Understanding Glare in Exterior Lighting, Display, and Related Applications 9781510654839, 1510654836

Citation preview

Understanding Glare in Exterior Lighting, Display, and Related Applications John D. Bullough

Understanding Glare in Exterior Lighting, Display, and Related Applications by John D. Bullough doi: http://dx.doi.org/10.1117/3.2639845 PDF ISBN: 9781510654839

Published by SPIE Press P.O. Box 10 Bellingham, Washington 98227-0010 USA Phone: +1 360.676.3290 Fax: +1 360.647.1445 Email: [email protected] Web: http://spie.org Copyright © 2022 Society of Photo-Optical Instrumentation Engineers (SPIE) All rights reserved. No part of this publication may be reproduced or distributed in any form or by any means without written permission of the publisher. The content of this book reflects the work and thoughts of the author(s). Every effort has been made to publish reliable and accurate information herein, but the publisher is not responsible for the validity of the information or for any outcomes resulting from reliance thereon.

Spotlight vol. SL63 Last updated: 9 August 2022

Table of Contents Preface

v

1 Introduction

1

2 What Is Glare and Why Does It Matter?

3

2.1 2.2 2.3 2.4

Disability glare Discomfort glare Glare recovery Reflected glare

3 4 6 8

3 What Factors Affect Glare? 3.1 3.2 3.3 3.4 3.5 3.6 3.7 3.8

8

Light source intensity and illuminance Light source luminance Background luminance Light source location Light source spectral distribution Glare exposure duration Observer age and other characteristics Psychological factors in discomfort glare

4 Modeling and Predicting Glare 4.1 4.2 4.3 4.4

32

Disability glare modeling Discomfort glare modeling Glare recovery modeling Reflected glare modeling

32 33 35 35

5 Hints for Designing to Control Glare 5.1 5.2 5.3 5.4

8 11 15 17 21 27 29 30

36

Disability glare control Discomfort glare control Glare recovery control Reflected glare control

36 36 37 38

6 Future Outlook 7 References

38 40

iii

SPIE Spotlight Series Welcome to the SPIE Spotlight series! This growing collection of concise eBooks serves as an entry point for particular topics in optics and photonics suitable for researchers, engineers, managers, executives, and educators. Spotlights fill the community need for timely and relevant references at a level of detail bridging the gap between in-depth journal articles and broad fundamental tutorials. Whatever your interest or need, we hope this series meets your expectations and encourage you to submit your own ideas for future Spotlights online. Craig Olson, Series Editor L3 Technologies

Associate Editors

Brian Sorg National Cancer Institute Biomedical Optics/Medical Imaging

Richard Blaikie University of Otago Semiconductor, Nano-, and Quantum Technology

Keith Kasunic Optical Systems Group Aerospace and Defense

Matthew Jungwirth CyberOptics Corp. Optical Design and Engineering

Ronian Siew inopticalsolutions Aerospace and Defense

Wei Li Stanford University Energy and the Environment

,

Preface When I tell people I work in lighting, I often get a blank stare in return, or the question “Why?” When I mention that a lot of my research has to do with seeing better while driving at night, those blank stares often turn into impassioned pleas for “doing something about those bright headlights!” Like many things, lighting is something that is noticed mainly when it’s doing something wrong. Glare is an unfortunate part of lighting that has given me an opportunity to “do something” to help make walking and driving at night safer and more comfortable by identifying glare’s causes and countermeasures. My investigations of glare have been published across many different journals, conference proceedings and technical reports, so I am grateful to SPIE Press for allowing me to consolidate them into this Spotlight. I hope this eBook can be a genesis for an Integrated Glare Metric (IGM) to help us tame the glare from headlights, streetlights, displays and windows. Many thanks to Mark Rea and Mariana Figueiro at the Mount Sinai Light and Health Research Center (LHRC) for urging me to submit my book proposal to SPIE, and to Mark for feedback on my drafts. I also appreciate the support from the members of the LHRC’s Light for Transportation Safety Partnership (ams OSRAM, Audi, General Motors, Hella, Lumileds, Marelli, and Varroc) that made it possible for me to complete this project.

v

Bullough: Understanding Glare in Exterior Lighting, Display, and Related Applications

1

1 Introduction Over the past two decades, lighting for the built environment has under-gone nothing less than a revolution. The advent of new lighting technologies, particularly through solid state lighting systems including the development of light emitting diode (LED) sources, has resulted in dramatic increases in lighting system efficacy and operating life.1 Early on,2 solid state lighting technologies had primarily been used for limited applications including signal indicators, signage, and low-light uses such as pocket flashlights. At present, LED sources are used in every general lighting application in homes, workspaces, outdoor areas, and vehicles. At the same time this lighting revolution has been underway, there has also been a growing, perhaps renewed3 awareness of some of the problems that suboptimal lighting can create, including glare. One well-known example of an application where glare has been viewed as a widespread problem is that of automotive lighting, especially vehicle headlighting. 4 Up until the mid-to-late 1990s, most vehicle headlights used filament (incandescent) sources, mostly halogen lamps and bulbs. These produced a yellowish-white illumination color seen as “warm” white. Around this time, some vehicles began to appear that had high-intensity discharge (HID) headlights,5 using sources such as metal halide lamps with xenon fill gases to facilitate rapid starting. (Metal halide lamps had also been widely used for outdoor lighting in parking lots and streets; these lamps could take several minutes to achieve full light output upon starting,6 a situation that would be unacceptable for vehicle headlights!) Not only did HID headlights have a “bluer” or “cooler” appearance than halogen headlights, but they also had higher efficacy, meaning they could produce roughly twice the light output (in lumens) as 55 W halogen headlights while using only 35 W of power. Vehicle headlights are strictly controlled by federal regulations7 in terms of their maximum intensity in directions that correspond to the roadway directly ahead, but there are few limits on their intensities that correspond to peripheral directions. This may be advantageous to forward visibility8 but could result in higher intensities reaching the eyes of oncoming drivers. More recently available LED headlights also tend to have a “bluer” color appearance than halogen headlights and yet higher efficacy than HID sources. As shall be seen later in this Spotlight, there is a correlation between “blue” light content and glare sensations. Two other trends related to automotive lighting (at least in the United States) also occurred simultaneously with the shift in lighting technology. One is the increased proportion of pickup trucks and sport-utility vehicles (SUVs) in the overall vehicle fleet. These vehicles are taller than most passenger cars, and because headlight intensity standards7 are made without regard to the mounting height of the headlights for consumer vehicles, the same headlight can produce higher intensities toward an oncoming driver when it is mounted on a pickup or SUV than when it is mounted on a passenger car. The second trend is that a

2

Bullough: Understanding Glare in Exterior Lighting, Display, and Related Applications

greater proportion of states have dropped requirements to check headlight aim as part of the vehicle safety inspection process.9 Consequently, measurements of vehicles currently on the road have shown that the majority of them have at least one headlight that is misaimed,10 either too low (which limits forward visibility) or too high (which can increase glare to oncoming drivers). Together, all of these factors have created a “perfect storm” of conditions that has made instances of severe glare while driving at night more frequent. Perhaps the most noticeable of them to an oncoming driver is the “bluer” color of many new headlights, and this characteristic has received a lot of attention from aggrieved drivers. The National Highway Traffic Safety Administration (NHTSA), which writes the federal vehicle headlighting standards, issued in 2001 a call for public comments11 related to nighttime driving and headlight glare when complaints began to trickle into this agency about glare. Whereas NHTSA might receive a few hundred comments from the public on “hot topics” such as tire blowouts or child safety restraints, over 5000 responses were submitted to NHTSA’s call for comments! Many of these referenced the “blue” color of the new headlights compared to the “yellowish” or “warm” color of halogen lights. Clearly, the prospect of experiencing glare while driving at night can animate a substantial amount of interest in this topic among the public. Concerns about glare from lighting installations are not limited to vehicle headlights. The proliferation of LED streetlights has also brought with it a growing awareness of glare12 for both drivers and pedestrians at night. Unlike “cobrahead” types of high-pressure sodium streetlights, which have been available for half a century and have been largely standardized in terms of their shape and optics, LED streetlights have a wide variety of configurations and optical designs. In some LED streetlights, the LED sources are directly visible, and this has elicited unpleasant sensations of glare.13 Another lighting application where glare has been cited as an important concern is the use of flashing lights on emergency and maintenance vehicles.14 These have evolved toward an increasing use of LED sources, and the increased efficacy has resulted in brighter flashes of light of highly saturated colors. Current standards for these lights15,16 specify minimum intensity levels to ensure that they can be easily seen, especially during bright daytime conditions, but there are no upper limits for intensity, which means they can be very glaring at night. Nor is glare necessarily limited to outdoor and nighttime applications of lighting. In fact, there is a growing interest in benefits of daylighting. Consider that outdoor light levels from daylighting can range between 10,000 and 100,000 lux, whereas interior light levels from electric lighting systems often range from 100 to 500 lux.6 This disparity in brightness can result in glare from windows and skylights.17 Consider also the potential for building lighting to support not only occupants’ vision but also their circadian rhythms.18 It is likely that lighting schemes to address this objective will require higher illuminances in building interiors during the daytime (perhaps in combination with lower

Bullough: Understanding Glare in Exterior Lighting, Display, and Related Applications

3

illuminances at night), possibly increasing glare. As will be seen, glare in interior lighting is a smaller concern than outdoors at night, and this Spotlight will focus more on glare at night and in otherwise dim environments, but the same principles that can help mitigate nighttime glare can also do so for interior, daytime lighting. The goal of this Spotlight is to provide the reader with a basic understanding of the definitions of different types of glare and the impacts it can have on occupant visibility and visual comfort. The emphasis will be on exterior lighting and visual display applications, although the discussion is relevant to interior lighting as well. The factors that can influence glare will also be described, with special reference to calculation models that can help the reader predict how different lighting configurations will impact occupants of lighted areas and spaces. A few considerations for the design and implementation of lighting to minimize glare will also be provided, with a brief discussion of the future for glare modeling and control, potentially leading to an Integrated Glare Metric (IGM).

2 What Is Glare and Why Does It Matter? It could be argued that glare is a subjective response to lighting conditions and that one person’s glare could be another person’s useful illumination. Of the 5000+ responders to NHTSA’s request for public comments on headlight glare,11 it seems unlikely that most of them have a formal definition of glare in mind. Rather (with apologies to the late Justice Potter Stewart), most people “know it [glare] when they see it.” In this chapter, the definitions for several different aspects of glare are provided. It may be possible for more than one of these aspects to be present in any given situation. 2.1 Disability glare Disability glare is defined6 as the reduction in contrast caused by light entering the eye that is scattered by the eye’s optical media (e.g., cornea and lens). Some of this light falls on the retina, creating a luminous “veil” that is superimposed over the retina and over the retinal images of objects in the field of view. Figure 1 shows an image of a letter “C” that is not obscured, and the same letter obscured by a luminous veil. The luminance (or “brightness”) of the background and the letter are both increased by the superimposed veil, and as a result its contrast is reduced. It is this contrast reduction that causes some objects to become less visible or even to disappear in the presence of a bright light source. As will be seen, the magnitude of the reduction in contrast depends upon the locations of the object being viewed, and of the source of glare, within the field of view. Importantly, disability glare can just as readily be produced by a large, diffuse source of light as by a small, intense point source, and consequently be unnoticed by an observer. For example, while driving at night, a distant but bright intense

4

Bullough: Understanding Glare in Exterior Lighting, Display, and Related Applications

Figure 1 (a) Image without a luminous veil. (b) Image with a luminous veil superimposed over it.

light source or the dashboard lights of the driver’s vehicle might produce a similar veiling luminance over an object such as a pedestrian walking along the roadway. Each would reduce the contrast between the pedestrian and the pedestrian’s background by an equal amount and hence would reduce visibility equally, but the effect would likely be less noticeable for the dashboard lights because of their diffuse luminous appearance. This will be addressed further in the next chapter of this Spotlight. Interestingly, every object in the field of view that emits or reflects even a small amount of light theoretically contributes to disability glare because all light entering the eye is scattered to some degree. Under most conditions, the brightness of the luminous veil created by scattered light is nearly negligible relative to the brightness or luminance of the objects and surfaces in the line of sight. Only in the case of very low ambient light conditions, such as nighttime viewing outdoors, or in the case of very intense light sources, such as the sun during the daytime, will the relative intensity of a potential glare source compared with the overall field of view be large enough to contribute to a meaningful degree of disability glare. 2.2 Discomfort glare Discomfort glare (DG) is defined6 as the sensation of annoyance or even pain that can be experienced when a very bright, intense source of light is present in the field of view. Although it must be based in physiology, it is observed as a subjective response to light. For this reason, it has sometimes been given the name “psychological glare” in contrast with disability glare, which has sometimes been called “physiological glare.”19 Indeed, one of the most common ways that DG has been evaluated in research studies has been through subjective rating scales, such as the scale devised by De Boer.20 In this nine-point scale, odd-numbered values

Bullough: Understanding Glare in Exterior Lighting, Display, and Related Applications

5

Table 1. Rating Scale for Discomfort Glare (Reference 20). Rating value

Descriptor

9

Just noticeable (or no) glare

8

—

7

Satisfactory

6

—

5

Just permissible

4

—

3

Disturbing

2

—

1

Unbearable

are given descriptors that individuals can use to assess the level of glare they are experiencing at a particular moment (Table 1). In the rating scale listed above, respondents can use even-numbered responses to indicate conditions between the bracketing odd-numbered values; for example, a rating value of 2 would correspond to glare that is beyond a criterion of “disturbing” (3) but does not reach the threshold for a criterion of “unbearable” (1). A perhaps counterintuitive feature of this scale is that lower numerical values correspond to greater sensations of visual discomfort. This is caused by the scale’s origin as a figure of merit;20 higher numerical values correspond to greater quality of the lighting and visual conditions (e.g., greater visual comfort). It might be noticed that disability glare (Section 2.1) and DG described in the present section can both be present simultaneously, such as when a driver at night views the bright headlights of an approaching driver’s vehicle. The illumination from the oncoming headlights can produce scattered light, which against the relatively dark nighttime roadway environment could produce a luminous veil sufficient to make roadside hazards more difficult to see by reducing their contrast. And nearly all people with driving experience at night have experienced the uncomfortable sensation of oncoming headlights, perhaps when the other driver has forgotten to switch their headlights from high to low beams, or if their headlights are misaimed. Yet disability and DG can also be decoupled. Taking the example of bright dashboard lights while driving at night, these are rarely so bright that they would cause discomfort to the driver, yet as discussed in the previous section, they can contribute to disability glare. Figure 2 illustrates a typical nighttime driving scene. In principle, any of the sources of light (e.g., headlights, streetlights, traffic signals, or dashboard lights) could serve as sources of disability glare, but usually,

6

Bullough: Understanding Glare in Exterior Lighting, Display, and Related Applications

Figure 2 Sources of light in the nighttime driving visual environment that could serve as sources of disability and/or DG.

only those with a concentrated optical image (e.g., headlights, and some streetlights and traffic signals) would be likely to be major contributors to DG. The preceding argument does not imply that DG is rarer than disability glare, however. In fact, people generally seem to be more aware of DG than disability glare. It is likely that among the thousands of respondents to NHTSA’s request for public comments on headlight glare11 that most of them expressed concerns about DG rather than disability glare. Further, although the presence of DG does not by itself mean that someone has reduced visibility (a small light source viewed against a dark background could be uncomfortable in the field of view but might not produce a luminous veil bright enough to hamper visibility), this does not mean that DG has no implications for safety. A large body of research literature indicates that driving is safer when a driver maintains a constant speed along the road21 and a constant lateral position within the driving lane.22 Just as stressors such as thermal discomfort can lead to increased driving behaviors that are related to reduced safety, so can visual discomfort from the presence of headlight glare.23,24 For example, drivers may move their heads to look away from the road and fluctuate their driving speed in response to oncoming headlight illumination or a view of headlights in their rear-view mirrors. 2.3 Glare recovery Many people who have driven at night in the presence of uncomfortably bright oncoming headlights have undoubtedly breathed a sigh of relief after the offending vehicle has passed by, along with its associated disability or DG. Yet the effects of glare are not limited to the period of time when the so-called glare

Bullough: Understanding Glare in Exterior Lighting, Display, and Related Applications

7

source is visible in the field of view. Following exposure to a relatively bright source of light, the visual system requires a finite amount of time to re-adapt to a lower overall light level.6 An example of this is entering a darkened theater from a bright, sunny day. For a few moments, it may be very challenging to make one’s way around the theater while the visual system re-adapts from the bright outdoors to the relatively dim environment indoors. Qualitatively, this is illustrated by the dark adaptation curve in Fig. 3, which shows how the cone photoreceptors (which support vision at high light levels) and the rod photoreceptors (which support vision at very low light levels) in the eye can take over a half-hour to adapt from very high intensities like those on sunny days to extremely dark ones like completely unlighted, windowless spaces. Fortunately, the differences in adaptation that need to be overcome while driving at night are much smaller and can be overcome within seconds rather than minutes.25 (Figure 3 also shows that the cones and rods have different dark adaptation characteristics.) Similar phenomena can occur following exposure to a light source that produced either disability or DG, and this is what is meant by glare recovery. During a (hopefully) brief period of time after a bright light passes by or is extinguished, visual functioning can be temporarily reduced. This form of glare response is important for drivers to be aware of because one might assume that after any sensations of discomfort or any reductions in visibility have passed,

Figure 3 A dark adaptation curve showing the time course of adaptation from a very bright to a totally dark environment. The initial part of the curve shows the adaptation curve for the cone photoreceptors in the eye; the latter part shows the curve for the rod photoreceptors. Source, Wikimedia Commons (Creator: Dgtdsgn; Creative Commons License CC BY-SA 3.0).

8

Bullough: Understanding Glare in Exterior Lighting, Display, and Related Applications

Figure 4 (a) A wet roadway scene showing reflected glare on the road from overhead highway signs. (b) A computer display screen showing a reflected image of a light source, obscuring visibility of the text on the screen.

the eyes have returned to normal functioning when in fact, they may be still in the process of doing so. 2.4 Reflected glare A fourth type of glare is often described, known as reflected glare or veiling glare. It differs from disability glare in that the “veil” is entirely external to the visual system. Reflected glare is created on surfaces that are specular (or glossy) when illuminated from an angle that forms a mirror-angle among a light source, an illuminated surface, and the observer’s eyes. Essentially, an image (often a somewhat diffused or distorted image because many surfaces are not purely glossy but have semi-glossy characteristics) of a light source is visibly superimposed over a surface. The brightness of this image, like the veil from scattered light in the case of disability glare, obscures one’s ability to see objects on or along the surface. Figure 4 illustrates the phenomenon of reflected glare from overhead highway signs in a wet roadway scene and in a reflective computer display screen.

3 What Factors Affect Glare? Having described the various types of glare and their definitions in the previous chapter, the present chapter of this Spotlight provides a discussion of the characteristics of a lighting system or a lighting installation that can influence each of these forms of glare. 3.1 Light source intensity and illuminance Formally, the intensity (or luminous intensity) of a light source is characterized in units of candelas (cd) and is specific to a specific direction from the light source. In other words, the intensity of a beam of light from a flashlight is typically much

Bullough: Understanding Glare in Exterior Lighting, Display, and Related Applications

9

higher in the center of the beam (the “hot spot”), whereas the intensity in the periphery even a few degrees from the center of the beam from the same flashlight is likely to be much lower. The intensity of a light source is independent of the distance from which it is observed. For example, if the on-axis luminous intensity of a red traffic light viewed from 10 m away is 350 cd, the luminous intensity is also 350 cd when it is viewed from 100 m or even 1 km away, but the glare from the traffic light logically would be greatest for the shortest viewing distance. While the intensity does not change, what does change for these distances is the illuminance at the observer’s eyes, and this quantity rather than intensity is what affects glare. Figure 5 shows the relationships among luminous intensity, distance, and illuminance for an isotropic point source of light having a uniform intensity of 1 cd in all directions around the source. The luminous flux within 1 steradian (defined as the solid angle subtended by an area of 1 unit squared at a distance of 1 unit from the source; the units could be ft2 and ft, or m2 and m) from the point source in Fig. 5 is defined as 1 lumen (lm). The entire sphere around the source subtends a solid angle of 4π or ≈12.57 sr. For a distance of 1 ft from the source, the illuminance is defined as 1 footcandle (fc), defined as 1 lm/ft2.26 At a distance of 1 m from the source, the illuminance is 1 lux (lx), defined as 1 lm/m2. A sphere surrounding the light source with a diameter of 1 m would have a uniform illuminance of 1 lx along its interior surface.

Figure 5 Diagram showing a unit solid angle from a point source of light with an isotropic luminous intensity of 1 cd, the illuminance on the interior of the sphere is 1 fc if the sphere’s radius is 1 ft; it is 1 lx if the sphere’s radius is 1 m.

10

Bullough: Understanding Glare in Exterior Lighting, Display, and Related Applications

If the sphere had a diameter of 2 m, the illuminance on the interior surface would be 0.25 lx (1/22). This illustrates the inverse-square law whereby the illuminance from a light source (E, in lx) from a light source with a particular intensity (I, in cd) in a particular direction is inversely proportional to the square of the distance (d, in m), as expressed below:

E ¼ I∕d 2 :

(1)

As might be expected, when the illuminance from a light source is lower, it will produce less glare. With respect to disability glare, the brightness of the luminous veil of scattered light in the eye is directly proportional to the illuminance at the eyes.27 Thus, to the extent that the luminous veil’s equivalent luminance can be expressed in terms of units of cd/m2, doubling the illuminance at the eyes from a light source (by increasing its intensity or by reducing the distance between the source and the observer) would result in twice the equivalent luminance (in cd/m2) being superimposed over the retinal image for a particular scene. For light sources that subtend a relatively large area, such as a sign panel, it has also been shown that the illuminance at the eyes from the source is predictive of its contribution to disability glare. The same appears to be true for smaller sources of light. In a study of headlight glare,28 the size of the luminous area of oncoming headlights located 50 m ahead (Fig. 6) varied from 9 to 77 cm2, but all produced the same illuminance (1 lx) at observers’ eyes. Observers in that study were asked to respond to the presence of targets along the roadway scene by pressing a button while they performed a tracking task directly ahead of them, shown as the array of LEDs near the center of Fig. 6(a). As long as the illuminances from the sources were the same, the response times to press the button upon detecting the targets in the scene were nearly identical. The sets of headlights used in the study illustrated in Fig. 6 produced the same amount of disability glare.

Figure 6 (a) View of oncoming headlights in a target detection study of the impacts of glare. (b) Different headlight luminous sizes used in the study; the largest size is shown in the photograph in panel (a).

Bullough: Understanding Glare in Exterior Lighting, Display, and Related Applications

11

The illuminance from a glare source also can strongly influence the DG it elicits. Figure 7 shows the average ratings of DG (see Section 2.220) made in response to light sources producing various illuminances at the eyes, from 0.1 lx to nearly 100 lx, 29 with all other factors being equal. The numerical ratings decrease (indicating greater visual discomfort) as the illuminance from the sources increase, closely following a logarithmic function (note that the horizontal axis of Fig. 7 has a logarithmic scale). Similar logarithmic relationships between the illuminance from a light source and the subjective ratings of visual discomfort have been found in other studies of DG.30,31 The illuminance at the eyes has also been shown to be a factor in the glare recovery time. Krebs et al.32 measured response times to detect a pattern after exposure, for 1 s, to a bright yellow-green (wavelength, 550 nm) light source that produced between 0.06 and 77 lx at the observers’ eyes and found longer recovery times following exposure to the sources with higher illuminances. And as shown in Fig. 8,33 longer glare recovery times were measured following a 2.5 s exposure duration to 4 lx from a white light source, in comparison to the same duration at 2 lx. Regarding reflected glare, the illuminance from the glare source that is reflected on the surface is certainly a factor, but because this phenomenon is in large part the product of a reflected image on a screen or wet road surface, the physical size and especially the luminance of the source of light are the primary determinants of the obscuring effects of reflected glare (see Section 3.2). 3.2 Light source luminance The brightness of a light source can, as described in Section 2.1, be related to its luminous intensity and the resulting illuminance it produces at the eyes of an

Figure 7 Ratings of DG20 plotted as a function of illuminance at the eyes produced by the glare source.29

12

Bullough: Understanding Glare in Exterior Lighting, Display, and Related Applications

Figure 8 Glare recovery times following exposure for 2.5 s from an illuminance of 2 or 4 lx.33

observer, but the luminance characteristics of the source can also impact certain glare responses. For disability glare, however, the luminance distribution of the source of light producing scattered light in the eyes is relatively unimportant. Consider two sources with different luminance characteristics as illustrated in Fig. 9, one with a uniform luminance of 500 cd/m2 and one with a “checkboard” pattern having alternating cells with luminances of 0 and 1000 cd/m2, respectively. These sources would produce similar illuminances at the eyes of an observer and hence would produce similar amounts of scattered light onto the retinas of the eyes. Consider also the headlight sources illustrated previously in Fig. 6. The luminances of the headlights with luminous areas of 9, 26, and 77 cm2 were 1.4 million, 480,000, and 160,000 cd/m2, respectively, differing by

Figure 9 (a) Representation of a diffuse source with a uniform luminance of 500 cd/m2. (b) Representation of a nonuniform source with alternating regions having luminances of 0 and 1000 cd/m2. Both sources would produce similar illuminances at the eye.

Bullough: Understanding Glare in Exterior Lighting, Display, and Related Applications

13

as much as 9:1. Yet these sources all produced a similar degree of disability glare,28 resulting in similar impacts on target detection times. With respect to DG, the role of the source’s luminance characteristics is somewhat more complex. An important factor that helps to determine whether and the extent to which a light source’s luminance might influence visual discomfort is its size. One way to characterize an object’s visual size is by the angle it subtends from the observer’s viewing location. Returning again to the study of headlight glare illustrated in Fig. 6,28 the three headlights in that study had diameters of ∼3, 6, and 10 cm, from smallest to largest. At a distance of 50 m, they subtended 0.02 deg, 0.06 deg, and 0.1 deg, respectively. Despite their large differences in luminance (160,000, 480,000 and 1.4 million cd/m2), there were no statistically significant differences among these differently sized headlights when they were rated by observers using the De Boer (DB)20 rating scale for DG, while they performed the tracking task shown in Fig. 6(a). In comparison, a different study of DG34 was carried out using arrays of LED sources (with diameters of 5 cm), where the arrays were either viewed as bare LEDs, or behind circular diffusers located either 7 or 21 cm in front of the arrays. The arrays (Fig. 10) were viewed from a distance of ∼5 m in an otherwise dark room. The maximum luminances of the arrays with the diffusers were 15,000 and 50,000 cd/m2, and the maximum luminance of the array with no diffuser was 1,000,000 cd/m2. All of the arrays produced an illuminance of 2 lx at the eyes of observers in the study, but unlike the headlights in Fig. 6, the light sources in Fig. 10 elicited different ratings of visual comfort. The source with the lowest maximum luminance (15,000 cd/m2) resulted in the highest numerical rating values (the least amount of discomfort) and the source with the highest maximum luminance (1,000,000 cd/m2) elicited the lowest numerical ratings (the greatest amount of discomfort), as illustrated in Fig. 11. An important difference between the sources in each of these studies is their sizes. In the former study,28 the sources subtended no more than 0.1 deg at the eyes, whereas in the latter study34 the sources subtended at least 0.6 deg at the eyes. It would seem that the maximum luminance has an influence on DG for light sources that are relatively “large,”

Figure 10 (a) View of an LED array diffused to a maximum luminance of 15,000 cd/m2. (b) View of an LED array diffused to a maximum luminance of 50,000 cd/m2. (c) View of an LED array without a diffuser; the maximum luminance is 1,000,000 cd/m2.

14

Bullough: Understanding Glare in Exterior Lighting, Display, and Related Applications

Figure 11 DG ratings for LED arrays varying in maximum luminance. All arrays produced the same illuminance of 2 lx at the observers’ eyes.

whereas when the light sources are relatively “small,” only the illuminance from the source matters. In a similar follow-up study, 35 a single bare LED (with a maximum luminance of 1,000,000 cd/m2) in the array illustrated in Fig. 10(c) was shown to observers, as well as a single LED diffused to have a lower maximum luminance of 200,000 cd/m2. Each LED was viewed from a distance at which it produced 1 lx at the observers’ eyes, with the bare LED subtending 0.03 deg of visual angle and the diffused LED subtending 0.2 deg. Unlike the sources in Fig. 10, DG rating values in response to these single LEDs were nearly identical despite their different luminances (differing by a factor of 5). Thus, sources with visual sizes of 0.1 deg and 0.2 deg seem to be “small” whereas those subtending 0.6 deg seem to be “large.” The precise breakpoint between “small” and “large” was investigated in a study by Rosenhahn and Lampen36 who found that when a light source subtended a visual angle of no more than 0.2 deg, it was “small” and the amount of DG it elicited was dependent only upon the illuminance it produced at the eyes. When the source size was greater than 0.2 deg in terms of visual angle, the maximum luminance also influenced the amount of DG people experienced. The impact of the light source luminance on glare recovery has not been investigated independently of illuminance. For example, Irikura et al.37 measured recovery times following exposure to light sources varying in luminance between 300 and 3300 cd/m2, which were presented along the line of sight. They found longer recovery times in response to the highest luminances. Because the size of the light source did not change in this study, the illuminance from the source at observers’ eyes was directly proportional to the source’s luminance. As described in Section 2.1, recovery times increase with increasing illuminance from the glare source. .

Bullough: Understanding Glare in Exterior Lighting, Display, and Related Applications

15

Reflected glare is strongly influenced by the luminance of the glare source. In a study to assess the impacts of reflected glare in displays,38 ratings of disturbance from a light source visible in the reflected screen were always consistently greater when the luminance of the source was 850 cd/m2 and lower when the luminance was 200 cd/m2. 3.3 Background luminance Lighted devices often have dimmers and brightness controls because what is necessary for visibility under high ambient lighting levels (producing high background luminances) may be uncomfortable when the overall lighting level (and hence, the background luminance) is much lower. A simple example from the area of transportation is the appearance of vehicle headlights during daytime and nighttime. During the daytime, vehicle headlights are not usually considered to be major contributors to glare, but the same headlights with the same intensity and luminance characteristics are often quite glaring when viewed at night, when the background luminance is substantially lower. The impact of the background luminance on disability glare can be understood by considering the impact of a luminous “veil” on the contrast between an object and its background at two different light levels. Suppose a corridor contains a “wall pack” light fixture (luminaire) at one end of the corridor that is switched on 24 h per day and produces an illuminance of 10 lux at a particular point on the floor, and suppose the corridor also contains windows that provide 1000 lx of illumination at the same location on the floor. Suppose also that the floor is made of worn concrete with a reflectance (ρf) of 0.239 and a black tool bag with a reflectance (ρt) of 0.1 has been dropped on the floor in that location. Assuming an object is matte and not glossy, its luminance (L, in cd/m2) can be estimated from its reflectance (ρ) and from the illuminance (E, in lx) falling on it according to the following equation:

L ¼ ρE∕π:

(2)

Under daytime viewing conditions (with a total illuminance of 1010 lx), the floor would have a luminance (Lf) of ∼64 cd/m2, and the tool bag would have a luminance (Lt) of ∼32 cd/m2, and the resulting contrast (C) would be calculated as follows:

C ¼ jLf − Lt j∕ maxðLf , Lt Þ ¼ j64 − 32j∕64 ¼ 0.5:

(3)

Similarly, the luminances of the floor (Lf) and tool bag (Lt) would be ∼0.64 and 0.32 cd/m2, respectively, at night under the illumination from only the wall pack luminaire (10 lx), using Eq. (2). Using these values of Lf and Lt in Eq. (3) would also result in a contrast (C) of 0.5.

16

Bullough: Understanding Glare in Exterior Lighting, Display, and Related Applications

Suppose now that the scattered light in the eye from the wall pack luminaire produces a luminous veil with an equivalent luminance of 0.3 cd/m2. The values of Lf and Lt will increase to 64.3 and 32.3 cd/m2 for the daytime condition and will increase to 0.94 and 0.62 cd/m 2 for the nighttime viewing condition. Inserting these values into Eq. (3), the contrast (C) for the daytime conditions is virtually unchanged (C = 0.498 ≈ 0.5), but for the nighttime conditions the contrast is reduced substantially (C = 0.34). Thus, while the equivalent luminance of the luminous veil produced by a light source does not change in different ambient environments, it will clearly have a greater impact on visibility (i.e., contrast) when the light levels are lower. The impacts of different background luminances on DG have been investigated for a variety of contexts. One example is signage and display applications, where panels containing organic light-emitting diodes (OLEDs) are becoming an increasingly common technology. 40 An investigation of the DG from OLED panels 41 was carried out under a range of background luminances between 1 cd/m 2 , representing a typical background luminance for nighttime scenes outdoors, and 215 cd/m2, representative of a brightly lighted interior wall. The luminances of the OLED panels varied from 500 to 7000 cd/m2. Figure 12 shows the DG ratings for each combination of OLED panel luminance and background luminance. It can be seen that for each OLED panel luminance, the numerical rating values are lower (indicating greater discomfort) when the background luminance is lowest, and higher (less discomfort) when the background luminance is highest. The same panel will be rated as more or less glaring depending upon the luminance of the background surrounding it. In the previously mentioned study that investigated the influence of the maximum luminance of LED arrays (with and without diffusers) on DG,34 the impact

Figure 12 Ratings of DG in response to uniform OLED panels varying in luminance, against three background luminances.

Bullough: Understanding Glare in Exterior Lighting, Display, and Related Applications

17

Figure 13 (a) LED arrays viewed in an otherwise dark environment. (b) LED arrays adjacent to an illuminated surface serving as part of the background.

of illuminating a surface adjacent to the LED arrays was also investigated. As illustrated in Fig. 13, the adjacent surface, which had a luminance of 45 cd/m2, contrasted with the darkness (luminance of 0.2 cd/m2) surrounding of the arrays when the surface was not lighted. For each LED array, the numerical DG ratings were consistently about 0.5 units higher when the illuminated background was present, compared with when it was not (the ratings when no illuminated background was present are shown in Fig. 11; see Section 3.2). Findings consistent with these examples have been published in other reports as well.42,43 The background luminance can also affect the glare recovery time following exposure to a bright light. In their study of glare recovery times, Irikura et al.37 assessed responses to lights presented against background luminances of 0.1, 0.3, or 1 cd/m2. The recovery times for people to be able to see after exposure to bright light were shortest for the highest background luminance of 1 cd/m2 and longest for the lowest background luminance of 0.1 cd/m2. In some sense, this is evident from inspecting Fig. 3 (see Section 2.3). The dark adaptation process takes longer, the lower the dark adaptation state that is eventually reached. The background luminance also plays a role in the perception of reflected glare. Early computer screens typically had negative display polarity (characterized by light characters against a dark background). As computer screens became more ubiquitous in the workplace, people experienced visible reflections in their screens more frequently. Positive display polarity screens with higher background luminances and with dark characters are more resistant to reflections.44 3.4 Light source location The location of a light source in the field of view can impact the degree to which it could cause glare. When considering disability glare, it has been shown27 that scattered light from a light source in the visual field is not uniform throughout the retinal image. Rather, the equivalent luminance is highest very close to the

18

Bullough: Understanding Glare in Exterior Lighting, Display, and Related Applications

Figure 14 Equivalent veiling luminance for different angular distances from a light source producing 3 lx at an observer’s eyes. (Figure derived from the veiling luminance equation in Ref. 27.)

light source (in terms of visual angle from the observer’s point of view) and drops off rather rapidly as the angular distance from the light source increases. Figure 14 illustrates the rapid reduction in the equivalent veiling luminance from a light source that produces 3 lx at the eyes of an observer based on an analysis of scattered light in the eye carried out by Fry.27 Farther than about 10 deg from the light source, the veiling luminance is relatively low and gradually decays at even larger angles. At locations closer than 5 deg from the light source, the veiling luminance escalates very rapidly. As a result of this effect, objects very close to a glare source, even a glare source that may not be especially bright, can be difficult to see, especially when the ambient light level is low. Understanding the influences of light source location on DG is somewhat more complicated than for disability glare, in part because evaluating the equivalent luminance from scattered light in the eye is ultimately a physical measurement process. In comparison, there is not yet a firm explanatory mechanism that has been validated for DG. Indeed, attempts to link DG to physiological responses such as pupil size fluctuations45,46 or electromyography (electrical measurements of muscle tension)47 have not been successful, leaving subjective ratings20 as the state-of-the-art in measuring DG. Another complication in defining the impacts of light source location in DG has to do with the natural inclination of an observer to look toward a bright light in the field of view,48 even temporarily. Certainly, it has been shown that when an observer fixes their gaze at a specific point in the field of view, DG is reduced when a light source is moved further away from their line of sight, even if the illuminance at the eyes is kept constant.49 This situation perhaps is applicable

Bullough: Understanding Glare in Exterior Lighting, Display, and Related Applications

19

when driving along a highway where the line of sight is maintained along a fairly narrow region along the road ahead. But when driving in an urban location with the potential for pedestrians and other hazards to be located throughout the scene (observe for example Fig. 2 in Section 2.2), or for pedestrians whose direction of gaze fluctuates more widely, it is likely that the observer will eventually, even if briefly, view a possible glare source directly. Some studies of DG have modeled the angular distance effect in DG by a term that includes the reciprocal of the square of the angle.50 However, when one is viewing a glare source directly, the angle between the line of sight and the source is 0 deg, resulting in an expression that trends toward infinity! Clearly this would be a nonsensical result, and further, it does not match what happens in practice. In one study designed to compare a fixed-gaze (fixating on a location 5 deg from a glare source) versus a free-gaze (looking generally toward and possibly at the glare source) viewing protocol,43 the DG ratings were actually very similar (Fig. 15). Possibly, subjects glanced at the glare source even when instructed to look at a point 5 deg away from it. At any rate, since observers are likely to gaze directly at bright sources in a scene, it seems reasonable to allow them to do so in experiments of DG, except for specific tasks where the gaze must be held in a specific location. The effect of the light source location on glare recovery times, if any, are not well understood. In general, when a glare source was used to evaluate the time needed to regain visual functioning after exposure, the source was located near the center of the line of sight.32,37,51 Presumably, if a glare source were presented in the peripheral field of vision, it would have a smaller impact on glare recovery. The location of a light source is critical in whether it will produce reflected glare. As illustrated in Fig. 16, the geometric relationships among the locations of the light source, the surface or object being viewed, and the observer’s eyes

Figure 15 (a) DG ratings in response to halogen and HID headlights against high and low background luminances when observers fixated 5 deg from the location of the glare source. (b) Discomfort ratings when observers could view the glare source directly and shift their gaze freely throughout the visual scene.

20

Bullough: Understanding Glare in Exterior Lighting, Display, and Related Applications

Figure 16 Illustration of the mirror angle where an image of a light source at point P will appear at point O when the observer is at point Q. Source, Wikimedia Commons (Creator: Arvelius; Creative Commons License CC BY-SA 3.0).

will determine where reflections can occur and whether they will obscure the visual task the observer is performing. While not strictly a question of the location of a potential glare source, questions about the number of sources and the simultaneous impacts of multiple light sources in various locations in the field of view have also been investigated, at least for disability and DG. Because disability glare is essentially an accounting of scattered light in the eye from a source,27 it is a simple matter to treat multiple sources individually and add the resulting equivalent veiling luminances (for each source’s angular distance from the line of sight) together. For DG, Schmidt-Clausen and Bindels 31 found that when two glare sources were separated in the visual field by an angle no larger than 5 deg, they could be treated as equivalent to a single source producing the sum of the illuminances that each source produced individually at an observer’s eyes. However, Bullough et al.29 found that DG responses to one light source located on the line of sight was not affected by another light source if it was located at least 9 deg from the line of sight. Possibly, some angular separation between 5 deg and 9 deg could serve as a boundary between glare sources whose discomforting effects can be “merged” and those that can be treated independently of other sources in the visual field.

Bullough: Understanding Glare in Exterior Lighting, Display, and Related Applications

21

3.5 Light source spectral distribution In recent years, the spectral or color characteristics of lighting have received a great deal of attention among the driving public in terms of their impacts on glare. As described earlier (see Section 1), as lighting technologies have evolved and sources such as LEDs, which often (but do not always) have a “bluer” color appearance than earlier light sources such as incandescent and halogen filament sources, which appear “yellower” in color. This color or spectral difference is often more evident than other differences that may exacerbate glare, and this may explain why spectral differences have received so much attention. The role of spectrum in disability glare has been investigated in numerous studies. Some of these studies have focused on the practical implications of the transition in vehicle headlighting technologies when some vehicle models were beginning to use HID headlights instead of halogen filament sources. Figure 17 shows the relative spectral distributions of these two types of light sources as used in vehicle headlights. It can be seen from Fig. 17 that the halogen source is weighted toward longer wavelengths, consistent with its yellower color appearance, whereas the HID source in Fig. 17 has a more uniform distribution of wavelengths across the visible spectrum, appearing whiter or even bluer than the halogen source. (Both light sources are still considered to be “white” in color appearance, despite their differences.) Flannagan et al.52 found that in the presence of the same amount of illumination from a halogen or an HID light source with spectra similar to those in Fig. 17, the ability to see a simulated pedestrian target was the same under each source. Similar results were found by others.28,49 Of course, both halogen and HID light sources produce white light, so it could be argued that the spectral

Figure 17 Relative spectral distributions of halogen and HID sources used in automotive headlighting systems.

22

Bullough: Understanding Glare in Exterior Lighting, Display, and Related Applications

differences between them are not large enough to be able to measure a reliable difference between them in terms of disability glare. Steen et al.53 measured the differences in disability glare from light filtered by a short-wavelength pass and by a long-wavelength pass filter, resulting in blue and red colored light. When the resulting illumination was matched for photopic intensity, the two sources resulted in an equivalent amount of visual disability. A recent investigation of red and blue LED sources like those used for emergency vehicle flashing lights was carried out,54 where a scale model police vehicle (Fig. 18) was fitted with red or blue flashing LEDs and observers had to identify whether a scale model police officer figure was positioned to the left or right of the vehicle. It was found that increased intensity of the flashing lights resulted in poorer performance, but there was no difference in performance in the presence of red or blue lights; they made the police officer equally difficult to see. When the same lights in Fig. 18 were rated by observers in terms of the discomfort they elicited, the blue LED lights were judged as substantially more uncomfortable than the red ones (Fig. 19), even though they were matched for luminous intensity and had equivalent impacts on disability glare.54 These results provide strong evidence that disability and DG are truly independent phenomena, likely influenced by different channels in the visual system.55 Because red and blue lights are dominated by wavelengths at either end of the visible spectrum (long wavelengths for red light; short wavelengths for blue light), results such as those in Fig. 19 do not elucidate the visual mechanisms underlying the spectral sensitivity for DG. To lay a foundation for such a spectral sensitivity function, Bullough56 reported on an experiment to measure discomfort from narrowband light at different wavelengths (from 450 to 700 nm) throughout the visible spectrum and producing a range of illuminances (0.05, 1.3, or 2.6 lx) at the eyes of observers. The light sources were viewed from visual angles of 5 deg and 10 deg from the line of sight.

Figure 18 Photograph of scale model police cars with flashing LED lights having the same intensity but differing in color,54 blue on the left and red on the right. (Apparent brightness differences may reflect different sensitivity to blue/red light by the camera used to take the photographs.)

Bullough: Understanding Glare in Exterior Lighting, Display, and Related Applications

23

Figure 19 Ratings of DG to scale model flashing lights using red and blue LEDs matched for luminous intensity. Lower numerical ratings correspond to greater discomfort.20

As shown in Fig. 20(a) (which shows the results for sources located 5 deg from the line of sight), data for the shortest wavelengths, especially 450 nm, were offset (having lower numerical DG ratings20) from the values for the stimuli at

Figure 20 (a) DG ratings for stimuli presented 5 deg off-axis and varying in wavelength and illuminance. (b) Ratings for the same stimuli, plotted as a function of illuminance quantities based on the V DG (λ) spectral sensitivity function 56 combining photopic and S cone sensitivity.

24

Bullough: Understanding Glare in Exterior Lighting, Display, and Related Applications

other wavelengths, suggesting that the photopic illuminance is not an ideal metric for characterizing the spectral sensitivity for DG. Nor was large-field photopic illuminance,57 which has increased short-wavelength sensitivity based on the sensitivity of the peripheral retina. Rather, as shown in Fig. 20(b), quantities based on a spectral sensitivity function derived from adding the photopic large-field function to a specific proportion of the spectral sensitivity for shortwavelength (S) cones (peaking at 440 nm) were able to rectify the DG responses in this experiment. This spectral sensitivity function was named VDG (λ), the subscript DG referring to DG. The spectral sensitivity function VDG(λ) that was able to rectify DG responses in Fig. 20 is defined as follows:

V DG ðλÞ ¼ V 10 ðλÞ þ k · SðλÞ:

(4)

In Eq. (4), V10(λ) is the large-field photopic luminous efficiency function,57 S(λ) is a luminous efficiency function based on the short-wavelength cone fundamental defined by Smith and Pokorny,58 and k is a constant that equals 0.18 for an angular distance of 5 deg from the line of sight (in multiple experiments at 5 deg, k ranged from 0.15 to 0.19), and 0.72 for an angular distance of 10 deg from the line of sight;57 see Fig. 21. The increase in short-wavelength sensitivity for the peripheral retina is consistent with similar increases in short-wavelength spectral sensitivity that have been found in other studies of visual responses to light.59,60 The stimuli used to derive the VDG(λ) spectral sensitivity functions shown in Fig. 21, as described above, were narrowband sources each centered around a particular wavelength. Yet, many of the light sources that might serve as sources of discomfort in real-world conditions are broadband, “white” sources of illumination, such as headlights (see Fig. 17), streetlights, and interior lighting systems. Does the VDG(λ) function apply to those as well? This was tested in a study of LED aviation signal lights61 varying in correlated color temperature (CCT) from 2700 K (“warm” white) to 5900 K (“cool” white). (K refers to kelvins, a unit of

Figure 21 (a) Spectral sensitivity for DG from sources located 5 deg off-axis. (b) Spectral sensitivity for DG from sources located 10 deg off-axis.

Bullough: Understanding Glare in Exterior Lighting, Display, and Related Applications

25

temperature, and CCT is the temperature of a blackbody radiator most similar to a light source of a particular CCT.) The appearance of white signal lights having a luminous intensity of 10,000 cd viewed from distances of 30 to 180 m (producing illuminances at observers’ eyes between 0.3 and 10 lx) was simulated in a scale model display scene, and observers rated their level of visual discomfort on the DB20 scale. The observers in this study were not constrained to viewing the glare sources from a particular viewing angle; rather they could look toward the lights as they felt they needed to, in order to make a DG judgment. The 5900 K CCT source produced more DG (resulting in lower numerical ratings20) than the 2700 K source when they produced the same illuminance at the eyes as shown in Fig. 22. Further, the rating values were found to be best rectified by the VDG(λ) function with a value for k of 0.72, even though the observers did not look at a fixed location 10 deg from the glare source in this experiment but could look freely wherever they wished.61 The results do indicate, however, that DG does exhibit increased short-wavelength sensitivity for broadband “white” light sources just as it did for narrowband colored light sources. The spectral distribution of the background surrounding a glare source has also been investigated in terms of its impacts on visual discomfort. Sweater-Hickcox et al.62 evaluated the effects of white, yellow, or blue luminous backgrounds behind an array of white LED sources. The white LED sources produced between 4 and 12 lx at the observers’ eyes, and the luminous background produced between 1 and 3 lx at their eyes. The relative contribution of the luminous background ranged between 5% and 20% of the total

Figure 22 Ratings of DG to light sources varying in CCT and in the illuminance they produced at observers’ eyes.61

26

Bullough: Understanding Glare in Exterior Lighting, Display, and Related Applications

illuminance at observers’ eyes. As would be expected from the previously described impacts of a luminous background on DG (Section 3.3), the backgrounds of all three colors (white, yellow, and blue) reduced visual discomfort, but blue reduced discomfort less than the other two colors.62 Sweater-Hickcox et al.62 speculated that the short-wavelength light from the blue surround may have actually contributed to DG in addition to mitigating the glare from the white LED array. Another study of the spectral content of a luminous background and its influence on DG was made by Nagare and Bullough,63 where the glare source produced 2 lx at observers’ eyes and the background simulated a light level typical of an outdoor scene such as a roadway at night, with a luminance of 0.44 cd/m2 and producing 0.017 lx at the eyes, less than 1% of the total illuminance at the eyes. The luminous background was either “warm” white with a CCT of 3000 K or “cool” white with a CCT of 6500 K. At this ambient light level, the dim background was substantially brighter in appearance when the CCT was higher than when it was lower. Under these conditions, in contrast with the findings from Sweater-Hickcox et al.,62 the CCT of the background had no influence on the DG ratings. A possible explanation for the difference is the much lower overall contribution of the luminous background in the latter study63 compared with the former one.62 Because glare recovery is often critical under low light levels, when the visual system is adapted to relatively dark conditions, and because the visual system exhibits a Purkinje shift6 toward short visible wavelengths as the light level is reduced and rod photoreceptors contribute to vision in greater proportion (see Section 2.3), questions about the possible influence of light source spectrum on glare recovery have been made. Not all of these studies have yielded consistent results, however. Some studies carried out in the 1960s, at a time when the country of France required vehicles to have yellow-tinted headlights, suggested that recovery times following exposure to white headlights were shorter than after people were exposed to yellow headlights,64 while others found similar recovery times for both headlight colors.65 More recently as LED sources for streetlighting have become more commonplace, comparisons of recovery times following exposures to LEDs varying in CCT between 1900 and 5000 K were made,66 and there was a systematic increase in the necessary time to recover dark adaptation as the CCT increased. One possible explanation for the apparent contradictions in the literature could be the nature of the visual tasks used in different experiments to evaluate glare recovery times. In a recent experiment,67 the impact of LED glare sources having CCTs of 3000 and 6500 K was investigated. Each glare source was exposed to observers, producing either 0.5 or 1 lx at the eyes. After the source was extinguished, the subjects were asked to identify the orientation of a small low-contrast target in the center of a display close to the location of the glare source. In this study, greater illuminances resulted in longer glare recovery times,

Bullough: Understanding Glare in Exterior Lighting, Display, and Related Applications

27

Figure 23 Influence of glare source CCT on recovery times for (a) a fixed-location visual recovery task and (b) a variable-location visual recovery task.

but there were no differences in recovery time between the source CCTs of 3000 or 6500 K, as shown in Fig. 23(a). A follow-up experiment was carried out with a similar methodology,51 but the glare recovery task involved detecting a small target located either above, below, to the left, or to the right of the location of the glare source. For this task, there were reliably longer recovery times when the CCT of the glare source was 6500 K as illustrated in Fig. 23(b). Possibly, the latter recovery task involved peripheral vision to a much greater extent than the former task where the target location was fixed. At low light levels, the visual system’s dependence on rod photoreceptors for vision results in a spectral sensitivity to shorter wavelengths than at higher levels because visual sensitivity shifts from cone-dominated (with a photopic spectral sensitivity peaking near 555 nm) toward rod-dominated (with a scotopic spectral sensitivity peaking near 507 nm).6 This shift impacts peripheral vision only68,69 because the central retina (fovea) is populated only by cones and no rods. There do not seem to have been any systematic investigations of the spectral content of reflected glare on the perception of visual tasks. Reflected glare from “white” light sources tends to wash out colors of visual tasks where it appears70 and can negatively impact the performance of tasks that require fine color discrimination. 3.6 Glare exposure duration The length of time that one is exposed to a source of glare can impact the degree of glare that is experienced under certain circumstances. With regard to disability glare, because the scattered light within the eye is a physical phenomenon, the equivalent luminance of the luminous “veil” that a light source produces is independent of the length of time it is present. Exposure duration therefore has little influence on disability glare as it is characterized. DG, on the other hand, does appear to exhibit some influence from the duration of exposure. Sivak et al.71 measured DG in observers who were exposed to

28

Bullough: Understanding Glare in Exterior Lighting, Display, and Related Applications

stimuli ranging in illuminance at the eyes from 0.5 to 8 lx and having durations of 0.125 to 2 s. In addition to finding the expected relationship between DG and the light source’s illuminance (see Section 3.1), they found that duration also exerted an effect on DG, with the longer-duration stimuli being reported as more uncomfortable. Overall, ratings of discomfort were found to be correlated with the “dose” of light exposure, defined by the product of illuminance (lx) and duration (s) and having units of lx·s. However, in a study by Van Derlofske et al.33 in which DG ratings to stimuli varying in illuminance (from 1 to 4 lx at observers’ eyes) and in duration (from 2.5 to 5 s) were assessed, the duration of exposure did not impact the DG ratings. Possibly, the duration matters in terms of DG up to a duration of about 2 to 2.5 s after which it no longer impacts visual discomfort, but this would require further testing to verify. Glare recovery is the response that appears to have the strongest relationship with the duration of exposure to a light source. Irikura et al.,37 Van Derlofske et al.,33 and Skinner and Bullough25 all found that recovery times following glare exposure were strongly correlated with the dose as defined in the preceding paragraph (i.e., illuminance × duration). Interestingly, this relationship with the glare dose also seems to interact with the recovery task (on- or off-axis) and the spectral distribution of the glare source51 as described in Section 3.5. Figure 24 shows recovery times following exposure to sources with different CCTs (3000 and 6500 K), and it illustrates the relationship between recovery and the glare dose, albeit a different one for each glare source CCT. Like disability glare, reflected glare is a physical phenomenon, and further takes place entirely within the realm of physics, occurring outside the eye. Consequently, reflected glare is present for the same duration that the offending light source is present.

Figure 24 Glare recovery times for an off-axis visual task as a function of the glare dose for two glare source CCT values.51

Bullough: Understanding Glare in Exterior Lighting, Display, and Related Applications

29

3.7 Observer age and other characteristics The present and subsequent sections of this chapter move away from the characteristics of the potential glare source and the luminous environment in which it is located, and toward some of the characteristics of the observers themselves. As our population grows increasingly older, age becomes a more relevant parameter for lighting applications. Age strongly influences visual performance because the lens of the eye grows thicker and yellows throughout the life span and because the pupil size tends to grow smaller with age.6 Therefore, a reduction in contrast is more likely to affect an older individual than a younger one, in terms of visual performance. There is also evidence that due to the thickening of the lens it grows less transparent and scatters more light in older adults than in young people.72 As a consequence of this, the equivalent luminance of the luminous “veil” that is used to quantify disability glare would be expected to be larger in older than in younger populations. The age of the observer has been incorporated as a parameter in some glare evaluation formulae.73 It should be noted, though, that the amount of individual variation in scattered light for a particular age can be as large as the variation that can exist between individuals differing by nearly 50 years of age!72 Thus, while it is not incorrect to say that the older population may be, on average, more affected by disability glare compared with the younger population, the value of applying an age-related correction to any individual or small population of individuals is questionable. In a similar way, there have been some measurements undertaken to show that eye color can affect the amount of light scattered within the eye, with darker iris colors associated with less scattered light, and lighter blue eyes associate with more. These factors have also been included in some equations for equivalent veiling luminance,73 although the influence of eye color is small compared with the illuminance produced by a glare source at an observer’s eyes. Studies of age and DG have also been carried out, with wide variations in their results. In the context of nighttime driving, which is one of the most challenging for DG, a small increase in DG in older individuals has sometimes been found.30 However, a small, nonsignificant effect in the opposite direction has also been found in a difference study of DG from streetlights.74 Comparing discomfort to light sources varying in illuminance and spectral distribution in individuals averaging 30 years old in one case and 60 years old in another case, another study found nearly identical average ratings of DG for each age group.75 As with disability glare, age-related correction for DG does not seem to reduce uncertainties in an especially helpful way. Recovery glare is perhaps the glare response that most consistently and clearly demonstrates a strong influence of age, with older individuals having longer glare recovery times than younger ones in response to the same stimulus. Reading64 found a very strong correlation between observer age and recovery times. In a study of visual recovery following exposure to headlight glare conditions,25 older observers with an average age over 50 had significantly longer recovery times than

30

Bullough: Understanding Glare in Exterior Lighting, Display, and Related Applications

Figure 25 Average glare recovery times following headlight glare for observers in two age groups (50 years).

younger observers with an average age under 30 (Fig. 25). And Schieber76 similarly found an age-related trend with longer recovery times for older individuals in response to glare than for middle-aged and younger individuals. Clearly, older individuals are more susceptible to longer glare recovery. Reflected glare is not specifically impacted by the observer’s age, but as with disability glare, contrast reductions caused by reflected light source images on a surface will have a greater impact on the visual performance of an older person than the same reduction in contrast would have in someone who is younger. 3.8 Psychological factors in discomfort glare Most of the preceding sections discuss factors that are in some way quantifiable and have a connection to the photometric and colorimetric properties of a lighting installation or display. In the immediately preceding section, we have discussed age as an observer-based factor, which is still quantifiable in terms of its impacts on glare because of optical changes in the visual system that occur over years. In general, these kinds of factors are sufficient to characterize many glare-related effects and responses, such as disability glare, glare recovery, and reflected glare. However, DG is unique in that not only those measurable factors about the luminous environment, stimuli, and observer influence visual discomfort but so do others that can often only be defined qualitatively at best, yet can exert a substantial influence on DG. These are psychological factors that may relate to the type or difficulty of task an observer is performing when the glare source is present or to the aesthetics of the glare source. Sivak et al.77 were among the first researchers to investigate whether, and how, the nature of the visual task being performed by an observer might affect

Bullough: Understanding Glare in Exterior Lighting, Display, and Related Applications

31

that observer’s sensation of DG from a light source in the field of view. They asked observers to identify the orientation of a small gap in an image of a square—either on the top, bottom, left, or right side. They did this while in the presence of a light sufficient in intensity to cause visual discomfort, and they modulated the size of the gap to make it easier or more difficult to detect. They found systematically lower numerical rating values (indicating greater discomfort) in response to the glare source when the gap size was smaller. This suggests that when a visual task is more difficult, people will tend to judge the impact of a glare source as leading to more visual discomfort. These findings were replicated in other investigations, as well. Theeuwes et al.78 mounted a glare source simulating the appearance of distant vehicle headlights onto a vehicle and could adjust the illuminance it produced at drivers’ eyes while they drove along a predefined route along public roadways. The same glare illuminance was rated as producing more discomfort when the drivers were on narrow, winding roads than when they were on wide, straight, and flat roads.78 Van Derlofske et al.28 measured DG from a set of facing headlights located 50 m ahead while observers were performing a target detection task with peripheral targets varying in reflectance (from 0.2 to 0.4). Observers pressed a button as soon as they could detect a target in their peripheral field of view and then rated the level of DG they experienced during each experimental trial. Ratings decreased (indicating greater discomfort) as the amount of illuminance produced by the headlights at observers’ eyes increased, but they also did so when the reflectance (and the contrast) of the target was lower, making it harder to see,28 as shown in Fig. 26.

Figure 26 DG ratings as a function of the illuminance from facing headlights, for high- and low-contrast targets.

32

Bullough: Understanding Glare in Exterior Lighting, Display, and Related Applications

The perceived aesthetics of a glare source can also influence the degree to which observers might experience DG. The degree of visual discomfort from windows with different view characteristics was investigated.79 Controlling for the light levels produced by the windows, they either were diffused to eliminate the view through them completely, had a mundane view of a concrete block wall, or had a view of a natural park scene. The visual interest of these scenes were rated in a separate experiment to increase from no view, to the view of the wall, to the natural view. People’s responses to the views showed that they experienced less DG when there was an attractive natural view, and they experienced the most DG when the view was obscured by the diffusers, even if all of the windows produced the same light levels at observers’ eyes.79 This demonstrates that people are more able to tolerate DG if the environment is more attractive or interesting.

4 Modeling and Predicting Glare In the present chapter of this Spotlight, a few models of glare are briefly described. These models can be used by lighting specifiers and optical engineers to help make predictions of various aspects of glare that might be experienced by individuals in different lighting applications. Not all glare aspects have been extensively modeled. 4.1 Disability glare modeling There have been several equations published to characterize the properties of the scattered light in the eye that acts like a luminous veil with a particular equivalent luminance. The primary factors that influence disability glare are the illuminance it produces at observers’ eyes and the angular distance between the source of glare and the observer’s line of sight. As described in the previous chapter, some formulae that can be used to calculate the equivalent veiling luminance produced by scattered light in the eye include factors such as the age of the observer and the observer’s eye color.73 However, the inherent variability among individuals makes those formulae less useful from a practical standpoint. A simple equation for the equivalent veiling luminance (Lv, in cd/m2) from by a light source that produces a particular illuminance (E, in lx) at the eyes, and which is located at a particular angular distance (θ, in degrees) from the line of sight is as follows:

Lv ¼ 9.2E∕½θðθ þ 1.5Þ:

(5)

Equation (5) should be used for each potential source of glare in the lighted environment; the equivalent veiling luminances from each source can be added together to estimate the overall equivalent veiling luminance that would be experienced. The equivalent veiling luminance is added to the luminances of both the object to be seen and to the immediate background to adjust its contrast so it can be evaluated in terms of visibility.

Bullough: Understanding Glare in Exterior Lighting, Display, and Related Applications

33

4.2 Discomfort glare modeling In their study of the factors (e.g., illuminance, maximum luminance, illuminance from the immediate background and ambient environment, number of light sources) that affect DG, Bullough et al.29,34,35 developed a mathematical model to predict the numerical DG rating on the DB scale20 from a particular light source. The model uses four input variables: l l l l

The illuminance from the light source (EL, in lx); The illuminance from the area immediately surrounding the source (ES, in lx); The illuminance from the ambient environment (EA, in lx); The maximum luminance of the light source (LL, in cd/m2).

The model is based on outdoor viewing conditions and assumes that the observer is viewing the glare source directly, so there is no term related to its angular location from the line of sight. The model is for a single source being viewed from a specific location; the same light source could produce a different amount of DG to an observer in a different location. The total illuminance (ET) at the observer’s eyes (in lx) at the location of interest is the sum of EL, ES, and EA. The definition of the illuminance from the surround (ES) is somewhat imprecise; based on the additivity of glare sources that are located within 5 deg of each other,31 and the independence of sources that are 9 deg from each other,29 a radius of 7 deg (splitting the difference between these two values) is suggested as an approximation of the immediate surround. The ambient illuminance (EA) is defined as the amount of light coming from other sources in the general vicinity of the observer, usually from properties other than the one on which the glare source is located. As an alternative, especially in situations where measurements of the ambient illuminance cannot be made in the field, suggestions for ambient illuminances for different types of areas are suggested below29 based on field measurements from Li et al.:80 l l l

Urban area: EA = 2 lx Suburban area: EA = 0.2 lx Rural area: EA = 0.02 lx

The quantities can be measured in an existing lighting installation. 81 To estimate EL, orient an illuminance meter toward the suspect light source (when the installation is switched on) to measure the total illuminance (ET) at a particular location. Using a small circular baffle to shield the illuminance meter from the light source in question, measure the illuminance at the same point and in the same direction. A 5-cm diameter baffle located about a meter in front of the illuminance meter is effective at blocking the direct illuminance from the light source. Subtract the shielded illuminance from the total illuminance to estimate EL. If it is possible to switch off the lighting installation (including the suspect

34

Bullough: Understanding Glare in Exterior Lighting, Display, and Related Applications

light source), do so and measure the ambient illuminance (EA) from the same location and orientation. If it is not possible to switch off the lighting installation, estimate E A from the bulleted list above. Subtracting E A from the shielded illuminance made with the baffle will provide an estimate for ES. The illuminance quantities EL and ES can also be estimated using photometric lighting calculation software, in conjunction with an estimate for EA.81 For the maximum light source luminance (LL), use a luminance meter with a 1/3-deg aperture to measure the highest luminance value anywhere on the surface of the light source. If the light source is small enough that it does not fill the aperture of the luminance meter, the glare source size is small enough that its maximum luminance need not be considered for DG (see Section 3.3). From the illuminance quantities, an interim DG quantity can be calculated as follows:

DG ¼ logðEL þ ES Þ þ 0.6 logðE L ∕ES Þ − 0.5 logðE A Þ:

(6)

If the light source is too small to measure its maximum luminance as described above, the corresponding DB20 rating value for visual discomfort is calculated from DG as follows:

DB ¼ 6.6 − 6.4 logðDGÞ:

(7a)

If the maximum light source luminance was able to be measured, the following equation can be used to estimate the corresponding DB rating value:

DB ¼ 6.6 − 6.4 logðDGÞ þ 1.4 logð50, 000∕LL Þ:

(7b)

The use of Eqs. (6) and (7) to estimate DG from a light source is appropriate to “white” sources of illumination. If these equations were used to compare light sources that vary widely in their color appearance (e.g., a “blue” versus a “yellow” source), the relative DG ratings could differ, as discussed in Section 3.5. Nonetheless, the model embodied in Eqs. (6) and (7) has been used successfully to predict DG ratings from LED streetlights82 and from cap lamps worn by coal miners.83 It should be noted that DG models, such as the unified glare rating (UGR) system,84 for evaluating lighting in interior spaces such as offices and classrooms, are based on very much the same factors as models to address glare from exterior lighting. These include the illuminance from a glare source, its luminance, the background luminance, and the angle between the glare source and the observer’s line of sight. The UGR model has sometimes predicted more DG than was actually reported,41 even for conditions representative of interior lighting.

Bullough: Understanding Glare in Exterior Lighting, Display, and Related Applications

35

4.3 Glare recovery modeling Although fundamental work to establish dark adaptation curves like the one shown in Fig. 3 (Section 2.3) has been undertaken, and strong correlational relationships between measured recovery times and glare exposure doses (in terms of lx·s) have been established for specific visual tasks (and for different locations in the field of view), modeling and predicting glare recovery times would be very challenging because of the strong influence of the task one is performing that is used to define recovery. For a very easy task like detecting a large, very bright object, recovery could be almost instantaneous even following a very intense exposure to light for a long duration. If the task is small and/or low-contrast and occurs at a very low light level close to total darkness, recovery times can be very long. Comparisons of two glare source exposures under otherwise identical conditions (location in the field of view, background light level, size and contrast of the visual task) can be made by comparing the exposure doses of the two glare conditions. 4.4 Reflected glare modeling In a study of various interior lighting conditions that included a wide variety of computer display types,85 a simple model of subjective acceptability of reflections in the screens was made for three types of computer display screens: l l l

Positive-polarity (white background), glossy-finish screens Positive-polarity, diffuse-finish screens Negative-polarity (dark background), diffuse-finish screens

Negative-polarity, glossy-finish display screens were not included because these almost always exhibit unacceptable and very conspicuous screen reflections. Observers rated the acceptability of images of ceiling luminaires reflected in the screens on a seven-point rating scale where 1 = never acceptable, 4 = sometimes acceptable, and 7 = always acceptable. Ratings of acceptability (R) for each type of screen were related to the luminous intensity (I, in cd) from the luminaire in the direction of the display85; for positive-polarity, glossy-finish screens were predicted as follows:

R ¼ 5.595 − 0.329I 0.4 :

(8a)

For positive-polarity, diffuse-finish screens, ratings of acceptability (R) were predicted by the following equation:

R ¼ 7.392 − 0.139I 0.4 :

(8b)

For negative-polarity, diffuse-finish screens, ratings of acceptability (R) were predicted as follows:

36

Bullough: Understanding Glare in Exterior Lighting, Display, and Related Applications

R ¼ 6.107 − 0.307I 0.4 :

(8c)

These predictive models were developed around the turn of the millennium.85 As display finishes have improved in recent years to provide reduced glossiness while maintaining image sharpness, and as they have increased in screen luminance, the formulae in Eq. (8) could be considered conservative because brighter or less glossy screens than those used to develop the model would be expected to be more acceptable than the model equations would predict.

5 Hints for Designing to Control Glare The preceding chapters of this Spotlight describe some of the technical details and predictive models that can be employed by the designer of a lighting system or by an optical engineer to reduce the negative impacts of glare. The present chapter will not provide an exhaustive treatment of detailed methods for glare control, but rather will briefly summarize a few approaches that might reduce the negative impacts of disability glare, DG, glare recovery, and reflected glare. 5.1 Disability glare control Methods for limiting the impacts of disability glare include: l

l

l

Locate luminaires away from the line of sight (e.g., for outdoor lighting, mount luminaires on poles as high as possible while maintaining appropriate illumination on the roads/sidewalks) to reduce the veiling luminance as shown in Fig. 14. Select “cutoff” types of outdoor luminaires86 or luminaires that produce the lowest amount of high-angle light. The Illuminating Engineering Society (IES) has published a rating system for outdoor luminaires87 that characterizes the distribution from luminaires by the acronym BUG, where “B” refers to light behind the luminaire (where it might illuminate adjacent properties rather than the road), “U” refers to uplight that can contribute to sky glow, and “G” refers to the high-angle light that can contribute to glare.87 Select luminaires with the lowest “G” rating value. Increase overall light levels making visual tasks easier to perform, so that contrast reductions from scattered light will have a smaller effect on visual performance. When feasible, use high-contrast, large visual tasks (e.g., contrasting tape colors on the edges of stairs to make them more conspicuous).

5.2 Discomfort glare control Methods for reducing the negative impacts of DG include:

Bullough: Understanding Glare in Exterior Lighting, Display, and Related Applications

l

l l l l l

37

Aim or orient luminaires, or use shields, to block a direct view so that they do not directly illuminate observers’ eyes, but rather the visual tasks that observers are performing. On a divided roadway, for example, shielding could be incorporated into a guardrail separating oncoming traffic directions. Design luminaires to have a luminous background element that increases the illuminance from the immediate surround. Locate high-reflectance surfaces behind lights that might create objectionable DG (see Fig. 27). Use diffusers, lenses, or other optical elements to reduce the maximum luminances, especially those of bare LEDs, HID sources, and filaments. Consider using light sources with a lower CCT to reduce short-wavelength (blue) output. If possible, utilize lighting controls to dim or switch off lights during nighttime or when ambient light levels are low.

5.3 Glare recovery control Several ideas for reducing the impacts of glare recovery and helping facilitate shorter recovery times include the following: l

l

Use bright lights in the field of view (such as warning lights) for as brief a duration as possible; using lower duty cycles (defined as the percentage of time a light source is on) may help reduce glare recovery times from flashing lights. If the visual task for a specific location requires peripheral vision, consider using light sources with lower CCTs.

Figure 27 Two identical outdoor luminaires; the one on the left is positioned in front of a dark surface and the one on the right in front of a light-colored surface. Sensations of DG will be worse from the luminaire on the left with the lower background luminance.

38

Bullough: Understanding Glare in Exterior Lighting, Display, and Related Applications

5.4 Reflected glare control Consider the following methods for controlling reflected glare: l l l

Position the light source to avoid reflections in the visual task. Use printed materials with matte rather than glossy surfaces. If glossy computer screens with white backgrounds are present, limit the luminous intensity of luminaires in the direction of the screens to 180 cd to ensure an average acceptability rating value of at least 3 (see Section 4.4).85 If diffuse screens with dark backgrounds are present, limit the luminous intensity of the luminaires in those directions to 320 cd. If diffuse screens with white background are present, limit the luminous intensity in those directions to 5500 cd. (Luminous intensities of luminaires as high as 5500 cd may result in discomfort or disability glare!)

6 Future Outlook Hopefully, this Spotlight has provided a brief and basic introduction to the principles underlying glare, its causes, and its effects on performance and comfort in varied lighting applications. Recognizing the limitations primarily with respect to brevity, this chapter provides several thoughts on the current state of research to characterize glare, and methods for its prediction and control. Presently, research investigations of glare are almost always segregated by specific applications. In this Spotlight, much of the primary focus has been on lighting applications that take place outdoors, such as vehicle headlighting and streetlighting, and on glare when using visual displays. As described earlier, these are the applications where glare, especially disability glare, DG, and glare recovery are most critical. It is likely that for these applications, where the basic effects of factors, such as light source intensity, luminance, and (when relevant) location are understood, that future work and development of lighting systems to control glare may focus on time as a critical factor. Adaptive lighting control schemes for vehicle headlights include automatic switching systems between high beam headlights and low beam headlights, and more recently, adaptive driving beam (ADB) headlighting systems; 88 see Fig. 28. ADB headlights allow a driver to have their high beam headlights on at almost all times, but use sensors and cameras to identify the locations of oncoming headlights and preceding taillights. With the angular locations of these lights, the intensity from the ADB headlights can be reduced specifically in those locations to reduce disability and DG for drivers located there. Adaptive streetlighting strategies have also begun to be used, mainly for the purpose of reducing energy use and light pollution. In adaptive streetlighting control, reducing the output of streetlights would also help reduce glare. At the time this Spotlight was produced, the temporal effects of lighting on glare were only beginning to be investigated.89

Bullough: Understanding Glare in Exterior Lighting, Display, and Related Applications

39

Figure 28 Appearance of oncoming high beam headlights (producing substantial glare), oncoming low beam headlights, and oncoming ADB headlights (adaptive matrix). The ADB headlights produce higher light levels along the roadway edge but appear no glaring than low beam headlights.

There is a large body of work that has investigated and continues to investigate the impacts of interior lighting and daylighting on glare in the built environment,17,90–92 and as previously mentioned, models to predict glare under these conditions.84 These results and models have rarely been applied to studies of nighttime glare, just as the results of nighttime glare studies have rarely been applied to interior lighting (see Ref. 41 for an exception). Yet there is no inherent reason that glare under both ranges of conditions should be substantially different. Certainly, the physiology of the visual system does not change based on ambient light levels. The same factors that impact glare outdoors at night will also impact glare indoors during the daytime, and there is some evidence that models devised for exterior applications can be successfully used to predict responses in interiors.41 And the growing understanding of the influence of spectral factors in glare, especially DG, could enlighten and be enlightened by applied research probing the intersection between visual neuroscience and lighting application.93 Perhaps it goes without saying, but the discussion of glare in this Spotlight has been exclusively focused on glare for human observers. Nonetheless, disability glare can be a major concern for machine systems like those illustrated in Fig. 28, when light from bright sources is scattered in camera lenses and other sensors. Many strategies to minimize glare for people, such as moving the glare source away from a camera’s line of sight when practical, will be helpful for machine systems. So could the use of polarizers of different orientations over potential glare sources and camera apertures.94 An Integrated Glare Metric (IGM; see Section 1) could help minimize confusion caused by numerous glare metrics and models tied to specific applications (e.g., pedestrian lighting, sports lighting, office lighting, daylighting, and visual displays) and perhaps even for different observers (e.g., humans and machine systems). Such a framework would need to include (or develop a rationale for ignoring) many of the factors described in this Spotlight. It is hoped that this brief overview discussion can help reduce uncertainty about the causes and implications of glare in all practical applications of light and lighting.

40

Bullough: Understanding Glare in Exterior Lighting, Display, and Related Applications

References 1. J. D. Bullough, Lighting Answers: Light-Emitting Diode Lighting Systems, Rensselaer Polytechnic Institute, Troy, New York (2003). 2. A. Bierman, “LEDs: from indicators to illumination?” Lighting Futures 3(4), 1–5 (1998). 3. W. S. Stiles, “The scattering theory of the effect of glare on the brightness difference threshold,” Proc. R. Soc. B: Biol. Sci. 105(735), 131–146 (1929).10.1098/rspb.1929.0033 4. National Highway Traffic Safety Administration, Nighttime Glare and Driving Performance: Report to Congress, U.S. Department of Transportation, Washington, D.C (2007). 5. K. Jost, “Anatomy of high-intensity discharge head-lamps,” Autom. Eng. 103(11), 38–44 (1995). 6. M. S. Rea, Ed., IESNA Lighting Handbook: Reference and Application, 9th ed., Illuminating Engineering Society of North America, New York (2000). 7. U.S. Department of Transportation, Standard No. 108: Lamps, Reflective Devices, and Associated Equipment, U.S. Department of Transportation, Washington, D.C (2004). 8. J. Van Derlofske, J. D. Bullough, and C. M. Hunter, “Evaluation of high-intensity discharge automotive forward lighting,” in Lighting Technology Developments for Automobiles, M. Flannagan, J. Jiao, and R. Karbowski, Eds., SP-1595, pp. 1–7, Society of Automotive Engineers, Warrendale, Pennsylvania (2001). 9. Texas Department of Public Safety, “1. Why don’t we aim headlights anymore?” in FAQ: Items of Inspection, Texas Department of Public Safety, Austin, Texas (2021). 10. N. P. Skinner, J. D. Bullough, and A. M. Smith, “Survey of the present state of headlamp aim,” in Transp. Res. Board 89th Annu. Meeting Proc., Transportation Research Board, Washington, D.C (2010). 11. National Highway Traffic Safety Administration, Glare from Head-lamps and Other Front Mounted Lamps, Docket No. NHTSA-2001–8885, U.S. Department of Transportation, Washington, D.C (2001). 12. J. Bullough and J. Van Derlofske, “Shedding light on glare,” Opt. Eng. Mag. 3(6), 17–18 (2003). 13. Anonymous, “Eight-track players,” Lighting Des. Appl. 41(4), 72–19 (2011). 14. M. Yontz, “Research reveals ‘visual chaos’ in service truck warning lights,” Serv. Truck Mag. 3(6), 1–8 (2016). 15. Society of Automotive Engineers, Optical Warning Devices for Au-thorized Emergency, Maintenance and Service Vehicles, J845, Society of Automotive Engineers, Warrendale, Pennsylvania (2021). 16. Society of Automotive Engineers, Directional Flashing Optical Warn-ing Devices for Authorized Emergency, Maintenance and Service Vehicles, J595, Society of Automotive Engineers, Warrendale, Pennsylvania (2021). 17. A. A. Nazzal, “A new evaluation method for daylight discomfort glare,” Int. J. Ind. Ergon. 35(4), 295–306 (2005).10.1016/j.ergon.2004.08.010 18. Underwriters Laboratories, Design Guideline for Promoting Circadian Entrainment with Light for Day-Active People, DG 24480, Underwriters Laboratories, Northbrook, Illinois (2019). 19. D. A. Schreuder, White or Yellow Light for Vehicle Head-Lamps? Institute for Road Safety Research, Voorburg (1976). 20. J. De Boer, Public Lighting, Philips Technical Library, Eindhoven (1967). 21. B. Wali et al., “How is driving volatility related to intersection safety? A Bayesian heterogeneity-based analysis of instrumented vehicles data,” Transp. Res. Part C: Emerg. Technol. 92(1), 504–524 (2018).10.1016/j.trc.2018.05.017 22. J. M. Cooper, N. Medeiros-Ward, and D. L. Strayer, “The impact of eye movements and cognitive workload on lateral position variability in driving,” Hum. Factors, 55(5), 1001–1014 (2013).10.1177/0018720813480177

Bullough: Understanding Glare in Exterior Lighting, Display, and Related Applications

41

23. J. D. Bullough and J. Van Derlofske, “Methods for assessing the impact of oncoming glare on driving behavior,” in Lighting Technology and Human Factors, J. Jiao and M. Flannagan, Eds., SP-1932, pp. 21–26, Society of Automotive Engineers, Warrendale, Pennsylvania (2005). 24. J. D. Bullough, “Investigation of the influence of headlight glare and aim on risk-related driving behavior,” in Society of Automotive Engineers World Congress Experience, Society of Automotive Engineers, Warrendale, Pennsylvania (2017). 25. N. P. Skinner and J. D. Bullough, “Visual recovery and discomfort following exposure to oncoming headlamps,” in Automotive Lighting Technology, J. Jiao, Ed., SP-2223, pp. 73–77, Society of Automotive Engineers, Warrendale, Pennsylvania (2009). 26. J. D. Bullough, “Measurement of light and color,” in Handbook of Measurement in Science and Engineering, M. Kutz, Ed., Vol. 3, pp. 2043–2074, Wiley, New York (2016). 27. G. A. Fry, “A re-evaluation of the scattering theory of glare,” Illum. Eng. 49(2), 98–102 (1954). 28. J. D. Bullough et al., An Investigation of Headlamp Glare: Intensity, Spectrum and Size, DOT HS 809 672, National Highway Traffic Safety Administration, Washington, D.C (2003). 29. J. D. Bullough et al., “Predicting discomfort glare from outdoor lighting installations,” Lighting Res. Technol. 40(3), 225–242 (2008).10.1177/1477153508094048 30. P. L. Olson and M. Sivak, “Discomfort glare from automobile headlights,” J. Illum. Eng. Soc. 13(3), 296–303 (1984).10.1080/00994480.1984.10748879 31. H.-J. Schmidt-Clausen and J. T. H. Bindels, “Assessment of discomfort glare in motor vehicle lighting,” Lighting Res. Technol. 6(2), 79–88 (1974).10.1177/096032717400600204 32. T. Krebs et al., “Recovery time of the human eye after exposure to a glare source at various intensities,” Lighting Res. Technol. 26(4), 199–201 (1994).10.1177/096032719402600405 33. J. Van Derlofske et al., “Headlight glare exposure and recovery,” SAE Trans. J. Passenger Cars – Mech. Syst. 114(6), 1974–1981 (2005). doi: 10.4271/2005-01-1573 34. J. D. Bullough, “Luminance versus luminous intensity as a metric for discomfort glare,” in Automotive Lighting Technology and Human Factors in Driver Vision and Lighting, J. Jiao and M. Flannagan, Eds., SP-2300, pp. 31–35, Society of Automotive Engineers, Warrendale, Pennsylvania (2011). 35. J. D. Bullough and K. Sweater Hickcox, “Interactions among light source luminance, illuminance and size on discomfort glare,” SAE Int. J. Passenger Cars – Mech. Syst. 5(1), 199–202 (2012).10.4271/2012-01-0269 36. E.-O. Rosenhahn and M. Lampen, “New investigation of the subjective glare effect of projection type headlamps,” SAE Trans. J. Passenger Cars – Mech. Syst. 113(6), 847–853 (2004). doi: 10.4271/2004-01-1281 37. T. Irikura, Y. Toyofuku, and Y. Aoki, “Recovery time of visual acuity after exposure to a glare source,” Lighting Res. Technol. 31(2), 57–61 (1999).10.1177/096032719903100204 38. C. J. Lloyd, M. Mizukami, and P. R. Boyce, “A preliminary model of lighting-display interaction,” J. Illum. Eng. Soc. 25(2), 59–69 (1996).10.1080/00994480.1996.10748148 39. J. Kaufman, Ed., Lighting Handbook: Reference Volume, Illuminating Engineering Society of North America, New York (1984). 40. U.S. Department of Energy, Solid-State Lighting R&D Plan, DOE/EE-1228, U.S. Department of Energy, Washington, D.C (2015). 41. X. Mou et al., “Preliminary evaluation of discomfort glare from organic light-emitting diode and edge-lit light-emitting diode lighting panels,” J. Biomed. Opt. 22(5), 055004 (2017).10.1117/1.JBO.22.5.055004 42. W. Kim and J. T. Kim, “Effect of background luminance on discomfort glare in relation to the glare source size,” Indoor Built Environ. 19(1), 175–183 (2010).10.1177/1420326X09358006 43. J. D. Bullough et al., “Discomfort glare from headlamps: Interactions among spectrum, control of gaze and background light level,” in Lighting Technology, J. Jiao and M. Flannagan, Eds., SP-1787, pp. 21–25, Society of Automotive Engineers, Warrendale, Pennsylvania (2003).

42

Bullough: Understanding Glare in Exterior Lighting, Display, and Related Applications

44. Canadian Centre for Occupational Health and Safety, Computer Monitors and Display Colours, Canadian Centre for Occupational Health and Safety, Hamilton, Ontario (2021). 45. J. M. Fugate and G. A. Fry, “Relation of changes in pupil size to visual discomfort,” Illum. Eng. 51(7), 537–549 (1956). 46. G. A. Fry and V. M. King, “The pupillary response and discomfort glare,” J. Illum. Eng. Soc. 4(4), 307–324 (1975).10.1080/00994480.1975.10748533 47. S. M. Berman et al., “The influence of spectral composition on discomfort glare for large-size sources,” J. Illum. Eng. Soc. 25(1), 34–41 (1996).10.1080/00994480.1996.10748131 48. J. J. Vos, “Reflections on glare,” Lighting Res. Technol. 35(2), 163–175 (2003).10.1191/ 1477153503li083oa 49. J. D. Bullough, Z. Fu, and J. Van Derlofske, “Discomfort and disability glare from halogen and HID headlamp systems,” in Advanced Lighting Technology for Vehicles, J. Jiao, M. Flannagan, and R. Karbowski, Eds., SP-1668, pp. 1–5, Society of Automotive Engineers, Warrendale, Pennsylvania (2002). 50. L. L. Holladay, “The fundamentals of glare and visibility,” J. Opt. Soc. 12(4), 271–319 (1926).10.1364/JOSA.12.000271 51. N. P. Skinner and J. D. Bullough, “Influence of intensity, duration and spectral characteristics on glare recovery for peripheral visibility,” in Society of Automotive Engineers World Congress Experience, Society of Automotive Engineers, Warrendale, Pennsylvania (2020). 52. M. J. Flannagan, Subjective and Objective Aspects of Headlamp Glare: Effects of Size and Spectral Power Distribution, UMTRI-99-36, University of Michigan, Ann Arbor, Michigan (1999). 53. R. Steen et al., “Effects of filters on disability glare,” Ophthalm. Physiol. Opt. 13(4), 371–376 (1993).10.1111/j.1475-1313.1993.tb00493.x 54. J. D. Bullough, N. P. Skinner, and M. S. Rea, “Impacts of flashing emergency lights and vehicle-mounted illumination on driver visibility and glare,” in Society of Automotive Engineers World Congress Experience, Society of Automotive Engineers, Warrendale, Pennsylvania (2019). 55. M. S. Rea and J. D. Bullough, “Two types of glare: two visual channels,” Scand. J. Optometry Vis. Sci. 11(2), 8 (2018). doi: 10.5384/SJOVS.vol11i2p1-8 56. J. D. Bullough, “Spectral sensitivity for extrafoveal discomfort glare,” J. Mod. Opt. 56(13), 1518–1522 (2009).10.1080/09500340903045710 57. Commission Internationale de l’Eclairage, CIE 10 Degree Photopic Photometric Observer, CIE 165, Commission Internationale de l’Eclairage, Vienna, Austria (2005). 58. V. C. Smith and J. Pokorny, “Spectral sensitivity of the foveal cone photopigments between 400 and 500 nm,” Vision Res. 15(2), 161–171 (1975).10.1016/0042-6989(75)90203-5 59. R. A. Weale, “Spectral sensitivity and wave-length discrimination of the peripheral retina,” J. Physiol. 119(2–3), 170–190 (1953).10.1113/jphysiol.1953.sp004837 60. B. R. Wooten, K. Fuld, and L. Spillmann, “Photopic spectral sensitivity of the peripheral retina,” J. Opt. Soc. Am. 65(3), 334–342 (1975).10.1364/JOSA.65.000334 61. J. D. Bullough and Y.-W. Liu, “Response to white light emitting diode aviation signal lights varying in correlated color temperature,” Transp. Res. Rec. 2673(3), 667–675 (2019).10.1177/ 0361198119834007 62. K. Sweater-Hickcox, N. Narendran, and J. D. Bullough, “Effect of different coloured luminous surrounds on LED discomfort glare perception,” Lighting Res. Technol. 45(4), 464–475 (2013). 10.1177/1477153512474450 63. R. M. Nagare and J. D. Bullough, “Influence of background spectral distribution on perceptions of discomfort glare,” SAE Int. J. Adv. Curr. Pract. Mob. 2(5), 2581–2584 (2020).10.4271/202001-0637

Bullough: Understanding Glare in Exterior Lighting, Display, and Related Applications

43

64. V. Reading, “Yellow and white headlamp glare and age,” Trans. Illum. Eng. Soc. 31(4), 108–121 (1966).10.1177/147715356603100402 65. A. E. Smart, “An investigation of glare from white and yellow sources,” Lighting Res. Technol. 1(2), 106–109 (1969).10.1177/14771535690010020101 66. H. Jin et al., “Research on the lighting performance of LED street lights with different color temperatures,” IEEE Photonics J. 7(6), 1601309 (2015).10.1109/JPHOT.2015.2497578 67. N. P. Skinner and J. D. Bullough. “Influence of LED spectral characteristics on glare recovery,” in Society of Automotive Engineers World Congress Experience, Society of Automotive Engineers, Warrendale, Pennsylvania (2019). 68. Y. He et al., “Evaluating light source efficacy under mesopic conditions using reaction times,” J. Illum. Eng. Soc. 26(1), 125–138 (1997).10.1080/00994480.1997.10748173 69. J. D. Bullough and M. S. Rea, “Simulated driving performance and peripheral detection at mesopic and low photopic light levels,” Lighting Res. Technol. 32(4), 194–198 (2000).10.1177/ 096032710003200403 70. J. A. Worthey, “Effect of veiling reflections on vision of colored objects,” J. Illum. Eng. Soc. 18(2), 10–15 (1989).10.1080/00994480.1989.10748755 71. M. Sivak et al., “The influence of stimulus duration on discomfort glare for persons with and without visual correction,” Transp. Hum. Factors 1(2), 147–158 (1999).10.1207/sthf0102_4 72. J. K. IJspeert et al., “The intraocular straylight function in 129 healthy volunteers: dependence on angle, age and pigmentation,” Vision Res. 30(5), 699–707 (1990).10.1016/0042-6989(90) 90096-4 73. Commission Internationale de l’Eclairage, CIE Equations for Disability Glare, CIE 146, Commission Internationale de l’Eclairage, Vienna, Austria (2002). 74. C. H. Rex and J. S. Franklin, “Visual comfort evaluations of roadway lighting,” Highway Res. Board Bull. 255, 101–116 (1960). 75. P. Dee, “The effect of spectrum on discomfort glare,” M.S. thesis, Rensselaer Polytechnic Institute, Troy, New York (2003). 76. F. Schieber, “Age and glare recovery time for low-contrast stimuli,” Proc. Hum. Factors Ergon. Soc. Annu. Meeting 38(9), 496–499 (1994).10.1177/154193129403800908 77. M. Sivak et al., “Discomfort glare is task dependent,” Int. J. Veh. Des. 12(2), 152–159 (2014).10.1504/IJVD.1991.061677 78. J. Theeuwes, J. Alferdinck, and M. Perel, “Relation between glare and driving performance,” Hum. Factors 44(1), 95–107 (2002). doi: 10.1518/0018720024494775 79. N. Tuaycharoen and P. R. Tregenza, “View and discomfort glare from windows,” Lighting Res. Technol. 39(2), 185–200 (2007).10.1177/1365782807077193 80. Q. Li et al., “A survey of the luminance distribution in the nocturnal environment in Shanghai urban areas and the control of luminance of floodlit buildings,” Lighting Res. Technol. 38(3), 185–189 (2006).10.1191/1365782806lrt160oa 81. J. A. Brons, J. D. Bullough, and M. S. Rea, “Outdoor site-lighting performance: a comprehensive and quantitative framework for assessing light pollution,” Lighting Res. Technol. 40(3), 201–224 (2008).10.1177/1477153508094059 82. Y. Tyukhova and C. E. Waters, “Discomfort glare from small, high-luminance light sources when viewed against a dark surround,” Leukos 14(4), 215–230 (2018).10.1080/ 15502724.2018.1434415 83. J. J. Sammarco et al., “Discomfort glare comparison for various LED cap lamps,” IEEE Trans. Ind. Appl. 47(3), 1168–1174 (2011).10.1109/TIA.2011.2126016 84. Commission Internationale de l’Éclairage. Discomfort Glare in Interior Lighting, CIE 117, Commission Internationale de l’Éclairage, Vienna, Austria (1995).

44

Bullough: Understanding Glare in Exterior Lighting, Display, and Related Applications

85. N. J. Miller, P. R. Boyce, and P. Ngai, “A metric for judging acceptability of direct luminaires in computer offices,” J. Illum. Eng. Soc. 30(2), 12–29 (2001).10.1080/00994480.2001. 10748350 86. Illuminating Engineering Society of North America, American National Standard Practice for Roadway Lighting, RP-8-00, Illuminating Engineering Society of North America, New York (2000). 87. Illuminating Engineering Society, Luminaire Classification System for Outdoor Luminaires, TM-15-20, Illuminating Engineering Society, New York (2020). 88. J. D. Bullough, N. P. Skinner, and T. T. Plummer, “Assessment of an adaptive driving beam headlighting system: visibility and glare,” Transp. Res. Rec. 2555, 81–85 (2016).10.3141/ 2555-11 89. J. Girard, C. Villa, and R. Bremond, “Discomfort glare from a cyclic source in outdoor lighting conditions,” Leukos. doi: 10.1080/15502724.2021.1954531 90. C. Quek et al., “Comparing performance of discomfort glare metrics in high and low adaptation levels,” Build. Environ. 206(1), 108335 (2021).10.1016/j.buildenv.2021.108335 91. C. Pierson, J. Wienold, and M. Bodart, “Review of factors influencing discomfort glare perception from daylight,” Leukos 14(3), 111–148 (2018).10.1080/15502724.2018.1428617 92. L. M. Geerdinck et al., “Discomfort glare perception of non-uniform light sources in an office setting,” J. Environ. Psychol. 39(1), 5–13 (2014).10.1016/j.jenvp.2014.04.002 93. M. S. Rea, Value Metrics for Better Lighting, SPIE Press, Bellingham, Washington (2013). 94. C. Walker, “Filters reduce glare in automotive canister inspection,” Vision Syst. Des. 19(5), 17–19 (2014).

Bullough: Understanding Glare in Exterior Lighting, Display, and Related Applications

45

John D. Bullough John D. Bullough, Ph.D., FIES is a Program Director, Population Health Science and Policy, at the Light and Health Research Center, part of the Icahn School of Medicine at Mount Sinai. He manages the center’s research programs in transportation lighting, safety and human factors. He earned his doctorate in Multidisciplinary Science from Rensselaer Polytechnic Institute and has written or co-written more than 500 articles and technical publications on lighting.