiple ray paths from the source to
points on the to-be-generated optical surfaces of the
device. Refraction, reflection, and TIR are
used in conjunction to generate the multiple
surfaces and provide the optimal output
angular spread from the optic. SMS is part of
LED a family of optics called hybrid optics that
Hybrid use many different optical phenomena for
Optic
their operation. A primary example is the
pseudo-collimating lenses used for high-
brightness LEDs.
82 Illumination

Displays––Overview

A multitude of existing displays incorporate different

illumination strategies to provide a lit screen. Optical
display technologies include backlighting, projection,
and organic LED (OLED).
Backlit displays use large liquid crystal (LC) modules
that are lit from the rear by small sources coupled to a
TIR element that spans the extent of the screen. The TIR
is frustrated by structures placed on a surface of this
element. Sources used in backlit displays include cold-
cathode fluorescent lamps (CCFL) and LEDs, in
which the ejected light proceeds through many additional
layers, including polarizers, the LC, and diffusers.
Additional layers may include a brightness
enhancement film (BEF), which recirculates ejected
light until it is in the desired angular range. The next few
pages describe the components of a backlit display in
more detail.
Projection displays use smaller SLMs in different
spectral ranges to multiplex a full-color image. The
illumination components include a broadband source (e.g.,
a narrow-gap arc lamp or LED), a reflector to capture the
emitted radiation, lenslet arrays (often called fly’s eyes),
and dichroic filters to separate the light into the desired
spectral ranges (typically red, green, and blue). There are
both front-projection displays and rear-projection
displays. Front-projection displays use distinct spectral
channels to illuminate the screen; however, this increases
cost and can reduce tolerances. Rear-projection displays
fold the system in order to maintain a smaller display
depth. Projection displays are discussed in more detail
later.
Unlike backlit and projection displays, OLED displays
deposit pixel emitters onto a substrate. These emitters
provide both the illumination and display information, so
the design demands for the illumination engineer are
negligible. OLED modules can be used in projection
displays.
Displays 83

Backlit Display Components

Standard components of a backlit LCD include:

• Source: Typically CCFL, LED, or electro-
luminescent (EL).
• Injector: A specular or diffuse reflector that captures
and injects the light into the lightguide.
• Lightguide: A dielectric, typically acrylic, that
captures the injected light via TIR. Features are
placed on the backside of the lightguide to break the
TIR condition. The lightguide is also wedged using
decreasing thickness with increasing distance from
the injector.
• Features: Paint patterns or geometric structures to
frustrate the TIR. The density and/or depth of the
features increases with distance from the injector to
provide uniform illumination over the screen. The
geometric structures can be holes (extending into the
lightguide) or bumps (extending out of the lightguide).
• Back reflector: A diffuse or specular reflector placed
below the features to capture and recirculate any light
that is emitted from the lightguide backside.
• Polarizers: Two crossed linear polarizers placed on
the display output side with an LC placed in between.
• Liquid crystal: Sandwiched between the two crossed
linear polarizers to rotate the polarization by 90 deg
for a pixel that has information content. Closely
placed pixels provide for the color content (e.g., three
pixels to provide red, green, and blue).
• Diffusers: Sometimes placed on the output side of the
lightguide to provide better angular uniformity from
the display.
• Brightness enhancement film: A microstructure,
such as a prism, to select a desired angular output
range while the higher angular content is recirculated
to increase display brightness.
Injector and Polarizers, LC, BEF, and
Source Diffusers

Features and
Lightguide Back Reflector
84 Illumination

Backlit Display: Source and Injector

Small sources are preferable for backlights to reduce the

overall display volume. The source can be located to the
side of a lightguide or placed directly behind the
polarizers and LC. The former allows for thin displays at
the expense of lightguide complexity; the latter increases
the depth due to the removal of the lightguide, and
careful design is required to provide uniform luminance.
Three standard sources are used: CCFLs, LEDs, or EL
films. CCFLs are small-diameter lamps that run the
length of one or more sides of the display. For LED
backlights, multiple LEDs are used to provide the
required luminance level. They can be white-light
emitters, a combination of multiple colors (e.g., red, green,
and blue), or a combination of color-emitting LEDs and
white-light emitters. EL films provide single-color
background displays with information shown in black.
Examples include watch faces and automotive
dashboards. They work by passing current through the
EL material, which then emits spatially uniform
Lambertian light. EL backlights have no need for a
lightguide because the EL is mated to the back of the LC-
polarizer module.

An injector is standard for any type of backlighting

scheme. For a backlight whose source is located to the
side of a lightguide, either diffuse or specular reflectors
are placed around the source to better capture the emitted
radiation. Standard shapes for a CCFL include spherical,
parabolic, and elliptical troughs. For LEDs, dielectric
(especially acrylic) couplers akin to the hybrid optics
presented are used. The output aperture of the injector is
mated to the input aperture of the lightguide. For
backlighting without a lightguide, reflectors are often
placed around the sources to assist in directing the light
and to provide uniform luminance from the display. The
simplest case is the lightbox, which is a highly reflecting,
diffuse material placed around the sources over the extent
of the screen backside. Lightboxes are analogous to
integrating spheres.
Displays 85

Backlit Display: Lightguides, Features, Reflectors

Plastics are best suited for backlight

lightguides because they can take
advantage of injection molding. The
thickness at the injector end of the L
lightguide depends on the screen size, with i
larger displays requiring thicker lightguides. g
The lightguide is thinned with increasing h
distance from the injector. This thinning t
assists with ejection of the trapped TIR
light, because as the lightguide cross-
sectional area decreases, the conservation of
étendue demands an increase in the angular
extent.
Structure or features are added to the
backside of the lightguide, i.e., on the
wedged surfaceside because the backside has Bumps
more distance for the light to spread over the
spatial extent of the display. Initially, paint Holes
patterns were used to cause the ejection;
however, the paint must undergo a separate
and costly process, and the paint spots
provide little direct control of the resulting
angular distribution. For this reason,
replicated geometrical structures are added
during injection molding either as holes,
which extend into the lightguide, or bumps,
which extend out of the lightguide.
Geometrical shapes, as shown in the figure,
include hip roofs, spheres, and ellipsoids.
The density and/or depth of these features
increases with distance from the injector.
The design of such feature patterns is from
the diffusion equation, followed by Backside
optimization for improved performance. Reflector

Light leaked through the back feature-side of the

lightguide is caught with a reflector, diffuse, or specular,
which is placed below the lightguide. The backside
reflector provides recirculation and better efficiency.
86 Illumination

Backlit Display: Polarizers, LC, and BEF

There are a number of additional components that

comprise an LCD, including two linear polarizers and
the twisted-nematic LC module. The first polarizer
passes linear polarization at one orientation, while the
second passes linear polarization orthogonal to the first.
The LC is sandwiched in between these two polarizers,
and in a pixel’s transmissive state, it rotates the light that
exits the first polarizer by 90 deg. A pixel in a
nontransmissive state absorbs the incident radiation. The
LC module also has glass substrates on both sides of the
LC, and on each glass substrate there are transmissive
indium-tin oxide (ITO) electrodes. A spectral filter
mask is inserted to provide color output from the display.
Typically, a three-color mask is used, where neighboring
subpixels pass red, green, or blue. The combination of
these pixels forms a display through resolution
considerations of the viewer. There are a number of
color-pixel patterns including:
• Triangular or delta: better for motion pictures;
• Stripes: better for television; and
• Diagonal: better for motion pictures.

R G B R G B R G B B R G B
G B R G B R G B
R G B R G B R G B B R G B
G B R G B R G B
Triangular Stripe Diagonal
A BEF, a replicated structure of microprisms, recirculates
emitted light until it is in the desired angular range. The
BEF is situated just below the polarizers and LC. A dual
brightness enhancement film (DBEF) incorporates
the polarization into the optic such that the first linear
polarizer can be removed.
LC Module

BEF LG
Displays 87

Projection Displays

Projection displays typically use three channels— red,

green, and blue—to develop the object to be displayed by
the projection lens. Each of the channels uses a spatial
light modulator (SLM) to generate this object. One-
channel systems use color filter wheels to temporally
generate the scene. There are essentially three options for
the SLM:
• Transmissive LCs akin to those used in backlighting;
• Digital light processing (DLP) modules, which
incorporate millions of micromirrors over their surface
area; or
• Reflective LCs, such as liquid crystal on silicon,
(LCoS), which integrate the LC with the circuitry.
The SLMs are microdisplays that use magnification from
the projection lens to generate the screen image. An X-
cube combines the three spectral channels, and the
resulting “object” is projected onto the screen.
The illumination components of a projection display
include the source, a reflector, and fly’s eye lenses and/or
straight lightpipes. The source is typically a narrow-gap
arc or even LEDs. The reflector (conic, edge-ray, or
faceted) is specular, and it captures most of the source
emission. The fly’s eye provides better spatial uniformity
over the SLMs by creating several images of the source.
The lightpipe mixes the light to provide better spatial
uniformity. The overall illumination system is typically
arranged in a Köhler scheme to hide the source structure.
Projection
X-cube
lens
Reflector and Fly’s
source eye

Dichroics
88 Illumination

Mapping Flat-Fielding Sources

High-performance camera systems such as airborne and

satellite cameras generally go through a process known as
flat-fielding. The camera is presented with a large-sized
extended light source that has nearly perfectly uniform
radiance. Since the flat-field source is uniform, any pixel-
to-pixel nonuniformities in the camera are inherent to the
camera and can be remedied with image processing.

Generally, flat-fielding sources are realized by internally

illuminated integrating spheres. Spheres with exit ports
of about a 50-cm diameter are common. Ports of over a
meter in diameter are sometimes needed depending on
the aperture of a single large camera or the combined
apertures of an array of smaller cameras. Radiance
uniformities of 98% or 99% or better are the norm.

To verify that the flat-fielding sources have been designed

properly and that there are no deficiencies in their
manufacture, they are mapped for radiance uniformity.
The mapping is done with a radiance meter, which is
often photopically filtered for no other reason than
commercial availability and the desire to band-limit the
silicon detector to a region of good sensitivity.

The radiance meter is operated either in a collimated

mode or is focused on a small spot in the plane of the exit
port. Keeping the viewing direction constant, the meter is
scanned in two directions to create a radiance map of the
source.
Integrating sphere
flat-fielding source Radiance meter
focused on the
plane of the port,
and scanned in
two directions
Characterizing Illumination Systems 89

Goniophotometers

Light sources designed to produce useable irradiance

(automobile headlamps, roadway luminaires, and interior
lighting fixtures) as well as those designed to produce
useable intensity (automobile tail lights, traffic signals,
aircraft and marine running lights) are all characterized
by goniophotometers—devices used to measure the
directional distribution of light from sources.

A goniophotometer consists of a small detector placed at a

distance from the source where intensity is meaningful,
(i.e., the inverse-square law applies). Except for highly
collimated sources such as searchlights, a distance from
the source of five to ten times the largest dimension of the
source is usually sufficient. The lamp or detector (or a
combination of the two) is moved to map the intensity
distribution of the source.

Goniophotometers are classified as type A, B, or C

depending on how they are constructed. This can be
confusing because, in addition to three types of physical
construction, there are three variations of spherical
coordinates for reporting data that are also called types A,
B, and C. These usually, but not always, match the type of
goniophotometer used. Details of the three coordinate
systems are shown on the next page.

Types A and B goniophotometers are similar in that the

luminaire is mounted on a device with horizontal and
vertical axes and a distant fixed detector.

Type C goniophotometers move the detector around the

luminaire on a horizontal axis and rotate the luminaire
on a vertical axis. Sometimes, for large luminaires,
involving large distances, the detector is fixed and a large
high-quality mirror moves on a horizontal axis, directing
the light to the detector.

Type C goniophotometers are necessary for measuring

lamps that are sensitive to the burning position.
90 Illumination

Types A, B, C Goniometer Coordinate Systems

All are spherical coordinate systems.

Type A spherical coordinates:

Polar axis: vertical
Label on vertical angles: Y
Label on horizontal angles: X
Range of Y: –90 (nadir) to +90 (zenith)
Range of X: –180 (left, from luminaire) to +180
Straight ahead: Y = 0, X = 0
Primary uses: optical systems, automotive lighting

Type B spherical coordinates:

Polar axis: horizontal
Label on vertical angles: V
Label on horizontal angles: H
Range of V: –180 to +180
Range of X: –90 (left, from luminaire) to +90
Straight ahead: V = 0, H = 0
Primary uses: floodlights

Type C spherical coordinates:

Polar axis: vertical
Label on vertical angles: V
Label on horizontal angles: L (lateral)
Range of V: 0 (nadir) to 180 (zenith)
Range of L: 0 (along primary axis of luminaire) to 360
Straight down: Y = 0, X = 0
Primary uses: indoor lighting, roadway lighting
Characterizing Illumination Systems 91

“Snapshot” Goniophotometers

Conventional goniophotometers take a long time to

produce an intensity mapping. In addition, they must
have precise motion control to achieve the desired angular
resolution. As such, they are well suited for characterizing
luminaire designs, but not really useful for quality control
or sorting of LEDs, for example. For these applications,
several versions of rapid “snapshot” goniophotometers
have been developed:

• Rapid-scan goniophotometers
• Multiple-detector goniophotometers
• Tapered fiber bundle goniophotometers
• Camera-based goniophotometers

Rapid-scan goniophotometers are small devices used to

characterize LEDs and the output of optical fibers. They
operate on similar principles to the conventional type C
goniophotometers, but motions are much faster, making
measurements in seconds rather than minutes.

Multiple-detector goniophotometers place numerous

discrete detectors in the intensity field of interest and
capture the entire intensity distribution at one time. The
angular resolution is restricted to the spacing of the
detectors.

A tapered fiber optic bundle can be manufactured with

one concave spherical face with all the fibers directed
toward the source. At the other end of the bundle, the
fibers can be aligned with the pixels of a detector array.
The detector array captures the entire intensity
distribution at one time.

Camera-based goniophotometers place a diffuse reflecting

surface (flat or concave) at an appropriate distance from
the source, and view the light reflected from the surface
with an imaging photometer that, together with the
reflecting surface, is calibrated to capture the entire
intensity distribution in one “snapshot.”
92 Illumination

Software Modeling Discussion

Outside the laboratory, software programs are used to

model, optimize, and tolerance optical systems. Two types
of codes exist in the optical design arena: lens design
codes and optics analysis codes. The former are used
primarily to design the lenses used in optical systems.
They include robust analysis tools such as point spread
function graphs, spot diagrams, and modular transfer
function curves; optimization tools to improve upon the
performance of the imaging system; and tolerancing tools
to ensure manufacturability. Increasingly, these lens
design codes include nonsequential ray tracing.
Nonsequential ray tracing is required for a number of
illumination systems, especially those based on
nonimaging optics. In standard lens design, rays follow a
prescribed sequence of optical interfaces. Thus, the traced
rays know the sequence of surface intercepts, which
reduces the computation load since the algorithm does not
need to determine which surface is struck next by a ray.

Optics analysis codes are based around nonsequential ray

tracing such that computation time must be spent to
determine which surfaces are struck by each ray.
Nonsequential ray tracing is inherently slower than
sequential ray tracing. Analysis codes are further broken
down into two geometry types: surface-based geometry
and solid-based geometry. Surface-based codes require
the user to generate each surface, assigning the optical
properties on the two sides of each interface. Solid-based
codes develop enclosed objects that allow the user to
assign volume-based properties such as the type of
material (e.g., BK7) and surface-based properties (e.g., a
silver mirror).

Optical design codes incorporate more computer-aided

design (CAD) into their capabilities. This feature allows
the codes to import mechanical design formats such as
IGES and STEP. Certain industries such as the
automotive and architectural industries have specialized
codes. The list of codes is extensive and always changing.
Architectural Illumination 93

Role of Light in Architecture

The illumination of buildings is a design process aimed at

orchestrating light for the user’s well-being. The layering
and patterning of light is considered successful when
complex physiological and psychological responses are
satisfied. Such responses are centrally conditioned by
vision: the medium through which information and
perceptions about a given space are recorded and
interpreted. Economics and energy efficiency play a
critical role in design decisions, but the satisfaction of
vision requirements is of overriding importance.

The characteristic features of an architectural space only

come to life with light. Hence, no light no architecture. At
the same time, light is not neutral: The way it is arranged
gives a particular appreciation of the space and generates
specific emotive and aesthetic responses.

The electric illumination of an architectural space is

simply the result of transmitted or reflected light
emanating from distant and immediate surrounding
surfaces. Therefore, the lighting designer can influence
the interface between light and matter to meet these
visual requirements and sensations. Hence, only with a
proper understanding of physiological and psychological
factors and a familiarity with available technologies can
lighting decisions be made for proper effect.

Despite some setbacks in the 1970s caused by an

advocacy for windowless buildings to save energy, light
available from the sun and sky has regained the attention
of lighting designers for the many benefits it brings to
users. When available and well controlled, daylight is by
far the preferred source of illumination. Today, the
common design approach combines the contribution of
both electric and natural lights for increased work
productivity, and reduced absenteeism or visual fatigue.
96 Illumination

Lighting Design––Layering of Light

For both energy conservation and visual variety, lighting

design is implemented in layers to properly distribute
light throughout the architectural space.
The horizontal ambient layer is maintained to 1/3 to
2/3 the task illumination level. Lower bound levels (1/3)
for horizontal ambient light may be appropriate for a
museum or boutique store to emphasize a display. Upper
bound levels (2/3) are more relaxing for most casual
activities where a 25- to 35-fc ambient light level is
sufficient and relates well to tasks requiring 50 to 60 fc.
The vertical ambient layer is critical in keeping
vertical tasks glare free, such as washout of video display
terminals (VDTs). In addition, when people look away
from a task, the line of sight is then the vertical average
luminance from the walls and ceiling. Wall washing and
grazing are some of the techniques used to reinforce the
sensation of spaciousness, clarity, and pleasantness.
A task layer supplements the ambient illumination to
fulfill lighting requirements for critical activities. Energy
is saved by (1) locating the source near the task to provide
the light level recommended by the Illuminating
Engineering Society of North America (IESNA), (2)
reducing ambient light levels, and (3) turning off the task
light when not in use. The scene presents varied lighting
instead of the monotonous atmosphere, resulting from the
general illumination approach.
The accent or focal layer gives the space its identity
and mood by highlighting or spotlighting certain
architectural elements and objects, such as paintings,
sculptures, and landscapes. Downlighting, accent
lighting, and backlighting are some techniques used to
produce such effects on various elements in the space.
The ornamental layer introduces elements that add
sparkle to the space with effects similar to those of
Christmas lights. Chandeliers, candles, and sconces can
be considered for this purpose.
94 Illumination

Eye Adaptation and Visual Fields

Eye adaptation to the visual environment is the eye’s

response and sensitivity to the ambient light level as the
person moves from one environment to the next, such as
walking from the bright and sunny outdoors to the dark
indoors. If the difference between the two light levels is
extreme, the person may feel like he or she has moved
into a totally black environment. Slowly, the sensitivity of
the eye attunes itself to the dark environment and details
become increasingly distinguished. It takes 20 to 30
minutes for the eyes to completely adapt to a dark
environment and grasp the details. Conversely, eyes
adapt to a sunny environment in 2 to 3 minutes.

Transient adaptation is the ability of the visual system

to adapt in short intervals to the different luminances
prevailing in a fixed visual field, for instance, when
looking through a bright window and down to a desk. Due
to such variations, the iris constantly adjusts the aperture
to control the light entering the eye. Large variations
between luminances in a scene are considered detrimental
to visual comfort and lead to eye fatigue.

Visual fields refer to the direction of the eyes’ line of

sight. When looking down, the viewer apprehends a
horizontal field, and when looking up, a vertical field.

Vertical Field

Horizontal Field
Visual Fields
Lighting and Visual Performance 95

Apparent Brightness

Vision is stimulated by brightness mapped on the retina

as a byproduct of light reflected from an opaque surface or
transmitted through a transparent medium (the glass
bulb of a lamp, for example). A distinction must be made,
however, between photometric brightness or
luminance and apparent brightness. Luminance or
photometric brightness is calibrated in relation to the
eye’s sensitivity to various wavelengths, while apparent
brightness is perceived in the context of the ambient light
level to which the eye is adapted. Hence, the brightness
of an object relative to the retinal image is by no means a
complete specification of its visual appearance. This may
be understood by considering the blinding brightness of a
car’s headlights during the night that is barely
perceivable during the day, even though a light meter will
register the same photometric brightness.
At an ambient surrounding of 3.4 cd/m2 (1 fL), a measured
luminance of 34 cd/m2 (10 fL) appears to be 340 cd/m2
(100 fL). At a low ambient level, the difference perceived
between two surfaces is also reduced from a difference of
1:10 to 1:4.

eye adaptation levels

1000
Subjective brightness; apparent luminance––fl

3400
500 1:10

200 1:4
1:10 1:2.5
100 340
1.2
50 1:3:5 1.0 fL
1:2
2

20 10 fL
cd/m

100 fL 1:2
10 34
5 1:3

2 1:4 1000 fL

1 3.4
0.5

0.2
0.1
0.1 0.2 0.5 1 2 5 10 20 50 100 200 500 1000 fl
Measured luminance
0.3 3.4 34 340 3400 cd/m

Subjective brightness versus measured luminance. (Reprinted with

permission from Stein and Reynolds, copyright Wiley & Sons,
2005.)
Luminaire for Open-Plan Office 97

Photometric Report and VCP

Manufacturers provide a photometric report that

details the optical performance and characteristic light
distribution patterns of a luminaire. The candela
distribution curve (CDC), presented in either polar
(figure below) or rectilinear plots and in tables, shows the
luminous intensity distribution measured at different
angles, from 0 to 360 deg in increments of 5 deg. Using
the plot below, the luminous intensity can be found for a
specific direction. Rectangular luminaires (2 x 4 or 1 x 4)
require candela distribution curves in at least three
planes: crosswise, longitudinal, and 45 deg. These
luminous intensities can quickly reveal the potential for
glare.

Candela distribution curve.

The visual comfort probability (VCP), a rating system

for evaluating direct discomfort glare, is expressed as the
percent of occupants of a space who will be bothered by
direct glare. Standard data provided for a luminaire
specification include tables of its VCP ratings for various
room geometries, based on IESNA standard conditions.
These include a uniformly distributed illumination level
of 1000 lux (~100 fc), luminaire height, observer position,
and room surface reflectances (ceiling, 80%; walls, 50%;
and floor, 20%). In general, a minimum VCP of 70 is the
established limit for the viable use of a luminaire.
98 Lighting Design

The Layered Approach

Spacing criteria (SC) to achieve uniform ambient

illumination is the ratio of the spacing (S) distance
between the respective axis of parallel luminaires and
their mounting height (MH), i.e., SC = S/MH. For a
rectangular luminaire, SC is given along both axes,
lengthwise and crosswise. The distance between walls and
adjacent light fixtures is set at no more than one-half S.

The coefficient of utilization (CU) specifies the

proportion of lumens that reach the workplane from the
fixture for given room geometries and surface
reflectances. The CU gives some indication of the
luminaire’s efficiency.

When located in the field of view, bright light sources can

cause discomfort and disability glare. However, the
severity of glare depends on the angle at which the
luminaire is seen. The IESNA Handbook provides as
maxima the following luminances from a direct
luminaire according to the view or cut-off angle:

Maximum luminance according to view angle.

Daylight Compensation 99

Daylight Factor

Windows and skylights admit daylight as free

illumination, and the viewers have visual contact with the
outside. However, only a portion of outdoor light is
received inside a building. The sky’s condition, along with
the size, placement, and orientation of the window(s)
opening, the glazing type and transmittance, and shading
as well as room proportions, affect the quantity and
quality of received light.
The daylight factor (DF) accounts for these parameters
and is used to determine the percentage of outside light
that can be received inside a room with a specific
configuration. DF is the ratio of interior illuminance at a
given point on a given plane (usually the work plane) to
the exterior illuminance under minimal light conditions of
an overcast sky, e.g., the CIE overcast sky distribution.
For parallelepipedic-room buildings with
windows on one side only, a minimum DF
of 2% is generally recommended
throughout the work plane.
The DF general equation is
DF = Ei/Eo,
where Ei and Eo are the indoor and outdoor horizontal
illuminances, respectively.
Daylight should be allowed to
15’ 15’
reach through side windows for
the tasks performed at the rear of
the space. Shelves in side
Core
windows can help reflect light
deep into interior spaces and
shade the vision portion of the
window.
Office tasks can be lit at a depth of up to 15 feet from the
window. The next 15 feet may need to be supplemented by
electric light. Beyond 30 feet, little or no daylight is
available unless the space has windows on opposite sides.
Evenly distributed roof skylights provide uniform light
but for one story only, assuming proper glazing
transmittance and sunlight control.
100 Illumination

Daylight Strategies

IESNA RP-5-99 Recommended Practice of daylighting

emphasizes the following illumination practices:
1. Block direct sunlight in the vicinity of tasks. Blocking
direct sunlight at the window (e.g., with louvers) is the
first step toward glare control by allowing only light
from the sky and that reflected from the ground to pass
through the window glass.
2. Design windows to minimize direct glare. East, west, or
south facing windows can have too much glare if
excessive sunlight strikes the glass. With 100,000 lux
of light available on the outside on a sunny day, a glass
pane, with typically only 2% transmittance, can reach a
luminance up to 2000 cd/m2, which exceeds the
tolerated average of 850 cd/m2. Internal blinds
(structurally stable external shading devices with
adjustable louvers) can block direct sunlight and
reduce the luminance of the window or skylight. And
glazing that has a visible light transmittance of 25%
can be an acceptable trade-off of daylight availability,
the view to the outside, and minimize glare.
3. Zone electric lighting for daylight responsive control.
The electric light distribution system should be zoned
according to daylight availability inside an open-plan
office. Daylight zoning depends on the room
configuration, sky condition, and solar exposure. Large
open-plan offices are often subdivided into the
perimeter zones, the intermediate zone, and the core
zones based on daylight availability as indicated in
item 1.
4. Provide responsive lighting controls. Controls are at
the heart of efficient electric light operation and
daylight harvesting, specifically to accommodate the
time-dependent electric light demand. The variables
governing control strategies include the space layout,
configuration, orientation, the occupancy patterns,
lighting usage, and daylight availability. Controls
include tuning to reduce electric power while still
meeting each user’s needs, and adaptive compensation
to lower the light levels at night.
Exterior Lighting 101

Nighttime Visibility Criteria

The eye is capable of adapting to a wide range of light

levels but not at the same time. To function well, it must
be adapted to the prevailing light conditions. As
previously indicated for daytime conditions, our eyes use
photopic vision, which utilizes the eye’s cones and the
center of the visual field. The eye works differently when
it is adapted to low light levels. Under very dark, moonlit
conditions, our eyes use scotopic vision, which primarily
utilizes the eye’s rods, resulting in greater acuity in the
peripheral visual field.
For nighttime visibility in most urban and suburban
environments, our eyes use mesopic vision, which is a
combination of both photopic and scotopic. In nighttime
environments, the goal of the lighting design is to keep
the eye adapted to mesopic or scotopic vision, and not to
introduce high light levels that will create an imbalance
in the visual field and cause the eye to try to use photopic
vision. Recent research indicates that light sources rich in
blue and green (metal halide or fluorescent) improve
peripheral mesopic vision, clarity, and depth of field
better than sources rich in red and yellow, such as
incandescent and high-pressure sodium.
Extreme glare leads to loss of visibility. Glare is caused
by a high luminance ratio between the glare source and
the prevailing light conditions to which the eye is
adapted. In other words, insufficiently shielded light
sources generate direct glare. The following measures
reduce the luminance ratio and control nighttime glare:
• Uniform light distribution in a visual scene and
brightness ratios kept to 1:5 between average and
maximum luminance;
• Reduction of light levels and source brightness
using fixtures with low wattage;
• Shielding the light source and locating fixtures
to avoid glare. Fixtures near the property line should
have “house-side shielding” to prevent glare to
residential neighbors.
102 Illumination

Recommended Illuminance for Façades

Façade illumination aims to reproduce at night a

building’s aesthetic and formal characteristics that are
perceived during the day for the purpose of attracting
attention and creating a good impression. To this effect,
floodlighting is one technique employed, which treats a
building as a giant piece of sculpture for visual display.
Luminaires are typically mounted in close proximity to
buildings and are aimed to illuminate the structure.
Lighting the building from the top down reduces stray
uplight, and precisely aimed fixtures minimize trespass
light.

Effective illumination of facades is a complex and

subjective task. Results depend heavily upon surrounding
light levels, the surface finish of the intended target, the
spectral color distribution of the lamp source, mounting
location allowances, and viewers’ perceptions.

The following table lists the IESNA’s recommended

illuminance levels for the floodlighting of buildings and
monuments.

Target
Area Surface Average Target
Description Finish Illuminance
Bright Light 5 (50)
Bright Medium light 7(70)
Bright Dark 10(100)
Dark Light 2(20)
Dark Medium light 3(30)
Dark Medium dark 4(40)
Dark Dark 5(50)
th
Adapted from IESNA Hand-book, 9 Edition, copyright IESNA,
2000.
Exterior Lighting 103

Façade Floodlighting for Uniform Illumination

Floodlighting fixtures
can be mounted at
ground level, or on
stands and poles. They
can also be attached to
the building itself or to
adjacent structures.
The key lights are set
up for a modeling effect
but should be combined
with other color sources
to soften the strong
effects of shadows.
Floodlight categories
Adapted from IESNA Hand- are narrow beam
book, 9
th
Edition, copyright (types 1, 2, 3), medium
IESNA, 2000. beam (types 4, 5) and
wide beam (types 6, 7)
(IESNA RP-33). The further away the luminaire is from
the facade, the narrower the light beam must be. Aiming
and positioning ground-mounted floodlights for uniform
illumination depend first on the available setback in
relation to the building height. If the height is 2 times the
setback dimension, the center of a “wide beam” floodlight
aimed at 2/3 the height of the building is recommended. If
a building is 30 feet high, the recommended aiming point
is 20 feet high. Floodlight spacing along the facade
should not exceed 2 times the setback distance. If the
setback is 22.5 feet, the floodlights should be placed no
more than 45 feet apart, with the first floodlight at ½ to 1
of the setback dimension. As the building height increases
to 4 times the setback, a medium-beam floodlight with the
same aiming elevation is recommended. Buildings with
up to 6 times the setback require more narrow-beam
floodlights. Thus, one location on the ground may hold
multiple floodlights, each aimed at different building
elevations. The illumination from the ground of façades
with more than 6 times the setback is not recommended
due to the difficulty of achieving uniformity.
104 Illumination

Illumination of Outdoor Areas

Lighting building entries at night provides (1) vertical

illumination to comfortably light people’s faces, and (2)
horizontal illumination to light the pathway and any
changes in the light level. Such a pool of light comes from
a mounting position high on the building, on a pedestrian-
scaled post lantern, or on the underside of a canopy.
Emergency egress doors are provided with lighting on
the outside of the door threshold and extended for a
distance at least equal to the width of the door opening.
Softscape lighting is for private yards, patios, parks,
gardens, boulevards, entry markers, and other natural
features such as water. They are softly illuminated and
emit a minimum of glare, contrast, or spill light to the
neighbors. Some techniques used to light trees to achieve
the desired effect are frontlighting to highlight details,
texture, and color; backlighting to show form and
separate the plant from the background; sidelighting to
emphasize plant texture and create shadows; uplighting
to make branches glow; and downlighting for accent
details, colors, and texture. The illumination of tree
trunks along with canopies helps anchor them to the
landscape.
Hardscape lighting is for outdoor sculptures, fountains,
or vertical displays. A 3D sculpture is illuminated from
two directions to provide highlights and soften shadows.
The key light is focused on the mass of the sculpture with
light added to relieve shadows.
Stairs and ramps are hazardous in low light, so contrast
is essential for their safe use. Illuminated handrails, step
lights, or small fixtures in the balustrade provide light
differentiation between the step risers and threads. Other
techniques to complement light effects are coloring of the
step nosing and color differentiation between threads and
risers.
Walkways, sidewalks, and bikeways are illuminated
at levels recommended by the IESNA with lights placed to
provide visual information.
Exterior Lighting 105

Special Considerations for Outdoor Fixtures

Controls using astronomical time switches and/or

photosensors are deployed to ensure that exterior lighting
is not operated when sufficient daylight is available or
during nighttime except those fixtures for security.
Special considerations must be given to proper
installation of luminaires exposed to the outdoor
environment. If installed on ground they must have the
“wet location rated” label and if placed under canopies
but still exposed to the elements, they must be “damp
location rated.” In addition, durable with vandal-
resistant components and regularly maintained
luminaires to minimize dirt accumulation or to prevent
obstruction by grass, leaves, mud, and other debris,
ensure steady operation of exterior lighting. Separate
security fixtures are exclusively used to provide low light
levels for security cameras (.01 footcandle). The table
below summarizes the illuminance and luminance ratios
for various outdoor areas.
Horiz. Avg. Vert. Avg.
Illuminance Illuminance
Outdoor space type (fc) (fc)
Building entrance
(Active/Inactive) 5.0 / 3.0 3.0 / 3.0
Emergency lighting:
Egress Path 1.0
Roadside sidewalks &
Type A bikeways:
commercial,
intermediate, residential
areas 1.0; 0.6; 0. 2.2; 1.1; 0.5
Walkways distant from
roadside & Type B
bikeways 0.5 0.5
Loading dock 10.0 The plane of the task may be
horizontal, inclined, or vertical.
Storage yards,active-
inactive 10.0 / 1.0 3.0 / 0.3

(Adapted from IESNA Lighting Handbook, Ninth Edition.)

106 Illumination

Outdoor Luminaire––Transverse Light

Distribution

Luminaires’ beam pattern distributions are classified

by IESNA according to transverse and lateral projections.
(see the next two figures). Five types, shown below, are
illustrated according to the maximum candlepower and
the trace of the half-maximum value. A luminaire’s
transverse reach is expressed in MH units: type I (1
MH), type II (1.75 MH), type III (2.75 MH), type IV (6
MH), and type V (symmetric distribution in four
quadrants). Type V is usually best at the center of
parking lots. Type IV or forward-throw distribution is
best for wide, multilane roads and parking lot perimeters.
Type I has a long and narrow distribution that can be
applied to narrow roadways, walkways, or bike paths. It
also can be located at or near the center of a pathway,
approximately two MHs in width, or used as overhead
lighting in areas such as parking lots, plazas, courtyards,
and along walkways. Types II and III distribute light to
one side of the light source. These luminaires should
generally be used for street lighting to direct light to the
street side of the lamp but not shining into the building
side.

Type I Type II

Type III Type IV

Type V
th
Adapted from IESNA Handbook, 9 Edition, copyright IESNA,
2000.
Parking 107

Outdoor Luminaire––Lateral Light Distribution

Fixtures for roadway and parking applications are further

classified as short, medium, or long lateral
distribution. This classification relates the types of
fixtures, the spacing between them according to the point
of maximum candelas, and the MH. For a short-range
lateral throw, the maximum luminaire spacing is
generally less than 4.5 times the MH. A medium throw
allows a maximum spacing of generally less than 7.5
times the MH, and a long throw is generally less than 12
times the MH (see figure below).

Trace of 50%
House Side
Maximum Candela
Point of Maximum
Reference Line Candela

Type II
Street Side
1.75 MH

2.75 MH Type III

Type IV

Medium
Long Range Range Short Range
6.00 MH 3.75 MH 2.25 MH

Type III distribution of a luminaire. Half-maximum candelas

trace falls within 2.75MH. (Adapted from IESNA Handbook,
9th Edition, copyright IESNA, 2000.)
108 Illumination

Criteria For Roadway Lighting

Three principal criteria are used to design major roadway

lighting systems: luminance, illuminance, and the
newer concept of small target visibility (STV). The
illuminance inverse square law calculations are well
known. It was found, however, that illuminance levels do
not correlate well with visibility or driver performance.
IESNA Standard RP-8-00 addresses one shortcoming of
the illuminance method by adding a maximum veiling
luminance ratio (VLR) that is specifically intended to
limit glare from a luminaire. The luminance
determination is necessary to calculate the VLR.
Luminance describes the reflected light from the
pavement as seen when driving, so evaluating the quality
of a lighting system by how it looks at night is actually
the same as evaluating its luminance.

In reference to the figure on the next page, luminance at

point P is determined as the sum of contributions from all
n luminaires:

LP = Σr(βi,γ i)I(ϕi,γ i)/10,000 h2,

where r is the reflectance coefficient at angles β and γ.

The veiling luminance, Lv, and the VLR are also

necessary to limit the glare effect. The Lv can be
determined as follows:

Lv = Σ 10 Ev / (θ2 + 1.5 θ),

where: Lvis the veiling luminance at the observer’s

location, in cd/m2; Evis the vertical illuminance on the
plane of the observer’s eye; and θ is the angle between the
line of sight and the luminaires in degrees.

Values recommended for luminance and VLR are found in

the 9th Edition of the IESNA Handbook.
Roadway Lighting 109

Small Target Visibility

h
γ γ
P

φ β

α = 1°

Single fixture for luminance determination. (Adapted from IESNA

th
Hand-book, 9 Edition, copyright IESNA, 2000.)

The STV method was developed to account for the

contrast that must be present to allow drivers moving at
high speed to quickly detect hazards and react to them.
Indeed, a roadway lighting system may provide a high
and uniform road surface luminance, yet the visibility
threshold may be low due to the absence of contrast.
Three luminance components influence the visibility of a
target: the target luminance itself, the luminance of the
background, and the veiling luminance or glare. Given
these three luminances, all of which can be calculated, the
visibility of each target in the array can be determined in
terms of the visibility level (VL). VL is the ratio of
target contrast to the contrast of a similar target at
threshold, a measure of visibility that has been widely
used.

STV predicts the visibility of a standard object (18 × 18

cm) located on the roadway at a specific distance from the
driver, and accounts for the contrast between the
standard target and its background by considering the
driver’s age, viewing time, pavement reflectance, and
glare from the luminaire. The larger the STV number, the
more visible the object. For more details, go to IESNA RP-
8-00.
110 Roadway Lighting

Recommended Roadway Luminaires

After the desired minimum illuminance and the pole

height are initially set based on light spill, road cross-
section, pavement type, and roadway category (IESNA
Handbook), a few luminaires and sources can be tested to
achieve a design that meets the recommended light level,
uniformity, and acceptable glare. Common layouts include
luminaires on one side of the road, luminaires on both
sides, or luminaires in a center median. One-side and
median configurations often have the additional
advantage of requiring less wire and conduit, resulting in
lower construction costs.
Starting with any of the five previously described types of
luminaires is a convenient way to facilitate selection
according to roadway width and light spill control. Full
cutoff luminaires should be specified wherever possible
to prevent light pollution. Typically, Types I, II, and III
are appropriate for narrow roadways, while type IV is
proper for multilane roads. Lateral-medium-throw
luminaires are preferred over short-throw types because
fewer poles are required and over long-throw types
because then semicutoffs are not needed. Combining
transversal and longitudinal distributions helps the
designer select luminaires for even light distribution
based on roadway widths and pole spacing.
The architecture of such luminaires comes in many
shapes. Two are shown below. The most prevalent is the
“cobra head” luminaire, typically mounted on a 6-ft arm.
This fixture with a flat lens is recommended to minimize
disability glare and light trespass. Reflectors may be
formed or faceted aluminum. The rectilinear “shoebox”
luminaire, designed as a full cutoff, is also popular. The
reflector in this design is usually larger than that in a
cobra head to achieve better optical control.

(Reprinted with permission of Acuity Brands Lighting.)

Roadway Lighting 111

Recommended Lamps for Roadway Luminaires

Metal halide, high-pressure sodium, and induction

lamps are the most common lamps, ranging from 100 to
400 W depending on the MH and roadway cross-section.
Metal halide is used where color rendition is a concern or
white light is desired, but this lamp also emits light in the
blue-green portion of the spectrum, which enhances the
driver’s peripheral vision. A light-loss factor (LLF)
accounts for the depreciation in lamp output and
luminaire performance over time. Typical LLFs are 0.60
to 0.70 for high-pressure sodium lamps and 0.45 to 0.55
for metal halide lamps. Note that the LLF varies
depending on environmental conditions and maintenance
procedures.

Left: metal halide lamp. Above: 400W high pressure sodium

BT-37. (Images reprinted with permission from Sylvania.)

The isofootcandle procedure, not covered here, can be

used to determine luminaire spacing based on needed
illuminance and uniformity. With the help of software
programs, many iterations can be performed quickly to
compare various lighting systems and determine the
safest and most efficient solution.
112 Illumination

Equation Summary

Basic Quantities in Illumination

photopic luminous flux = 683∫ Φ λ ( λ ) ⋅ V ( λ ) ⋅ dλ

Color

Light source color:

X
X = ∫ Sλ ( λ ) ⋅ x ( λ ) ⋅ dλ x=
X +Y + Z
Y
Y = ∫ Sλ ( λ ) ⋅ y ( λ ) ⋅ dλ y=
X +Y + Z

Z = ∫ Sλ ( λ ) ⋅ z ( λ ) ⋅ d λ

u = 4 X ( X + 15Y + 3Z ) = 4 x ( −2x + 12 y + 3)
v = 6Y ( X + 15Y + 3Z ) = 6 y ( −2x + 12 y + 3)
u′ = 4 X ( X + 15Y + 3Z ) = 4 x ( −2x + 12 y + 3)
v′ = 9Y ( X + 15Y + 3Z ) = 9 y ( −2x + 12 y + 3)

Color temperature and CCT:

6
( MK )−1 = 10
CCT

Surface color:

X = k ∫ Sλ ( λ ) ⋅ R( λ ) ⋅ x ( λ ) ⋅ dλ

Y = k ∫ Sλ ( λ ) ⋅ R( λ ) ⋅ y( λ ) ⋅ dλ

Z = k ∫ Sλ ( λ ) ⋅ R( λ ) ⋅ z ( λ ) ⋅ dλ

100
k=
∫ S ( λ ) ⋅ y ( λ ) ⋅ dλ
λ
Appendix 113

Equation Summary (cont.)

Color of fluorescent surfaces:

780 780
X =k ∑ ∑ S ( µ ) D ( µ, λ ) x ( λ )
λ = 380 µ= 300
µ

780 780
Y =k ∑ ∑ S ( µ ) D( µ , λ ) y ( λ )
λ= 380 µ= 300
µ

780 780
Z =k ∑ ∑ S ( µ ) D ( µ, λ ) z ( λ )
λ = 380 µ= 300
µ

780
k = 100 ∑ S ( λ ) y( λ )
λ = 380
λ

Illumination Properties of Materials

BRDF and reflectance factor:

BRDF=L/E R=BRDF·π

Harvey/ABg scatter model:

g A
B = βrolloff BSDF = g
B + β − β0

Retroreflectors:

I RI RA L
RI = RA = RL = =
E⊥ A cos v E⊥

Illumination Transfer

Known intensity:
I cos ξ I
E= Φi = I ω L=
d2 Ar cos θ
114 Illumination

Equation Summary (cont.)

Known flux:

Φi Φi
I = Φi ω E = Φi Ai L= =
ω Ar cos θ Ω Ar

Known radiance:

Φ i L Ar Ω L Ar cos θ cos ξ
E0 = = = = L Ωi
Ai Ai d2

Known irradiance:
E ⋅ d2 R⋅E
I = Φ = E ⋅ Ai L=
cos ξ π

Relationship of configuration factor to projected

solid angle and of form factor to average projected
solid angle:
C=Ω π Fa to b = Ω a to b π

Projected solid angle of circular area from on-axis

point:
π r2
Ω= 2 = π sin 2 θ
r + d2

Projected solid angle of circular area from off-axis

point:
⎛ ⎞
π⎜ 1 + tan 2 δ − tan 2 θ ⎟
Ω = ⎜1 − 1 ⎟
2⎜
⎝ ( )( )
⎡tan 4 δ + 2 tan 2 δ 1 − tan 2 θ + sec 4 θ ⎤ 2 ⎟
⎣ ⎦ ⎠
⎛ ⎞
⎜ ⎟
( ) ( )
2 2

π⎜ 1+ d − r ⎟
Ω = ⎜1 − x x ⎟

{ }
1
2⎜ ⎡ 2
2⎤ 2 ⎟
( ) ( ) ( )
2 2
⎜
⎜ ⎢ 1+ dx + rx −4 r
x ⎥⎥ ⎟
⎟
⎝ ⎢
⎣ ⎦ ⎠
Appendix 115

Equation Summary (cont.)

Average projected solid angle from one circular

area to another, with both areas parallel and
centered on the same axis:

⎡ 1
⎤
⎢ 2 ⎧⎛ 2
⎞
2
⎫2 ⎥
⎛ ri ⎞ ⎪ ⎛ ri ⎞
⎢ 1 + ⎜d⎟ ⎜ 1 + ⎜d⎟ ⎟ 2⎪ ⎥

Ωr to i
π
= ⎢1 + ⎝ ⎠ − ⎪⎜1 + ⎝ ⎠ ⎟ − 4 ⎛ ri ⎞ ⎪ ⎥
⎨⎜ ⎟ ⎜ ⎟ ⎬ ⎥
2⎢ r
⎛ r⎞
2
⎪⎜ r
⎛ r⎞
2
⎝ rr ⎠ ⎪ ⎥
⎢ ⎜d⎟ ⎜d⎟ ⎟
⎢ ⎝ ⎠ ⎪⎜ ⎝ ⎠
⎟ ⎪ ⎥
⎢⎣ ⎩⎝ ⎠ ⎭ ⎥
⎦

Cosine fourth and increase factor:

Ei = π Lr sin 2 θ ⋅ cos4 δ ⋅ F ′

ω, Ω, NA, and f/# for a circular cone:

ω = 2π(1 − cos θ) Ω = π sin 2 θ

NA = n ⋅ sin θ f /# = 1 2sin θ

Illumination in Imaging Systems

Object and image radiance:

Li L
2
= τ ⋅ 2o
ni no

Image irradiance off axis:

Ei = π τLo sin 2 θ ⋅ cos4 δ ⋅ F ′

Image irradiance on axis:

π τ L0 π L0
Ei 0 = π τLo sin 2 θ = 2
=
4( f #) 4(T #)2
= π τ L0 NA 2 = τ L0 Ω

Image flux:
Φi = τ Lo ai Ωi = τ Lo a p Ω p
116 Illumination

Equation Summary (cont.)

Illumination in Nonimaging Systems

Generalized étendue:
E = n2 ∫∫
aperture
cos θdAs dω

Φ= ∫∫ L ( r, aˆ ) cos θdAs dω,

aperture

Φ = Ls ∫∫
aperture
cos θdAs dω

Ls E
=
n2
Concentration:
C = A A′

2
a n′ sin θ′ A ⎛ n′ sin θ′ ⎞
C2 D = = C3 D = =
a ′ n sin θ A ′ ⎜⎝ n sin θ ⎟⎠

2
n′ ⎛ n′ ⎞
C2D,opt = C3D,opt = ⎜ ⎟
n sin θa ⎝ n sin θa ⎠

Skew invariant:
dE ( s )
fskew ( s ) = ,
ds
where s = rmin kt

Fibers, Lightpipes, and Lightguides

Maximum acceptance angle:

1
sin θmax = n12 − n22
n0
Numerical aperture:
NA = n0 sin θmax = n12 − n22
Appendix 117

Equation Summary (cont.)

Étendue:
π2 2
E′tendue = d NA 2
4

Tapered single fiber:

ai·NAi2 ≈ ao·NAo2
Tapered bundle:

NAo = NAi = NAfiber

ao·pfo = ai·pfi

Uniform Illumination

Bent lightpipes:
R = r2 r1 = 1 + t r1 ≤ n

Integrating sphere radiance:

Φ 1
L= ⋅M M =
π ⋅ As 1−ρ

⎡ φ
⎤
θ ( φ ) = arctan ⎢ tan θ1 + ∫ I src (v ) dv ⎥
⎣⎢ φ1 ⎦⎥
⎡φ ⎛ s − θ (s) ⎞ ⎤
r ( φ ) = r1 exp ⎢ ∫ tan ⎜⎜ ⎟⎟ ds ⎥
⎢⎣ φ1 ⎝ 2 ⎠ ⎥⎦

Nonimaging Compound Concentrators

Preservation of étendue:

ai Ωi = ao Ω o

Maximum concentration ratio:

ai 1
(max) =
ao sin 2 θi
118 Illumination

Equation Summary (cont.)

Diameter of CPC used as a collimator:

do 1
=
di sin θmax

Length of CPC:
do + di
Length =
2 tan θmax

Equation for surface of CPC:

(r cos θmax + z sin θmax )2
+r ds (1 + sin θmax )2
−z ds cos θmax (2 + sin θmax )
ds2
− (1 + sin θmax )(3 + sin θmax )
4
= 0,
where:
r is the radius of the cone, perpendicular to the axis,
z is the axial position measured from the small end,
ds is the diameter of the small end, and
θmax is the maximum field angle at the large end.

Tailored edge ray design:

⎡ φ
⎤
θ ( φ ) = arctan ⎢ tan θ1 + ∫ I src (v ) dv ⎥
⎣⎢ φ1 ⎦⎥
⎧⎪ φ ⎡ s − θ ( s ) ⎤ ⎫⎪
r ( φ ) = r1 exp ⎨ ∫ tan ⎢ ⎥ ds ⎬
⎩⎪φ1 ⎣ 2 ⎦ ⎭⎪
Appendix 119

CIE Illuminants A and D65

λ CIE Ill. A CIE Ill. D65

(nm) Tungsten at 2856 K Sunlight
300 0.930483 0.03410
305 1.128210 1.66430
310 1.357690 3.29450
315 1.622190 11.76520
320 1.925080 20.23600
325 2.269800 28.64470
330 2.659810 37.05350
335 3.098610 38.50110
340 3.589680 39.94880
345 4.136480 42.43020
350 4.742380 44.91170
355 5.410700 45.77500
360 6.144620 46.63830
365 6.947200 49.36370
370 7.821350 52.08910
375 8.769800 51.03230
380 9.795100 49.97550
385 10.899600 52.31180
390 12.085300 54.64820
395 13.354300 68.70150
400 14.708000 82.75490
405 16.148000 87.12040
410 17.675300 91.48600
415 19.290700 92.45890
420 20.995000 93.43180
425 22.788300 90.05700
430 24.670900 86.68230
435 26.642500 95.77360
440 28.702700 104.86500
445 30.850800 110.93600
450 33.085900 117.00800
455 35.406800 117.41000
460 37.812100 117.81200
465 40.300200 116.33600
470 42.869300 114.86100
120 Illumination

CIE Illuminants A and D65 (cont.)

λ (nm) CIE Ill. A CIE Ill. D65

475 45.517400 115.39200
480 48.242300 115.92300
485 51.041800 112.36700
490 53.913200 108.81100
495 56.853900 109.08200
500 59.861100 109.35400
505 62.932000 108.57800
510 66.063500 107.80200
515 69.252500 106.29600
520 72.495900 104.79000
525 75.790300 106.23900
530 79.132600 107.68900
535 82.519300 106.04700
540 85.947000 104.40500
545 89.412400 104.22500
550 92.912000 104.04600
555 96.442300 102.02300
560 100.000000 100.00000
565 103.582000 98.16710
570 107.184000 96.33420
575 110.803000 96.06110
580 114.436000 95.78800
585 118.080000 92.23680
590 121.731000 88.68560
595 125.386000 89.34590
600 129.043000 90.00620
605 132.697000 89.80260
610 136.346000 89.59910
615 139.988000 88.64890
620 143.618000 87.69870
625 147.235000 85.49360
630 150.836000 83.28860
635 154.418000 83.49390
640 157.979000 83.69920
645 161.516000 81.86300
650 165.028000 80.02680
Appendix 121

CIE Illuminants A and D65 (cont.)

λ (nm) CIE Ill. A CIE Ill. D65

655 168.510000 80.12070
660 171.963000 80.21460
665 175.383000 81.24620
670 178.769000 82.27780
675 182.118000 80.28100
680 185.429000 78.28420
685 188.701000 74.00270
690 191.931000 69.72130
695 195.118000 70.66520
700 198.261000 71.60910
705 201.359000 72.97900
710 204.409000 74.34900
715 207.411000 67.97650
720 210.365000 61.60400
725 213.268000 65.74480
730 216.120000 69.88560
735 218.920000 72.48630
740 221.667000 75.08700
745 224.361000 69.33980
750 227.000000 63.59270
755 229.585000 55.00540
760 232.115000 46.41820
765 234.589000 56.61180
770 237.008000 66.80540
775 239.370000 65.09410
780 241.675000 63.38280
785 243.924000 63.84340
790 246.116000 64.30400
795 248.251000 61.87790
800 250.329000 59.45190
805 252.350000 55.70540
810 254.314000 51.95900
815 256.221000 54.69980
820 258.071000 57.44060
825 259.865000 58.87650
830 261.602000 60.31250
122 Illumination

x , y , z , V (λ), and V ′(λ)

Note: The photopic efficiency function, V(λ), is identical to

the ȳ standard observer function.

λ
(nm) x̄ ȳ , V z̄ V'
360 0.000130 0.000004 0.000606 0.000000
365 0.000232 0.000007 0.001086 0.000000
370 0.000415 0.000012 0.001946 0.000000
375 0.000742 0.000022 0.003486 0.000000
380 0.001368 0.000039 0.006450 0.000000
385 0.002236 0.000064 0.010550 0.001108
390 0.004243 0.000120 0.020050 0.002209
395 0.007650 0.000217 0.036210 0.004530
400 0.014310 0.000396 0.067850 0.009290
405 0.023190 0.000640 0.110200 0.018520
410 0.043510 0.001210 0.207400 0.034840
415 0.077630 0.002180 0.371300 0.060400
420 0.134380 0.004000 0.645600 0.096600
425 0.214770 0.007300 1.039050 0.143600
430 0.283900 0.011600 1.385600 0.199800
435 0.328500 0.016840 1.622960 0.262500
440 0.348280 0.023000 1.747060 0.328100
445 0.348060 0.029800 1.782600 0.393100
450 0.336200 0.038000 1.772110 0.455000
455 0.318700 0.048000 1.744100 0.513000
460 0.290800 0.060000 1.669200 0.567000
465 0.251100 0.073900 1.528100 0.620000
470 0.195360 0.090980 1.287640 0.676000
475 0.142100 0.112600 1.041900 0.734000
480 0.095640 0.139020 0.812950 0.793000
485 0.057950 0.169300 0.616200 0.851000
490 0.032010 0.208020 0.465180 0.904000
495 0.014700 0.258600 0.353300 0.949000
500 0.004900 0.323000 0.272000 0.982000
505 0.002400 0.407300 0.212300 0.998000
510 0.009300 0.503000 0.158200 0.997000
Appendix 123

x , y , z , V (λ), and V ′(λ) (cont.)

λ
(nm) x̄ ȳ , V z̄ V'
515 0.029100 0.608200 0.111700 0.975000
520 0.063270 0.710000 0.078250 0.935000
525 0.109600 0.793200 0.057250 0.880000
530 0.165500 0.862000 0.042160 0.811000
535 0.225750 0.914850 0.029840 0.733000
540 0.290400 0.954000 0.020300 0.650000
545 0.359700 0.980300 0.013400 0.564000
550 0.433450 0.994950 0.008750 0.481000
555 0.512050 1.000000 0.005750 0.402000
560 0.594500 0.995000 0.003900 0.328800
565 0.678400 0.978600 0.002750 0.263900
570 0.762100 0.952000 0.002100 0.207600
575 0.842500 0.915400 0.001800 0.160200
580 0.916300 0.870000 0.001650 0.121200
585 0.978600 0.816300 0.001400 0.089900
590 1.026300 0.757000 0.001100 0.065500
595 1.056700 0.694900 0.001000 0.046900
600 1.062200 0.631000 0.000800 0.033150
605 1.045600 0.566800 0.000600 0.023120
610 1.002600 0.503000 0.000340 0.015930
615 0.938400 0.441200 0.000240 0.010880
620 0.854450 0.381000 0.000190 0.007370
625 0.751400 0.321000 0.000100 0.004970
630 0.642400 0.265000 0.000050 0.003335
635 0.541900 0.217000 0.000030 0.002235
640 0.447900 0.175000 0.000020 0.001497
645 0.360800 0.138200 0.000010 0.001005
650 0.283500 0.107000 0.000000 0.000677
655 0.218700 0.081600 0.000000 0.000459
660 0.164900 0.061000 0.000000 0.000313
665 0.121200 0.044580 0.000000 0.000215
670 0.087400 0.032000 0.000000 0.000148
675 0.063600 0.023200 0.000000 0.000103
680 0.046770 0.017000 0.000000 0.000072
124 Illumination

x , y , z , V (λ), and V ′(λ) (cont.)

λ
(nm) x̄ ȳ , V z̄ V'
685 0.032900 0.011920 0.000000 0.000050
690 0.022700 0.008210 0.000000 0.000035
695 0.015840 0.005723 0.000000 0.000025
700 0.011359 0.004102 0.000000 0.000018
705 0.008111 0.002929 0.000000 0.000013
710 0.005790 0.002091 0.000000 0.000009
715 0.004106 0.001484 0.000000 0.000007
720 0.002899 0.001047 0.000000 0.000005
725 0.002049 0.000740 0.000000 0.000003
730 0.001440 0.000520 0.000000 0.000003
735 0.001000 0.000361 0.000000 0.000002
740 0.000690 0.000249 0.000000 0.000001
745 0.000476 0.000172 0.000000 0.000001
750 0.000332 0.000120 0.000000 0.000001
755 0.000235 0.000085 0.000000 0.000001
760 0.000166 0.000060 0.000000 0.000000
765 0.000117 0.000042 0.000000 0.000000
770 0.000083 0.000030 0.000000 0.000000
775 0.000059 0.000021 0.000000 0.000000
780 0.000042 0.000015 0.000000 0.000000
785 0.000029 0.000011 0.000000 0.000000
790 0.000021 0.000007 0.000000 0.000000
795 0.000015 0.000005 0.000000 0.000000
800 0.000010 0.000004 0.000000 0.000000
805 0.000007 0.000003 0.000000 0.000000
810 0.000005 0.000002 0.000000 0.000000
815 0.000004 0.000001 0.000000 0.000000
820 0.000003 0.000001 0.000000 0.000000
825 0.000002 0.000001 0.000000 0.000000
830 0.000001 0.000000 0.000000 0.000000
Appendix 125

Archaic and Arcane Units of Illumination

Luminous Intensity

The unit of luminous intensity is, and has been, the base
unit of photometry. Until recently, it suffered from a lack
of stable, reproducible standards. Over the past century
and a half, standards have ranged from actual candles
(wax or whale fat), gas lamps (pentane, isopropyl acetate),
vegetable oil lamps (colza, i.e. canola oil), carbon filament
lamps, blackbody furnaces, and finally, in 1979, a
radiometric standard. Because of the past difficulty in
realizing the standard, most of these expressions for
intensity are approximate at best.

1 Hefner candle ≈ 0.9 cd

1 candlepower* = 1 candela (cd)
1 new candle = 1 bougie nouvelle = 1 cd
1 candle (UK) ≈ 1 cd
1 decimal candle = 1 bougie decimal = 1.02 cd
1 international candle = 1.02 cd
1 Vereinskerze (German candle) ≈ 1.1 cd
1 pentane candle ≈ 10 cd
1 Munich candle ≈ 1.2 cd
1 carcel unit ≈ 9.8 cd
1 Violle ≈ 20.4 cd

Luminous Flux
1 spherical candlepower* (SCP) = 4π lumens (lm)
1 mean spherical candlepower* (MSCP) = 4π lm

Illuminance
1 nox = 0.001 lux
1 milliphot = 10 lux
1 footcandle* = 1 lm/ft2 = 10.764 lux
1 flame = 43.06 lux
1 cm-candle = 1 phot = 10,000 lux

* still in occasional use

126 Illumination

Archaic and Arcane Units of Illumination (cont.)

Luminance
Several units of luminance have the number π in the
denominator. This was done before the availability of
calculators and computers to facilitate the calculation of
the luminance of a Lambertian surface, which radiates
uniform luminance over a projected solid angle of π
steradians. For example, a Lambertian surface with
reflectance, ρ, receiving an illuminance of x lux has a
luminance of ρx apostilbs.
1 bril = 3.183 × 10-8 nit
1 skot = 3.183 × 10-4 nit
1 apostilb = 1 Blondel = 1 cd/πm2 = 0.3183 nit
1 millilambert = 3.183 nit
1 foot-Lambert* = 1 cd/πft2 = 3.426 cd/m2 = 3.426 nit
1 cd/ft2 = 10.76 nit
1 Lambert = 1 cd/πcm2 = 3183 nit
1 stilb = 1 cd/cm2 = 10,000 nit

CCT
1 mired* (microreciprocal degree) = 1 reciprocal
megaKelvin (MK)-1
1 mirek* (microreciprocal Kelvin) = 1 reciprocal
megaKelvin (MK)-1

Photons
1 Einstein† = Avagadro’s number of photons
1 Einstein† = 6.022 × 1023 photons
1 micromole† = 6.022 × 1017 photons in the 400- to 700-nm
band

Photon Radiance
1 Rayleigh† = 7.96 × 10-8 photons/sec⋅m2⋅sr

Wavelength
1 millimicron* = 1 nanometer

* still in occasional use

† still in occasional use in fields other than illumination
or optics
 127
Bibliography

I. Ashdown, “Chromaticity and color temperature for

achitectural lighting,” Proc. SPIE 4776, 51–60 (2002).
L. Audin, D. Houghton, M. Shepard, and W. Hawthorne,
Lighting Technology Atlas Series, Volume 1, E Source,
Boulder, CO (1994).
ASTM International, Standard Practice for Describing
Retroreflection, ASTM E808-01 (2002).
ASTM International, Standard Practice for Computing
the Colors of Fluorescent Objects from Bispectral
Photometric Data, ASTM E2152-01 (2001).
ASTM International, Standard Practice for Obtaining
Bispectral Photometric Data for Evaluation of Fluorescent
Color, ASTM E2153-01 (2001).
F. Bartell, “Projected solid angle and blackbody
simulators,” Appl. Opt. 28(6), 1055–1057 (1989).
R. Berns, F. Billmeyer, and M. Saltzman, Principles of
Color Technology, Wiley, New York (2000).
B. R. Boylan, The Lighting Primer, Iowa State University
Press, Ames, IA (1987).
J. Brogan “Light in architecture,” Architectural Design
67(3/4), March-April 1997.
a . Carlson and C. Clark, “Light sources for optical
.F
devices,” in Applied Optics and Optical Engineering, Vol.
1, R. Kingslake, Ed., 60–61, Academic Press, New York
(1965).
W. Cassarly, Design of Efficient Illumination Systems,
unpublished notes for SPIE short course (2000).
a ommission internationale de l'éclairage (CIE), Method of
.C
Measuring and Specifying Colour Rendering Properties of
Light Sources, CIE Publication 13.3-1995.
Commission internationale de l'éclairage (CIE),
Colorimetry, 3rd Ed., CIE Publication 15:2004.
128
Bibliography (cont.)

Commission internationale de l'éclairage (CIE),

Recommended File Format for Electronic Transfer of
Luminaire Photometric Data, CIE Publication 102-1993.
Commission internationale de l'éclairage (CIE), The
Photometry and Goniophotometry of Luminaires, CIE
Publication 121-1996.
Commission internationale de l'éclairage (CIE),
Measurement of LEDs, 2nd Ed., CIE Publication
127:2007.
C. De Cusatis, Ed., Handbook of Applied Photometry, AIP
Press, Woodbury, NY (1997).
C. Dimas, S. Read, and J. Kuta, “Integrating rod
homogeneity as a function of cross-sectional shape,” Proc.
SPIE 4768, 82–92 (2002).
M. D. Egan, Concepts in Architectural Lighting, McGraw
Hill Book Co., New York (1983).
Electro-Optics Handbook,.Burle Industries, Lancaster, PA,
www.burle.com (1974).
Eley Associates, Advanced Lighting Guidelines, 1993
(Revision 1), Final Report Prepared for California Energy
Commission (1993).
J. Flynn and S. Mills, Architectural Lighting Graphics,
Reinhold Publishing Corporation, New York (1962).
J. Flynn, A. Segil, and G. Steffy, Architectural Interior
Systems, Van Nostrand Reinhold Company, Inc., New
York (1988).
C. Gardner and B. Hannaford, Lighting Design, Design
Council, London (1993).
G. Gordon, Interior Lighting for Designers, 4th Edition,
John Wiley & Sons, Inc., Hoboken, NJ (2003).
J. Greivenkamp, Field Guide to Geometrical Optics, SPIE
Press, Bellingham, WA (2004).
 129
Bibliography (cont.)

A. Gupta, J. Lee, and R. J. Koshel, “Design of efficient

lightpipes for illumination by an analytical approach,”
Appl. Opt. 40, 3640–3648 (2001).
K. G. Holmes, Essays on Lighting, Crane, Russak &
Company, Inc., New York (1975).
R. G. Hopkinson, Architectural Physics: Lighting, Her
Majesty’s Stationary Office, London (1963).
R. Hopkinson and J. B. Collins, Ergonomics of Lighting,
MacDonald Technical & Scientific, London (1970).
R. G. Hopkinson and J. D. Kay, The Lighting of Buildings,
Faber and Faber, London (1972).
D. Hunten, F. Roach, and J. Chamberlain, “A photometric
unit for the airglow and aurora,” Journal of Atmospheric
and Terrestrial Physics (GB) 8, 345–346 (1956).
Illuminating Engineering Society of North America,
IESNA Handbook, 9th Edition, New York (2003).
Illuminating Engineering Society of North America,
Goniophotometer Types and Photometric Coordinates, LM-
75-01, New York (2001).
Illuminating Engineering Society of North America,
Lighting For Exterior Environments, IESNA RP-33-99
(1999).
Illuminating Engineering Society of North America,
Recommended Practice of Daylighting, IESNA RP-5-99
(1999).
Illuminating Engineering Society of North America,
Recommended Practice for Roadway Lighting, IESNA RP-
8-00 (2000).
W. Jankowski, The Best of Lighting Design, PBC
International, Inc., New York (1987).
W. Jankowski, Lighting the Workplace, PBC
International, New York (1988).
130
Bibliography (cont.)

D. Jenkins and H. Mönch, SID ‘00 Digest, Society for

Information Display (2000).
D. Jenkins and R. Winston, “Tailored reflectors for
illumination,” Appl. Opt. 35(10), 1669–1672 (1996).
M. E. Kaminski, K. J. Garcia, M. A. Stevenson, M. Frate,
and R. J. Koshel, “Advanced topics in source modeling,”
Proc. SPIE 4775, 46–57 (2002).
N. Karlen and J. Benya, Lighting Design Basics, John
Wiley & Sons, Inc., Hoboken, NJ (2004).
H. Kaufman and J. Christensen, IES Lighting Handbook,
1987 Application Volume, Illuminating Engineering
Society of North America, New York (1987).
R. Kingslake, “Illumination in optical images,” in Applied
Optics and Optical Engineering, Vol. 2, R. Kingslake, Ed.,
Academic Press, New York (1965).
R. J. Koshel, “Simplex optimization method for
illumination design,” Opt. Lett. 30, 649–651 (2005).
R. J. Koshel, lecture notes from the Illumination Seminar,
College of Optical Sciences, University of Arizona (2006).
R. J. Koshel and I. A. Walmsley, “Modeling of the gain
distribution for diode pumping of a solid-state laser rod
with nonimaging optics,” Appl. Opt. 32, 1517–1527 (1993).
R. J. Koshel and I. A. Walmsley, “Non-edge ray design:
improved optical pumping of lasers,” Opt. Eng. 43, 1511–
1521 (2004).
L. Larson, Lighting and Its Design, Whitney Library of
Design, New York (1964).
Y. LeGrande, Light, Colour, and Vision, Chapman and
Hall, London (1968).
J. Lindsey, Applied Illuminating Engineering, Second
Edition, The Fairmont Press, Inc, Lilburn, GA (1997).
J. A. Lynes, Developments in Lighting-1, Applied Science
Publishers LTD, London (1978).
 131
Bibliography (cont.)

M. J. J. J. B. Maes and A. J. E. M. Janssen, “A note on

cylindrical reflector design,” Optik 88(4), 177–181 (1991).
R. McCluney, Introduction to Radiometry and Photometry,
Artech House, Boston (1994).
M. Millet, Light Revealing Architecture, John Wiley &
Sons, Inc., Hoboken, NJ (1996).
P. Moon, Scientific Basis of Illuminating Engineering,
McGraw-Hill Book Company, New York (1936).
F. Nicodemus, Ed., Self-Study Manual on Optical
Radiation Measurements, National Institute of Standards
and Technology:
http://physics.nist.gov/Divisions/Div844/manual/studyman
ual.html
D. Phillips, Lighting in Architectural Design, McGraw-
Hill Book Company, New York (1964).
D. Phillips, Lighting Modern Buildings, Architectural
Press, an imprint of Butterworth-Heinemann, Boston
(2000).
M. D. W. Pritchard, Lighting, Environmental Physics
Series, Elsevier Publishing Company, New York (1969).
H. Ries and J. Muschaweck, “Tailored freeform optical
surfaces,” J. Opt. Soc. Am. 19(3), 590–595 (2002).
R. Rowlett, “How Many? A Dictionary of Units of
Measurement,” The University of North Carolina at
Chapel Hill http://www.unc.edu/~rowlett/units/index.html
(2006).
M. Ruda, Introduction to Illumination Design Techniques,
unpublished notes for SPIE short course (1997).
R. Rykowski and C. B. Wooley, “Source modeling for
illumination design,” Proc. SPIE 3130, 204–208 (1997).
R. Siegel and J. Howell, Thermal Radiation Heat
Transfer, 4th ed., Taylor and Francis, New York (2002).
132
Bibliography (cont.)

D. Sinclair, Phluometry, unpublished notes, Institute of

Optics, University of Rochester (1970).
F. K. Smith and F. J. Bertolone, Bringing Interiors to
Light, Whitney Library of Design, New York (1986).
W. Smith, Modern Optical Engineering, SPIE Press,
Bellingham, WA, and McGraw Hill, New York (2000).
G. Steffy, Architectural Lighting Design, Second Edition,
John Wiley & Sons, Inc., New York (2002).
G. Steffy, Architectural Lighting Design, Second Edition,
Van Nostrand Reinhold Company, New York (1990).
G. Steffy, Lighting the Electronic Office, Van Nostrand
Reinhold, New York (1995).
B. Stein and J. Reynolds, Mechanical and Electrical
Equipment for Buildings, 10th Edition, Wiley & Sons,
New York (2005).
M. A. Stevenson, M. Frate, M. E. Kaminski, and R. J.
Koshel, “Modeling filament-based sources for system
tolerancing,” Proc. SPIE 4775, 67–77 (2002).
W. Welford and R. Winston, High Collection Nonimaging
Optics, Academic Press, San Diego (1989).
R. Winston, J. C. Minano, and P. Benitez, Nonimaging
Optics, Elsevier Academic Press, Burlington, MA (2005).
R. Winston, and H. Ries, “Nonimaging reflectors as
functionals of the desired irradiance,” J. Opt. Soc. Am.
10(9), 1902–1908 (1993).
W. Wolfe, Introduction to Radiometry, SPIE Press,
Bellingham, WA (1998).
G. Wyszecki and W. Stiles, Color Science, Wiley, New
York (1982).
 133
Index

Abbe illumination, 58 candela distribution curve

Abel transform, 74 (CDC), 97
ABg method, 29 catadioptric, 81
absorbance, 27 center wavelength, 26
absorbing filter, 61 centroid wavelength, 26
absorptance, 27 chromaticity, 12
accent layer, 96 CIE 1924 luminous
accent lighting, 96 efficiency function, 5
acceptance angle, 51, 75 CIE 1931 chromaticity
apparent brightness, 95 coordinates, 12
arc lamp, 74 CIE 1960 UCS coordinate
area-solid-angle-product, 49 system, 13
average projected solid CIE 1976 UCS coordinate
angle, 37 system, 13
average projected solid CIE color matching
angle, 36 function, 5
average reflectance, 67 CIE color matching
averaged LED intensity, 11 functions, 12
cobra head, 110
back reflector:, 83 coefficient of retroreflected
backlight lightguides, 85 luminance, 32
backlighting, 82, 96, 104 coefficient of retroreflected
backlit displays, 82 luminous intensity, 32
backlit LCD, 83 coefficient of retroreflection,
backside reflector, 85 32
base type:, 73 coefficient of utilization
beam pattern distributions, (CU), 98
106 cold mirrors, 61
bend ratio, 66 cold-cathode fluorescent
bent lightpipes, 66 lamps (CCFL), 82, 84
bidirectional reflectance color, 26
distribution function color rendering index (CRI),
(BRDF), 28, 29 18, 19
bikeways, 104 color temperature, 14
brightness enhancement color-pixel patterns, 86
film, 83 compound elliptical
brightness enhancement concentrator (CEC), 76,
film (BEF), 82, 86 78
building entries, 104 compound hyperbolic
bulb type, 73 concentrator (CHC), 76,
bumps, 85 78

candela, 5
134

Index

compound parabolic entrance angle, 31

concentrator (CPC), 76, étendue, 49
77 exitance, 1
computer-aided design eye adaptation, 94
(CAD), 92
concentration (C), 51 f-number, 48
concentration ratio, 51 façade illumination, 102
condition A, 11 faceted reflectors, 80
condition B, 11 facets, 69
configuration factor, 37 features, 85
conservation of étendue, 50 features, 83
controls, 105 filament type, 73
correlated color flat-fielding, 88
temperature, 13 floodlighting, 103
correlated color flow-line method, 78
temperature (CCT), 14 fluorescence, 17
cosine-fourth law, 41 fluorescent lamps, 74
flux, 1, 33, 35, 49
damp location rated, 105 focal layer, 96
daylight factor (DF), 99 form factor, 36, 37
daylighting, 100 frontlighting, 104
delta, 86 front-projection displays, 82
diagonal, 86 full cutoff, 110
diffusers, 83
diffusion equation, 85 general color rendering
digital light processing index, 18
(DLP), 87 general illumination
dioptric, 81 approach, 96
displays, 82 generalized étendue, 50
dominant wavelength, 15, geometrical model, 71
26 glare, 101
donaldson matrix, 17 glazing, 96
downlighting, 104 goniometer, 71
downlighting, 96 goniophotometers, 89, 91
dual brightness
enhancement film hardscape lighting, 104
(DBEF), 86 Harvey method, 29
high-pressure sodium, 111
electroluminescent (EL), holes, 85
83, 84 horizontal ambient layer,
ellipsoidal mirror, 60 96
emergency egress doors, horizontal field, 94
104 hot mirrors, 61
 135
Index

hybrid optics, 81 liquid crystal on silicon,

(LCoS), 87
illuminance, 6, 7, 108 locating fixtures, 101
image irradiance, 49 long lateral distribution,
incandescent lamp, 73 107
increase factor, 41, 47 lumen, 5
indium-tin oxide (ITO), 86 luminaire’s transverse
induction lamps, 111 reach, 106
injection molding, 85 luminaires, 76
injector, 84 luminance, 6, 7, 108
injector, 83 luminances from a direct
integrating spheres, 67 luminaire, 98
intensity, 1, 3, 33, 89, 91 luminous efficacy, 9
interference filters, 61 luminous flux, 5, 6
inverse square, 3 luminous intensity, 6
inverse square law
calculations, 108 maximum overall length
involute, 78 (MOL), 73
irradiance, 1, 33, 65, 67, 75 medium beam, 103
isotropic source, 33 medium fixtures, 107
mesopic region, 8
known intensity, 34 mesopic vision, 101
Köhler illumination, 68 metal halide, 111
metal halide lamps, 111
Lambertian disk, 47 mixed bundle, 55
Lambertian radiator, 36 mixing rod, 65
Lambertian source, 3, 33
Lambertian surface, 27, 28, narrow beam, 103
29 nighttime visibility, 101
LED, 72, 84 nonedge-ray design, 81
lens design codes, 92 nonimaging compound
lenslet array, 68, 69 concentrators, 75, 76
light center length (LCL), nonimaging Fresnel lens
73 design, 81
lightbox, 84 nonsequential ray tracing,
lightguide, 83 92
lightguides, 53 numerical aperture (NA),
light-loss factor (LLF), 111 43, 48
lightpipes, 53
linear polarizers, 86 observation angle, 31
liquid crystal, 83 OLED displays, 82
optical density (OD), 27
optical fibers, 53
136

Index

optics analysis codes, 92 scotopic efficiency, 8

optimal concentration, 51 scotopic region, 8
organic LED (OLED), 82 scotopic vision, 101
ornamental layer, 96 searchlight, 63
semicutoffs, 110
packing fraction (pf), 55 setback, 103
paraboloidal mirror, 60 shape conversion, 55
peak wavelength, 26 shielding the light source,
perfectly reflecting diffuser 101
(PRD), 28 short fixtures, 107
photometric brightness, 95 sidelighting, 104
photometric report, 97 sidewalks, 104
photometry, 5 simultaneous multiple
photopic efficiency, 8 surfaces method (SMS),
photopic region, 8 81
photopic vision, 101 skew invariant, 52
Planckian locus, 12 small target visibility
plane angle, 2 (STV), 108, 109
polarizers:, 83 softscape lighting, 104
principal sections, 66 solid angle, 2, 3, 4, 35, 43,
projected area, 3 49, 54, 75
projected solid angle, 3, 4, solid-based geometry, 92
37, 43, 48 source, 83
projection, 82 source brightness, 101
projection displays, 82, 87 source model, 70
purity, 15 source modeling methods,
71
radian, 2 spacing, 103
radiance, 1, 3, 33, 65, 67 spacing criteria (SC), 98
radiance model, 71 spatial light modulator
ramps, 104 (SLM), 87
rated current, 21 spatial light modulators
rear-projection displays, 82 (SLMs), 82
rectilinear shoebox, 110 spectral filter mask, 86
reduction of light levels, spectral flux, 5
101 spectral intensity, 5
reference illuminant, 18 spectral irradiance, 5
reflectance, 27, 28 spectral radiance, 5
reflectance coefficient, 108 spherical mirror, 57
reflectance factor, 16, 28, 30 splitting the bundle, 55
retroreflectance factor, 32 stairs, 104
retroreflectors, 31 steradian, 2
 137
Index

stippled illumination veiling luminance ratio

pattern, 80 (VLR), 108
stripes, 86 vertical ambient layer, 96
structure, 85 vertical field, 94
surface-based geometry, 92 viewing angle, 26, 31
system model, 71 vignetting, 47
visibility level (VL), 109
T-number, 48 visual comfort probability
tailored-edge-ray, 80 (VCP), 97
tailored-edge-ray reflectors, visual fields, 94
79
task layer, 96 walkways, 104
test-color samples, 18 wall washing, 96
throughput, 49 wet location rated, 105
total integrated scatter wide beam, 103
(TIS or TS), 29 working f-number, 48
total irradiance, 5
total radiance, 5
total radiant flux, 5
total radiant intensity, 5
transient adaptation, 94
transmittance, 27, 45
triangular, 86
tristimulus values, 12
twisted-nematic LC
module, 86
type A spherical
coordinates, 90
type B spherical
coordinates, 90
type C spherical
coordinates, 90
type I, 106
type II, 106
type III, 106
type IV, 106

uniform circular
Lambertian disk, 40
uniform light distribution,
101
uplighting, 104
 Angelo Arecchi is the Director of
Systems Engineering at Sphere-
Optics, LLC. He also manages his
own consulting firm, Sunrise
Instruments, LLC. He was the Vice
President of Engineering at
Labsphere, Inc. for 14 years. In
addition to his work in the optics industry, Angelo
spent more than 20 years in the U.S. Coast Guard
where, among other assignments, he worked in visual
signaling, electronic communications and aids-to-
navigation. He was also on the faculty of the U.S.
Coast Guard Academy for several years. Angelo holds
an M.S. degree in Optics from the University of
Rochester, an M.B.A. from Plymouth State
University, and is a registered Professional Engineer.
He is an adjunct faculty member at Plymouth State
University and at Norwich University. Angelo is a
member of SPIE, OSA, and the Council on Optical
Radiation Measurements (CORM), where he serves on
its Board of Directors.

Tahar Messadi is professor at the

School of Architecture at the
University of Arkansas. He teaches
environmental building systems and
lighting courses, and manages a
design studio focused on environ-
mentally responsive architecture. His
research interests include the lighting and thermal
performance of buildings. Dr. Messadi has codirected
the smart façade research program at the Georgia
Institute of Technology. Other sponsored research
includes the development of a toolkit to monitor
lighting, IAQ, and comfort in high performance
schools. He is also the recipient of a number of funds
to support investigations conducted by students in
environmental technology, specifically, lighting,
thermal, and acoustics. Prof. Messadi has authored
and coauthored numerous publications in national
and international conferences and journals.
 R. John Koshel is the Senior Staff
Engineer at Lambda Research
Corporation and Adjunct Assistant
Professor at the College of Optical
Sciences, University of Arizona. At
Lambda he works on the TracePro
nonsequential optical analysis
code, especially in the field of illumination. At Arizona
he has taught courses on Radiometry and
Illumination Engineering, and he has worked on
illumination research projects. His primary research
areas are nonimaging optics, solid-state lighting,
optimization and tolerancing, and lit-appearance
modeling. He is active in SPIE and OSA. He was chair
of the International Optical Design Conference in
2006 and serves as a chair or member for a number of
SPIE conferences. He obtained his B.S. and Ph.D.
degrees from The Institute of Optics, University of
Rochester.