Audrey H. Gutherie, Ph.D.
Rehabilitation Research & Development Center
Atlanta VA Medical Center
Mark Uslan , M. Ed., M.S.
American Foundation for the Blind (AFB TECH)
Ronald A. Schuchard, Ph.D.
VA Rehabilitation R&D Center of Excellence
Atlanta VA Medical Center
Jack Smith, Ph.D.
Center for Environmental, Geotechnical and Applied Sciences (CEGAS)
Marshall University Research Corporation (MURC)
Advanced Engineering Systems
Mid-Atlantic Technology, Research and Innovation Center (MATRIC)
South Charleston, WV
The use of small visual displays embedded in electronic devices has become commonplace. Despite their widespread use, there are no published conventional standards which govern the characteristics of these small visual displays and persons who are visually impaired often have problems accessing them. This paper describes the characteristics of small visual displays, defines the visual variables that influence their accessibility, and discusses present-day efforts to develop display standards for people who are visually impaired.
The use of small visual displays embedded in electronic devices has become commonplace. Today, these displays are found on many household appliances such as microwaves, washing machines, and alarm clocks. Moreover, they are used in cellular phones, home health monitoring devices (including blood pressure monitors and blood glucose monitors), and are often found on ATMs and office equipment. Despite their widespread use, there are no published conventional standards which govern the characteristics of these small visual displays. Although this is a potential problem for individuals without vision loss, it is especially problematic for individuals with vision loss. As many as 20 million Americans have vision loss that may impact their ability to discern the information on small visual displays. Thus, as the use of these small visual displays becomes more prevalent, the need to standardize their development becomes greater.
There are published conventional standards which address the usability of computer monitors and televisions.1 Moreover, there is a large amount of research devoted to the understanding of how individuals with or without vision loss process the types of information found on visual displays (e.g. alphanumeric characters, objects, icons, etc.).2-5 However, there is a conspicuous lack of research applying what is known from this literature towards the development of small visual displays.
In the last decade, this problem has received much attention from governmental regulatory bodies, universities, and trade organizations. Guidelines for small visual displays, based on expert opinion and information gathered from visually impaired consumers, have been proposed by the Royal National Institute of Blind People (RNIB), the University of Wisconsin's Trace Research & Development Center, the Telecommunications and Electronic and Information Technology Advisory Committee (TEITAC) of the U.S. Access Board, and the Telecommunications Industry Association (TIA). See the section entitled "Further Reading" for more information. For example, RNIB has published several reports on their website,
http://www.tiresias.org, discussing the need for standardization of small visual displays, what is involved in standardization, and their recommendations. Despite these recommendations, there is an obvious lack of scientific research on what features of small visual displays are important for their accessibility.
Small Visual Displays and the Information They Convey
Before determining which characteristics of small visual displays influence their accessibility, it is important to first identify the types of information found on them and the different technologies used for their creation. Typically, the information contained on small visual displays is numbers, letters, short words or phrases, symbols, and/or simple icons. With the exception of cellular phones, small visual displays usually do not have the space to accommodate extended text or complex text menus. These constraints should be taken into consideration when testing their accessibility.
In recent years, a number of competing technologies have been used in the creation of these displays. However, small visual displays are most commonly composed of simple, passive, monochromatic Liquid Crystal Displays (LCDs). Unfortunately, these small LCDs are often the most problematic to see. They are generally comprised of only a few 7-segment alphanumeric characters, require ambient light to be viewed, and have a relatively small degree of image detail. However, they are relatively inexpensive compared to competing LCD technologies such as full 8-bit color Thin-Film Transistors (TFT) active-matrix displays. These displays are more complex than passive-matrix LCDs. They allow for more detailed images and have bright user-controllable backlighting (using either cold-cathode fluorescent lamps or light-emitting diodes). Moreover, TFT displays often have sophisticated coverings and coatings to help reduce glare and improve contrast at large viewing angles.
In addition to LCDs and TFTs, there are several display types currently in development that may become commonplace in the future, such as Interferometric Modulator (IMOD) full-color displays and Surface-conduction Electron-emitter Displays (SEDs). IMODs can create a backlit effect by reflecting ambient light to a high degree, and SEDs purport to have higher levels of brightness and color than any other small screen technology.6,7 To have wide application, display standards should be based on visual variables that underlie all of these technologies.
What Visual Variables Influence the Accessibility of Small Visual Displays?
Whether or not a person can discern a small visual display is contingent upon the resolution, luminance, contrast, spatial frequency, temporal frequency, glare, wavelength, font, and size of that display. However, some of these visual variables have a much larger impact on display accessibility (e.g., contrast and spatial frequency) than others (e.g., color).
Resolution—the ability of a display to produce fine detail. Resolution is a measure of the number of distinct pixels a display can produce in each geographical dimension; thus, a higher resolution allows for a greater amount of detail. Human visual resolution is the ability to resolve detail and is limited mostly by the number and organization of the photoreceptors of the eye. The cone photoreceptors have greater resolution than the rod photoreceptors, which is why (under photopic conditions) humans are better able to resolve fine details in their central vision where the cones are more densely packed. Human visual resolution can be measured with tests of visual acuity, the ability to distinguish discrete elements of a pattern like a letter or a symbol. Visual acuity can decline with age and visual disease; thus, both factors should be considered when determining the optimal resolution for small visual displays.
Luminance—a measure of the amount of light that is reflected from an object or image and is also referred to as brightness. The absolute luminance of an object is not as important as its luminance relative to the luminance of the background on which it is presented, i.e., its luminance contrast.
Contrast or Contrast ratio—the ratio of the luminance of the brightest color to that of the darkest color that the display is capable of producing, often measured by Michelson Contrast. For example, the contrast ratio for a monochromatic, black-and-white display is the ratio of the luminance of the whitest white to the luminance of the blackest black. Typically, objects on a digital display have a lower luminance value (are darker) than the background, e.g. black objects or text on a white background. However, some low vision individuals prefer reverse polarity—white text on a black background. Reverse contrast polarity seems to be especially beneficial to individuals with cloudy ocular media. Regardless of polarity, the contrast ratio of a display significantly affects a person's ability to discern the information on that display and is perhaps the most important characteristic of any visual display. It is commonly thought that the higher the contrast ratio the better a person is able to see a visual display. However, a person's sensitivity to contrast (contrast sensitivity) significantly varies with spatial frequency.
Spatial Frequency—a measure of the number of line pairs (e.g. a dark line against a light background) in a given area, measured in cycles per degree (c/deg) of visual angle. A higher spatial frequency would correspond to a display with a greater number of smaller lines (characters), whereas a low spatial frequency would indicate a fewer number of larger lines within the same amount of space. The human contrast sensitivity function (CSF), which is a plot of contrast sensitivity versus spatial frequency, reveals that a person's contrast sensitivity varies significantly with spatial frequency. Past research has found that individuals without vision loss are most sensitive to intermediate spatial frequencies (e.g. 4 c/deg) and that sensitivity drops off at lower and higher spatial frequencies.8 The human CSF is often measured using gratings of light and dark bars (in either sine-wave or square-wave patterns). See Figure 1 for an illustration of square-wave gratings of varying spatial frequency and contrast.
a) High contrast, high spatial frequency square-wave grating
b) High contrast, low spatial frequency square-wave grating
c) Low contrast, low spatial frequency square-wave grating
The relationship that exists between the spatial frequency and contrast of a display is expressed in the display's modulation transfer function (MTF), a plot which shows the correlation between the two display variables. Generally, as spatial frequency increases it becomes more difficult for the display to create a sharp edge, leading to a blurrier line and a lower contrast value. Using the MTF of a display, it is possible to determine what levels of contrast the display is able to create for any spatial frequency.
Temporal Frequency—how rapidly or slowly a stimulus or image changes across time. Temporal frequency is measured in cycles per second or hertz (Hz). A digital display's temporal frequency refers to the refresh rate of that display. A good guideline for the refresh rate of small visual displays should be based on carefully controlled studies applying what is known regarding human critical flicker frequency (CFF)—the fastest flicker rate a person can perceive. The refresh rate of large visual displays such as computer monitors and television screens are typically set well above the average human CFF (e.g. 60 Hz).
Glare—produced by light(s) which are brighter than the light to which the eye is adapted (e.g. headlights at night). Glare can significantly affect a person's contrast sensitivity and CFF. Thus, glare can pose a significant barrier to seeing information on a small visual display. The effects of glare can have a greater impact for individuals with ocular and macular diseases. For example, individuals with cloudy ocular media have increased light scatter in their eyes, which is significantly exacerbated by the effects of glare. Glare can be experienced on a small digital display if a bright ambient light is reflected off its surface. It can be reduced by altering the angle and material of the display surface.
Wavelength—the wavelength of the light emitted from a display determines its color. Typically, people perceive shorter wavelengths of light as bluish (e.g. 470 nm), middle wavelengths as greenish (e.g. 520 nm), and longer wavelengths as reddish (e.g. 650 nm). The visual system's sensitivity to different wavelengths of light is not uniform, but generally, people are most responsive to middle wavelengths.
Font—the size and the style of the characters displayed. Past studies have demonstrated that font style significantly affects the accessibility of alphanumeric information.9 Some of the differences between font styles include: segmented vs. pixelated characters, line width, serifs (i.e., fine lines finishing off the main strokes of a character), and the spacing between the characters. Different fonts have different line widths, and not all fonts have serifs. Moreover, some fonts are proportionally spaced (e.g. Adobe's Times-Roman) while others have a fixed width between letters (e.g. Adobe's Courier-Bold). Each of these factors contributes to display resolution (discussed above). Thus, user visual acuity should be a major determining factor for which font size and style to use on small visual displays.
Display Size—The size of a display significantly affects the density of the information that is contained on the display. Information density is an important consideration in determining the accessibility of a display, as a very low value would allow for too much space between characters, and a high value would lead to a cluttered display, which would make identification of specific characters difficult. However, one important characteristic of the displays we have been discussing so far is that they are small, which allows them to be placed on very small devices such as watches and hand-held blood glucose monitors. The need to maintain such versatility should be addressed when creating standards for small visual displays. Although a larger display size would allow for greater spacing of the characters or objects on the screen (i.e., a lower information density), it would significantly decrease the versatility of such displays.
To adequately address the issue of display size, research should be conducted on the minimal size necessary for small visual displays. The minimal display size would depend on the nature of the information that is to be displayed (e.g. single characters or symbols versus words, sentences, etc.) and each of the visual characteristics discussed above.
Efforts to Develop Standards for Small Visual Displays
The American Foundation for the Blind (AFB TECH) and the Atlanta VA Rehab R&D Center on Excellence for Aging Veterans with Vision Loss are collaborating on a research project to determine the efficacy of Barten's image quality metric, the SQRI, as the basis of a standard for small visual displays.10,11 Over a decade ago Barten developed the SQRI as a mathematical model to evaluate the quality of visual displays for individual observers. The SQRI is computed from a display's MTF and the individual's Contrast Sensitivity Function (CSF). Although the ability of this equation to predict image quality on large digital displays (e.g., computer monitors and projector screens) has been tested with individuals without vision loss, it has not been tested on small visual displays or with low-vision populations. Thus, the goal of the AFB/Atlanta VA project is to determine how well the SQRI can predict the accessibility of small visual displays by low vision users. The results of this project will help provide the foundation upon which to build design standards for small visual displays.
Once a standard is created, it will be necessary to routinely measure the key visual parameters in these displays. Work at AFB has been directed to that end, and an optical lab has been established to perform these measurements. At the core of this lab is a computer-controlled digital camera with a macro lens, an integrated light source, and image analysis software capable of measuring a variety of optical properties with about a 10-micron resolution, or almost 5 times the spatial resolution of the human eye at a 12" viewing distance. The lab is currently focused on hand-held devices with small LCD displays. Specifically, the AFB Optics Lab is investigating the contrast modulation (amplitude and spatial frequency) required for Barten's SQRI metric. Since most application-specific small LCD displays cannot be driven to display the special grid patterns traditionally used to measure contrast modulation, simpler measures have been used. Results of measures of 9 home blood pressure monitors that use passive monochromatic LCD displays were published in 2007.12 For displays that use more complicated fonts and/or graphics, a mathematical analysis (Fourier Transform) will be needed to determine the dominant spatial frequencies.8 In order to effectively deal with more complex displays and to measure variables besides reflective luminance, one needs to control and account for the interaction of multiple light sources (including ambient lighting), different angles of illumination and viewing, etc. These capabilities are currently under development. Details and continuing progress of the lab can be found at the AFB TECH LCD-MTF blog.
There are no published conventional standards for the development of small visual displays, and low vision individuals often have problems accessing such displays. AFB TECH and the Atlanta VA Rehab R&D CoE have formed a synergistic relationship in order to address this problem. It is hoped that the creation of standards for the development of small visual displays and the ability to benchmark them will lead to displays that are more accessible to the low-vision population.
The authors would like to acknowledge the assistance provided by Marshall University research intern Morgan Blubaugh.
- Video Electronics Standards Association Display Metrology Committee (2001). Flat-panel display measurements standard. Version 2.0. Milpitas (CA): Report No.: VESA-2001-6
2. Avidan, G., Harel, M., Hendler, T., Ben-Bashat, D., Zohary, E., & Malach, R. (2002). Contrast sensitivity in human visual areas and its relationship to object recognition. Journal of Neurophysiology, 87, 3102-3116.
- Braje, W. L., Tjan, B. S., & Legge, G. E. (1995). Human efficiency for recognizing and detecting low-pass filtered objects. Vision Research, 35(21), 2955-2966.
- Rubin, G.S. & Legge, G. E. (1989). Psychophysics of reading. VI—The role of contrast in low vision. Vision Research, 29(1), 79-91.
- Patching, G. R. & Jordan, T. R. (2005). Spatial frequency sensitivity differences between adults of good and poor reading ability. IOVS, 46(6), 2219-24.
- "Interferometric Modulator (IMOD) Technology Overview". Qualcom (May 2008). Retrieved on 2008-08-07. http://www.qualcomm.com/common/documents/
- "SED Surface-conduction Electron-emitter Display Technology." Retrieved on 2008-08-17. http://zainlcd.wordpress.com/2008/08/15/sed-surface-conduction-electron-emitter-display-technology/
- Campbell, F. W. & Robson, J. G. (1968). Application of Fourier analysis to the visibility of gratings. Journal of Physiology, 197, 551-566.
- Mansfield, J. S., Legge, G. E., & Bane, M. C. (1996). Psychophysics of reading: XV. Font effects in normal and low vision. IOVS, 37(8), 1492-1501.
- Barten, P. G. (1990). Evaluation of subjective image quality with the square-root integral model. Journal of the Optical Society of America, A, 7(10), 2024-2031.
- Schuchard, R., Uslan, M., & Wilson, T. (2008, July). SQRI image quality metric as a measure to predict character recognition for small visual displays. Poster session presented at biennial Vision conference, Montreal, Canada.
- Uslan, M. M., Burton, D. M., Wilson, T. E., Taylor, S., Chertow, B. S., & Terry, J. E. (2007). Accessibility of home blood pressure monitors for blind and visually impaired people. Journal of Diabetes Science and Technology, 1(2), 218-227.
Website for the Scientific Research Unit (SRU) of the RNIB Access and Innovation Group:
Consumer product guidelines project: Trace R&D Center, Madison, WI trace.wisc.edu/docs/consumer_product_guidelines/consumer.htm
TIA website: www.tiaonline.org
TEITAC recommendations for small visual displays: www.access-board.gov/guidelines-and-standards/communications-and-it/about-the-ict-refresh/background/teitac-report/6-the-recommendations
Blindness statistics from the American Foundation for the Blind: www.afb.org/Section.asp?SectionID=15&DocumentID=4398#numbers
AFB TECH LCD-MTF blog: afb-lcd.blogspot.com