Volume 36, Issue 2 p. 14-24
Frontline Technology
Free Access

Evaluating Display Reflections in Reflective Displays and Beyond

First published: 30 March 2020
Citations: 3


Although reflected ambient light can be disturbing when viewing some displays, it's essential for seeing reflective displays. Here, two experts explain how to properly measure display reflection to predict the visual performance of a display technology under realistic lighting conditions.

THE ADVENT OF PORTABLE DISPLAYS HAS GIVEN RISE TO a diverse array of mobile electronic devices, from smartphones to signage. Although this greatly benefits consumers, it requires that the information carried by the display be visible over a wide range of ambient lighting conditions, both indoor and outdoor. Ambient illumination comes mostly from diffuse light-scattering backgrounds (illuminated walls and ceilings; a blue or cloudy sky), with contributions from direct sources (lamps or the sun). Consumers expect their devices to be readable at every illumination level, from the dead of a moonless night to the direct sunlight of a clear day. In this article, we describe how best to evaluate a display's reflective characteristics for its suitability in a given illumination environment.

Emissive Versus Reflective Displays

Depending on the display technology, ambient light can be helpful or disturbing. Emissive displays show information by modulating emitted light. Because ambient light reflections disturb the emissive signal, these displays' native optical characteristics are measured in a darkroom—ambient light is only of interest to determine the effect of unwanted reflections.

Display manufacturers have explored surface treatments to minimize and extend the range of ambient illumination where the display can still be readable. These include antireflective (AR) coatings that reduce mirror-like specular reflections and antiglare (AG) surface treatments that diffuse specular reflections, reducing their brightness and dissipating distracting reflected images. However, these surface treatments become ineffective at higher illumination levels or in direct sunlight. In many cases, such as smartphones and smartwatches, the user can simply tilt the display to eliminate unwanted glare reflections. Sometimes, however, even the diffuse illumination background from the sky or in a bright room is so high that its reflection overwhelms the displayed information on an emissive display. The brute-force approach is to increase the backlight intensity of emissive displays at the cost of battery life.

Reflective displays show information by modulating reflected light. Therefore, an accurate characterization of their performance must specify the illumination used during the measurement. Any reflective display consists of at least two basic optical elements: a reflector (such as paper) and a reflection modulator (for example, pigment). The reflector's optical characteristics determine the appearance of the “white” background. To appear paper-like, its reflection should be as high as possible, spectrally uniform, and diffuse with a near-Lambertian scatter characteristic (independent of the viewing direction). The modulator's spectral characteristics determine whether the reflected light appears monochrome or colored.

The various reflective display technologies use different means to reflect and modulate ambient light.1 Electrophoretic displays (EPDs) use electrically charged pigments. With their paper-like appearance, low power consumption, and total absence of flicker, EPDs gained wide market acceptance and are known as electronic paper or ePaper. Opaque white pigments with near-Lambertian reflection make up the “paper” in ePaper, forming the background for the black or color pigments that absorb or spectrally modulate reflected light in the same way ink does on traditional paper.2 This display type is suited to a wide range of static and mobile applications, including e-readers, wearables, and signage for indoor and sunlight-readable outdoor environments.

Other reflective display technologies may use non-Lambertian reflectors, including mirrors or retroreflectors (gain reflectors), which are often combined with diffusers to appear less metallic and more like paper. Some technologies switch reflection by liquid crystal shutters or an electrophoretic black pigment that frustrates total internal reflection, then color the reflected light by using color filter arrays (CFAs). Others, such as solid-state reflective color displays, combine the functions of reflector and spectral modulators into color-changing mirrors.

Unfortunately, not all light reflected by traditional paper or ePaper is useful. Light reflected at the front surface does not reach the white reflector, meaning it cannot be modulated by the pigments of a reflective display or a glossy magazine (Fig. 1). As in emissive displays, such reflections not only produce disturbing mirror images of the ambient environment but also reduce ambient contrast. Display designers use AR and AG surface treatments to minimize this. Fig. 1(a) shows a monochrome e-reader with an AG surface receiving useful illumination from direct sunlight and diffuse light as well as the disturbing reflection from a table lamp into the specular direction. Fig. 1(b) shows glossy and matte printed traditional paper under the same mixed illumination, but only the glossy paper has the disturbing reflection from the table lamp.

Details are in the caption following the image

(a) Electronic paper (ePaper) and (b) glossy versus matte printed paper in an ambient illumination environment.

The General Light Reflection Model

To better comprehend the evaluation of display reflections, it helps to have a general understanding of the light-reflection process. In the simplest model, light reflected from a surface is comprised of three components:3

  • Mirror-like specular reflection (S), which produces a distinct virtual image of the source in the specular (or regular) direction. The luminance of this mirror image is proportional to the luminance of the source. Specular reflection occurs not only with a flat front surface (for example, the glass of a tablet display or glossy magazine paper) but with any internal interface where the refractive index changes.

  • Diffuse Lambertian reflection (DL), which scatters the incident light uniformly into all directions. The luminance of this scatter is independent of the directions of illumination and viewing and is proportional to the illuminance received from the light source. Lambertian reflection occurs at scattering surfaces—for example, copy paper or the white pigments used in EPDs.

  • Haze (DH), which appears as a fuzzy ball surrounding the specular image.4 It's distinct from the specular and Lambertian components, yet combines properties of both. Like specular reflection, its luminance peaks in the specular direction; but like Lambertian reflection, the reflected luminance is proportional to the illuminance.3 Haze reflection is a consequence of the matte AG surface commonly used on reflective e-readers and for emissive laptop and desktop displays.

Whether these three reflection components are observable singly or in combination depends on the display characteristics. A surface may have a summation of one, two, or all three of these scattering characteristics simultaneously. The reflection components can be observed in a dark room by shining a small-point light source such as a penlight on the display and examining the surface from various angles. A Lambertian surface will light up uniformly. A specular surface will produce a distinct virtual image of the source, but haze reflection will blur this image into a fuzzy ball. Fig. 2 illustrates the possible combinations.

Details are in the caption following the image

Illustrations of the reflected Lambertian, specular, and haze components separately and in combination.5

If the human observer is replaced by an imaging device with sufficiently high resolution and dynamic range, then the reflection's captured virtual images will closely resemble those shown in Fig. 2. A slice through the center of each image (see the left side of Fig. 3) will produce profiles of scattered luminance versus viewing direction. A common reflection metric often used to represent this angular scattering profile is the bidirectional reflection distribution function (BRDF).

Details are in the caption following the image

Conceptual example of the three-component light-scattering model.

In the BRDF profiles shown in Fig. 3, each reflection component is identifiable from the shape of its distribution. If all three components are present, the distributions add up to S+DH+DL. Fig. 4 shows this in more detail. The Lambertian reflection DL forms a flat background from which the bell-shaped haze distribution DH rises; this peaks at the specular reflection angle, and is topped by the sharp spike of specular reflection S.

Details are in the caption following the image

Example distribution of reflected luminance from a surface with specular, haze, and Lambertian scatter.5

Ideally, a high-resolution BRDF would represent the reflected scatter distribution for a point light source. However, real light sources are larger. The sun subtends an angle greater than 0.5°, and depending on the distance to the display, indoor light sources or windows can subtend 10° and more. Increasing the source's subtense can be conceptualized as clustering multiple adjacent point sources. Each point source's specular component is the same at each location in space (and angle), but the contributions from diffuse (haze plus Lambertian) components increase with the source's size as more scattered light comes from the adjacent points. Fig. 5(a) illustrates this concept in a one-dimensional simulation of a small source subtending 1°, as seen from the display. The 1° subtense light source can be constructed from a series of individual point-source luminance profiles. As the light source gets larger in Fig. 5(b), the diffuse haze component becomes more significant; but the specular component's size remains the same.

Details are in the caption following the image

One-dimensional simulation of the reflectance distribution from light sources of (a) 1° and (b) 0.25 to 15°, using the point source data from Fig. 4. In (a), the 1° source (dark blue) is the superposition of the individual point sources (colored curves).

The simulation results in Fig. 6 show that although the center luminance from specular reflection remains constant, the luminance from diffuse reflection increases with the reflected source's subtense (and illuminance). The rate of increase declines and almost levels off once the source subtense equals the haze distribution's width. Further increases in diffuse luminance arise mainly from the constant Lambertian reflection. However, for very small subtense sources (<1°) shown in the enlargement at the right of Fig. 6, the illuminance from the light source, and thus the luminance from diffuse reflection, goes to zero, leaving the constant specular component to dominate the center luminance value.

Details are in the caption following the image

A simulated relationship between the center luminance of specular and diffuse reflection and source subtense varied from 1 to 15°, using the data from Fig. 5. The chart on the right enlarges the range of very small apertures up to 0.5°.

The General Illumination Model for Viewing Displays

After establishing the general framework for describing reflected light, the next step is to define a model of the light sources that illuminate the display in order to assess it. In a general lighting environment, the display surface will be illuminated by several light sources. Ambient sources—for example, direct sunlight and diffuse skylight—are incoherent and, therefore, additive for each spectrum wavelength. The infinite complexity of ambient illumination from multiple sources can be approximated by linear combinations of two basic types of light sources: discrete directional and hemispherical-diffuse.6 In a natural outdoor environment, as Fig. 7(a) shows, directional illumination comes only from the disk of the sun that subtends 0.5°. Other light is hemispherical-diffuse—either blue skylight or light filtered through clouds on an overcast day. Inside, as Fig. 7(b) shows, light from lamps is directional, each with their own subtense and illumination spectrum, while light scattered from the ceiling and walls is diffuse.

Details are in the caption following the image

Ambient lighting environment for display viewing in (a) an outdoor environment5 and (b) an indoor environment.

Fig. 7b: Images courtesy of Vecteezy.com

To model a given indoor or outdoor illumination environment, the individual contributions from all sources must be tallied, each with their own illumination spectrum, direction of incidence relative to the display, and source subtense. The indoor and outdoor illumination should replicate the ambient lighting conditions specified by several standards. For example,7-9 reference outdoor illumination on an average clear sunny day contains 65,000 lux (lx) directional sunlight from an angle of 45°, with a spectrum approximating the CIE D50 illuminant, plus 15,000 lx of diffuse hemispherical skylight approximating D75. In an office, a display may receive 200 lx from a D50 ceiling light, plus 300 lx of indirect, scattered illumination. Standards concerned with establishing ergonomic workplace conditions with minimum disturbing glare specify the luminance and subtense of glare sources for viewing displays in diffuse ambient lighting.10

Defining the Reflection Geometries for ePaper

With an emissive display, all reflections from the illuminating ambient environment are potentially disturbing. For a reflective display, the reflection of ambient light can be useful (carries information), disturbing (carries no information), or a mixture of both. Therefore, it's important to determine which of the reflection components—Lambertian, haze, or specular—carries the reflective display information by defining reflection measurements with a geometry that can isolate useful, information-carrying reflection from disturbing reflection. The BRDF measurement is the most normal method to distinguish between the reflection components. The BRDF of EPD11 can guide the choice of standardized reflection measurement methods that may be employed.8, 9

Fig. 8 compares BRDF distributions of EPD modules measured with two different geometries: in-plane (with specular reflection included; see the blue curves in the figure) and viewing direction (with a display tilted to reflect the specular peak away by 45°; see the red curves). The in-plane BRDF show all three reflection components: specular, haze, and Lambertian. Lambertian reflection is useful because it changes with display color Q (where Q=W for white and k for black). Reflection from a glossy surface produces an intense and disturbing virtual image of the source, unchanged by the display color. A matte display surface blurs this unwanted image into haze with reduced peak luminance but wider distribution. However, the haze does not change with the display color, so it is not useful. The contrast reduction toward the specular angle in Fig. 8(c) occurs because of disturbing specular and haze reflections.

Details are in the caption following the image

Bidirectional reflection distribution function (BRDF) measured in-plane and in a viewing direction configuration on electrophoretic modules with (a) glossy and (b) matte surfaces, plus (c) contrast comparison with an illustration of an e-reader at 0, 45, and 60° viewing directions and in-plane with disturbing reflection.11

Just as the viewer of a handheld display instinctively tilts it to exclude virtual source images and avoid contrast loss, the viewing-direction BRDF geometry11 excludes disturbing reflection components while leaving the useful Lambertian unchanged (see the red curves in Fig. 8). Measuring the useful Lambertian reflection without including disturbing haze requires inclining the source or detector to 45° from the display's normal direction. Fig. 9 shows a display's BRDF with all three reflection components (S, DH, DL). It illustrates how measurement geometries can be chosen to capture these components separately or in combination. In each, the detector “views” the display from a near-normal direction, using a small measurement field angle (Ψd ≤ 1°). Each of the four measurement geometries (also detailed in the following list) is represented by an arrow, and each one's width corresponds to its source aperture subtense Ψs.

  • Directional or ring light 45°/0°. Fig. 10(a) shows the geometry that simulates directional illumination. With the source inclined at 45° and implemented as a ring light, disturbing glare reflection is effectively excluded from the normal viewing direction. This is the best choice for measuring the useful Lambertian reflection in EPD. It simulates viewing a display with the sun reflected away from the normal viewing direction in Fig. 11(a).

  • Hemispherical-diffuse reflection. Implemented with a sampling integrating sphere, the diffuse with specular component included di/8° geometry in Fig. 10(b) simulates diffuse illumination. The detector is positioned slightly off-normal to avoid self-reflection. Specular and haze components are reflected onto the sphere wall, and thus included in the measurement. Hemispherical-diffuse reflection (with the specular included) measures the sum of all useful and disturbing reflection components. This, too, is a useful measurement because it simulates viewing a display with the sky or white walls and ceiling reflected into an off-normal viewing direction in Fig. 11(b).

  • Specular reflection. The specular or regular reflection geometry with equal inclinations of source and detector in Fig. 12 examines disturbing reflections. By using a uniform source with a variable aperture subtense (variable aperture source or VAS), it should be possible to distinguish between specular and diffuse reflection components.12, 13 For a constant luminance source, the aperture size only changes the illuminance on the display. The diffuse component of the reflected luminance is proportional to the illuminance from the source, while the specular component of the reflected luminance is proportional to the source luminance.3 This measurement simulates viewing a display with a bright object (for example, a lamp or window) reflected into an off-normal viewing direction as shown in Fig. 11(c).

Details are in the caption following the image

Measurement geometries derived from BRDF.

Details are in the caption following the image

Reflection measurements for (a) directional with a 45°/0° ring light, and (b) hemispherical-diffuse with a di/8° sampling sphere.6

Details are in the caption following the image

Viewing a display (a) with the sun reflected away from the normal viewing direction, (b) with the sky, and (c) with a light source reflected into an off-normal viewing direction.

Details are in the caption following the image

Specular reflection geometry with a uniform source with (a) small 1° and (b) large 15° aperture.

Formalism to Predict Display Performance in Ambient Light

Having established the foundational concepts, we can apply them using a flexible and general formalism to predict the display performance under specific ambient illumination conditions. This requires measuring the display reflection characteristics and then combining the measured data with models that describe the display's ambient illumination environment. What we see on a display is “visible radiance” (luminance and colors), which is a combination of light emitted from the display plus ambient light reflected off the display.5 In other words, the total luminance or spectral radiance entering the eye from a displayed color (for example, a white, black, or colored screen) is the sum of the display's emission plus all the ambient light reflected by it. Predicting the display response to multisource ambient illumination requires the three steps listed below and illustrated in Fig. 13.

Details are in the caption following the image

Determining ambient display luminance from measured characteristics of display emission and reflection, plus ambient illumination models.

Image Measure each component of display reflection and emission. Simulate each type of illumination in the laboratory by selecting the appropriate measurement geometry: directional, hemispherical-diffuse, or specular reflection. Each measurement yields a spectral reflection coefficient that represents the display response to that illumination-detection geometry. If the display emits light— for example, an e-reader with front light—measure its spectral radiance in the darkroom.

Image Scale to the real-world ambient environment. The relative contribution of each type of illumination is determined by scaling the reflected light from laboratory-measurement illumination levels to those of the desired real-world situation. It is important to distinguish whether the light source reflection is diffuse or specular. If diffuse, then the reflected luminance (spectral radiance) is proportional to the display illuminance (spectral irradiance). If specular, then the reflected luminance is proportional to the light source luminance. In either case, each measured spectral reflectance component must be multiplied by the irradiance or radiance spectrum of the ambient daylight or indoor light to be simulated.

Image Linear superposition. The total ambient spectral radiance is the sum (at each wavelength) of contributions from all source reflections and display emissions (if applicable).

Scaling and linear superposition allow us to estimate the display response in any realistic multisource environment. Each reflection component—specular or diffuse, useful or disturbing—conveniently is measured in the laboratory without needing to simulate actual levels and spectra, such as those of solar illumination. The components of useful and disturbing reflection can be scaled individually to determine maximum acceptable levels of disturbing illumination. Then International Commission on Illumination (CIE) colorimetry14 is used to calculate display performance parameters such as luminance, chromaticity, contrast, and color gamut from the display's total ambient spectral radiance. This general measurement methodology for predicting the ambient contrast and color of both emissive and reflective displays6, 7 has led to the first measurement standards for electrophoretic ePaper displays8 and ePaper displays with built-in front illumination.9

Example Evaluation for ePaper Using Matte and Glossy Surfaces

The methodology outlined so far for predicting ambient contrast and color from a combination of measured display emission and reflection (with models of useful illumination) has been successfully applied to color EPD using CFAs15 and colored pigments.16 An example application of this methodology is illustrated for a black and white EPD. In this case, the effectiveness of a matte AG surface and glossy surface are compared under a combined (directional+diffuse) illumination that is both useful and disturbing.

Measuring the Reflection of ePaper in Useful Illumination

We evaluate the paper-like quality of EPD by using the modified illuminators described in prior work15, 17 to measure the reflection of directional and hemispherical-diffuse illumination for various viewing directions θd. Fig. 14 compares the viewing direction dependence for two monochrome electrophoretic display modules of directional reflectance RQ,dir(θd) to that of hemispherical-diffuse reflectance ρQ,hemi(θd) when displaying white (Q=W) or black (Q=k). The reflection coefficients for the 15° viewing direction are given in Table 1. The two modules are identical except for their surface: glossy versus matte AG. Both displays exhibited higher levels of reflectance under diffuse illumination.

Details are in the caption following the image

Viewing direction dependence of two electrophoretic display modules—one with a glossy surface, the other with a matte antiglare (AG) surface. (a) Directional and (b) hemispherical-diffuse reflectance.17

Table 1. Luminous reflectance factors and reflectance measured on monochrome electrophoretic display modules with glossy and matte AG surfaces.
45°/15° RQ,dir W(white) 0.31 0.30
k (black) 0.020 0.021
di/15° ρQ,hemi W 0.49 0.49
k 0.078 0.093
15°/15° ζspec W, k 0.036 0
RH + RQ,L W 0.059 0.063
k 0.052 0.057

Fig. 15(a) shows the contrast ratio (CR) of reflection versus viewing direction. Under directional illumination, both displays have a higher CR compared to that in hemispherical-diffuse illumination. For both types of illumination, the matte AG surface slightly reduces CR because of the light scatter it introduces. Fig. 15(b) shows that the ambient CR is a linear combination of those measured in pure directional or hemispherical-diffuse illumination, and thus falls between the two: lower than directional, but higher than hemispherical-diffuse CR. The reference indoor illumination with its lower amount of directional light (200 lx directional, 300 lx hemispherical-diffuse) has lower CR compared to the reference outdoor environment with more directional sunlight (65,000 lx) than diffuse skylight (15,000 lx). Where lighting is totally diffuse—for example, in a room with indirect lighting or outdoors under completely overcast skies or during snowfall—the ambient CR might drop to the lowest level measured under hemispherical-diffuse illumination.

Details are in the caption following the image

Viewing-direction dependence of the luminance contrast ratio of (a) directional versus hemispherical-diffuse reflectance, and (b) outdoor versus indoor illumination for monochrome electrophoretic modules with a glossy versus matte AG surface.17

In useful illumination, both outdoors and indoors, scatter from the matte AG surface reduces CR compared to the glossy surface. So, one might ask, what's the point of an AG surface?

Assessing the Reflection of ePaper in Disturbing Illumination

To understand the value of the AG surface, the EPD must be evaluated not only with useful illumination, but also with added disturbing glare illumination. This requires a reflection measurement with the light source in the specular (regular) reflection geometry. In Fig. 16 a ceiling light not only provides the useful illumination necessary to see the displayed information, but also causes a specular glare reflection that affects the contrast between white and black. Glare reflection from a glossy surface obliterates the information shown on the EPD, but the AG surface scatters glare so that the information remains recognizable.

Details are in the caption following the image

Specular reflection of a fluorescent ceiling light by the electrophoretic reflective display samples with glossy and matte AG surfaces.

The images in Fig. 17 show specular reflection for light source apertures ranging in size between 1° and 15°, and the corresponding reflectance characteristics of the two EPDs. This was obtained by variable aperture source (VAS) measurements using full screen black and white patterns. Because the displays are identical (save for their surface treatments), the change in VAS reflectance with source aperture size provides insight into each display's reflection characteristics in the specular direction. The VAS reflectance was measured relative to a black glass specular reflectance reference. Varying the light source aperture size changes only the illuminance falling on the display; the source luminance itself remains constant (after applying corrections for source variations18). Thus, the two different surfaces—one mirror-like, the other scattering—result in two fundamentally different VAS reflection characteristics for the displays (see Fig. 17). The glossy surface, Fig. 17(a), produces a sharp mirror image of the source aperture, and the reflection from this glossy display is predominantly specular. The specular reflected luminance does not change with aperture size. When displaying black, the measured VAS reflection is nearly flat as aperture subtense increases; it does not change with illuminance. However, when switched to white, there is a slight increase with aperture as a result of the additional Lambertian reflection of white pigment.

Details are in the caption following the image

Variable aperture source reflection characteristics for electrophoretic modules displaying white or black, with (a) glossy and (b) matte AG surfaces.18

In Fig. 17(b), the matte AG surface effectively dissipates the mirror-like reflection, turning it into a diffuse haze. With increasing source aperture size, the luminance of the blurry reflected aperture image seems to increase, and the measured VAS reflection increases sharply. Thus, the reflection from this matte display is predominantly diffuse, mostly from haze created by the AG scattering surface, but also from Lambertian scattering (as indicated by the increasing difference between white and black).

Estimates for specular and diffuse reflection are obtained from the measured VAS characteristics, using the smallest aperture to estimate the specular, and the largest to estimate the combined specular and diffuse contributions.6, 18 Table 1 summarizes the measured reflection properties of the two displays in useful and disturbing illumination, with directional and hemispherical-diffuse reflection factors RQ,dir and ρQ,hemi from Fig. 14, while specular and diffuse glare-reflectance components are estimated from VAS measurements in Fig. 17. For electrophoretic reflective displays, only the diffuse Lambertian component of glare reflectance is color-dependent.

Ambient Display Characteristics in Useful and Disturbing Illumination

When a display is viewed in an ambient environment where disturbing glare cannot be excluded, the glare term must be added to the two useful ambient reflection terms (directional and hemispherical-diffuse) in Fig. 13 to calculate the total spectral radiance LQ,T(λ).

Fig. 18 illustrates the geometry where both useful and disturbing light sources are present. Disturbing glare only occurs if the glare source is directly reflected in the viewing direction θd. When the display is viewed normally, θd = 0, the viewer's head will occlude any disturbing light coming from this direction. If the viewer's face is illuminated brightly, it can become a source of disturbing glare as well. If the display is viewed from an off-normal direction, then a light source or brightly illuminated object from the direction of mirror reflection will produce glare.

Details are in the caption following the image

Illumination-reflection geometry with directional, hemispherical-diffuse, and glare sources reflected in the viewing direction

When glare reflection is absent, the display's CR will appear as in Fig. 15. In the presence of glare, the luminance contribution of the glare source to the total luminance must be added (see Fig. 13). Although the luminance components from useful directional and hemispherical-diffuse reflection will be modulated by the display color Q, luminance components from glare reflection will largely depend only on the display's surface-reflection characteristics. A glossy surface will have a predominantly specular glare reflection, and the glare luminance will be proportional to the glare source luminance. If the display has a matte AG surface, its glare reflection will be predominantly hazy, and the glare luminance from the display would be proportional to the illuminance ES,spec from the glare source incident on the display. This requires specifying not only the luminance of the glare source but also its illumination geometry—that is, the source's solid angle as seen from the display. Fig. 19 compares the predicted ambient indoor CR of electrophoretic reflective displays with either glossy or matte AG surface to a minimum CR criterion of “artificial information” (text, graphics) in the presence of disturbing glare10 for the various ambient illumination conditions listed in Table 2. The analysis for these examples shows that the reflective display with a matte AG surface outperforms the glossy surface when bright glare sources cannot be avoided.

Details are in the caption following the image

Ambient contrast ratio versus luminance of small (1°) and extended (15°) glare sources, simulated for two electrophoretic reflective displays (glossy and matte AG surfaces) in different indoor environments: (a) IEC indoor8 and (b) ISO 9241-307 indoor with ambient “design illuminance.”10

Table 2. Illumination conditions for the ambient contrast simulations shown in Fig. 19.
ES,dir [lx] ES,hemi [lx] LS,spec [cd/m2]
Fig. 19(a) 200 300 10–5,000 IEC 62679-3-18
Fig. 19(b) 0 500 ISO 9241-30710 (Table 167)

In closing, the reflection measurement methodology developed for predicting the ambient contrast and color of both reflective and emissive information displays combines two components: the measurement of display reflection characteristics, and ambient illumination models that represent typical usage environments. The reflection measurement geometries must include the two fundamental ambient illumination components: directional and hemispherical-diffuse. To model handheld display viewing, directional reflection measurements should exclude disturbing reflection. The 45/0 directional source geometry effectively excludes disturbing specular and diffuse haze reflections, which was observed through BRDF measurements on electrophoretic ePaper.

The reflection methodology described in this article employed the concept of linear superposition, where the fundamental illumination components can be combined to model a specific ambient lighting environment. This methodology forms the basis of measurement standards for reflective displays8 and has been extended to reflective displays with integrated front lighting9 and emissive displays in mobile applications. Although there are many lighting environments where displays are used, these standards offer reference indoor and outdoor illumination geometries that can be used as a basis for comparison.

This article summarized the fundamental concepts, measurement best practice, and performance analysis employed by industry and codified in standards. We demonstrated how modern reflection methodologies can be used to evaluate common display treatments, such as matte AG or glossy surfaces. It was shown that disturbing reflections, in particular, are difficult to evaluate due to their overlapping scattering components, and their sensitivity to display structure and measurement geometry. The application of the VAS measurement provides a possible means to decompose the scattering components within the disturbing glare in order to predict its impact on visual performance. But other evaluation concepts, such as the point-spread function19, may also provide valuable information. This highlights the fact that as the multi-layer structures of modern display continue to increase in complexity, new methods will be needed to evaluate their performance.


  • Dirk Hertel, Ph.D., is a principal scientist at E Ink in Billerica, Massachusetts. He can be reached at [email protected].


  • John Penczek, Ph.D., is a senior research associate at the University of Colorado, Boulder, and a guest researcher at the National Institute of Standards and Technology in Boulder, Colorado. He can be reached at [email protected].