M.J. Ouellette is a senior technical officer in the Building Performance Laboratory of the Institute for Research in Construction.

Originally published in "Progressive Architecture" July 1993, p. 39-42

"There was no information anywhere about anything. It was totally uncoordinated," complained one survivor of the World Trade Centre bombing in 1993, in New York. She was but one of the tens of thousands of terrified and disoriented occupants who groped through the smoke and darkness, according to the New York Times. Their lives were in danger and they could not find their way out. The evacuation of these people would have been greatly facilitated if only the exit signs had been visible. The scenario is common. Only 8% of the survivors from 400 different fires remembered seeing exit signs during their escapes, according to J.L. Bryan of the University of Maryland.¹ Of course, post-incident studies of this kind are limited by the respondents' recollection skills. Nevertheless, some signs might have been extinguished by power failure, as happened in the World Trade Centre disaster. Some might have been lit, but were not sufficiently visible. Others still might have been visible but not conspicuous enough to stand out from their backgrounds, especially during the distractions of an emergency.

One might ask, "but if the signs conform with local building and fire codes, then surely they will be sufficiently visible and conspicuous during emergencies, right?" But the codes vary considerably from one jurisdiction to another. For example, in 1978, the British Standards Institution said that exit sign luminance shall be between 2 and 80 cd/m². On the other hand, the 1991 Life Safety Code of the U.S. National Fire Protection Association (NFPA) indirectly specifies a minimum luminance (roughly 14 cd/m²) but no maximum luminance. In Canada, the 1990 National Building Code specifies neither minimum nor maximum luminance, but it does say that signs should be illuminated and visible from the exit approach. Inconsistencies among the codes suggest that more needs to be understood about exit sign illumination and visibility, especially in conditions of smoke, when they are needed the most.

At the suggestion of the Canadian General Standards Board, NRC's Institute for Research in Construction began a study of the factors that affect exit sign visibility in order to develop recommendations that would help ensure that building occupants need never grope aimlessly when their lives are threatened by fire. Some of the results are summarized below and are discussed in the context of other relevant research in the field.

Figure 1. Exit sign visible in smoke-free fire

Figure 2. Exit sign nearly obscured in smoke. Visibility is greatly reduced by the glare from the overhead luminaire

Figure 3. Visibility of 13 different exit signs in smoke. Visibility is expressed as the minimum quantity of smoke needed to obliterate the sign. The figure shows that, in general, brighter signs tend to be more visible than dimmer ones. Variability in the data implies that other factors, such as lettering height, spacing between characters, contrast and glare, also come into play. Sign visibility is consistently greater in darkness compared to when the room lights are at emergency lighting levels.

The Effect of Smoke

The reason is that code requirements for exit signs are written for conditions of clear air. They imply that no one will experience smoke if the building conforms to the other requirements in the codes. These might include ventilation requirements and limits on the smoke-generating potential of materials.

Some argue that because smoke builds up so quickly in fires, we cannot possibly expect anyone to be alive to read the signs just minutes after a fire starts. So, they contend, "why bother making signs visible in smoke?" There is some logic to this reasoning, at least for the rooms where the fire starts. On the other hand, smoke growth in neighbouring rooms and exit routes throughout the building may not necessarily be as rapid.

The reality is that people do encounter smoke. Firefighters and rescuers are obvious examples. They are often unfamiliar with the building and they require as many visual cues as possible for orientation. Some say that an extra metre or two of exit sign visibility can sometimes make a large difference for these people - and the people they rescue.

Bryan reported that 53% of survivors from fires in institutions admitted to travelling through smoke. Similar findings were reported by P.G. Wood in the United Kingdom. Some people went only short distances through the smoke, but many of them advanced farther than they could see.³ Knowing that people do navigate through smoke, we have a responsibility to provide evacuation systems that are visible in smoke, wherever feasible. Exit signs are essential components of evacuation systems. Our research identifies simple ways for making exit signs more visible in smoke.

The Effect of Sign Brightness

In a follow-up experiment, we measured people's preferences for sign luminance in smoke. In a modest amount of smoke, more than 80% of our subjects preferred the highest luminance available (1200 cd/m²), regardless of whether the room lights were on, off, or at emergency lighting levels. After viewing the same sign at luminances ranging from 11 to 1200 cd/m², over 73% of our subjects reported that there exists no upper optimum level for sign luminance: the brighter, the better.

In clear air, the same subjects gave very different results. They appeared to have been bothered by glare from the brightest signs. Over 92% of them responded that there was, indeed, an optimum sign luminance, and that it was somewhere between 70 and 700 cd/m². This range challenges the upper limit of the U.K. code.

This presents a problem for the design community. In normal conditions, signs should neither be so bright that they cause visual discomfort nor should they clash with the architectural design. In smoke, however, they should be as bright as possible if they are to be visible. Bi-functional signs are now available which meet both constraints. They brighten and optionally flash upon a signal from the building's alarm system or from a built-in smoke detector.

The Effect of Ambient Illumination

Other light sources near the sign or between the observer and the sign will tend to obscure the sign by scattering the light. This includes not only the overhead lighting systems, but also unit emergency floodlights, such as those attached to exit signs themselves. It even includes the light-emitting apertures punched into the undersides of some signs in attempts to light the doorways below. All of these sources of illumination can degrade the visibility of a sign through smoke.

Glare from competing light sources can also compromise exit sign visibility in smoke-free conditions. This is especially true with sunlight and daylight. It is therefore important that evacuation routes be planned at the early design stages to ensure successful integration with windows, atria, skylights and luminaires. (See Figure 2)

Light Backgrounds vs. Dark Backgrounds

Figure 4. Three different types of illumimated exit sign. The top one, with its dark opaque background, is slightly more visible in smoke than the other two. The bottom one is the least visible in smoke.
Top: red lettering on black background
Middle: black lettering on red background
Bottom: red lettering on white background

Sign Colour: Red vs. Green

The human eye is most sensitive to the yellow-green region of the spectrum. Therefore, it takes less energy to produce green light than it does to produce red light of the same brightness. But this alone does not justify selecting green signs. What we see depends not only on the spectral efficiency of the eye but also the spectral energy of the light source and the spectral properties of all the media between the source and the eye. Incandescent lamps, the most common exit sign illuminants, produce much more red energy than they do green. This tends to compensate for the limitations of the human eye in the red end of the spectrum.

In our laboratory, we conducted an experiment in clear air and in several densities of smoke. Twelve normal-visioned subjects observed red and green signs. The signs were of exactly the same luminance and were identical except for their colours. We found no significant overall effects of colour: seven subjects did slightly better with red signs, while the remaining five showed a marginal advantage with green signs. We conclude that the effect of colour on visibility, if any, is much less than that of sign brightness, ambient illumination, or smoke density.

Colour might, however, play an important role in helping to ensure that the sign is noticed in the first place. It should contrast in colour from the surface it is mounted on. Likewise, the legend should contrast in colour and in brightness from its immediate background.

Some believe that North Americans overuse red in signage and that people might dismiss red signs as part of the visual clutter of urban spaces. Others argue that we are more likely to notice red exit signs because that is what we are conditioned to see. Yet others disagree, saying we might interpret red to mean danger ("stop") rather than the permission ("go") that we usually associate with green. More research is required to settle the debate. Meanwhile, I admit to particularly noticing the wild purple exit signs used throughout one building in Orlando, Florida.

Legends and Arrows

These signs might suffice for people with normal vision when viewed from a short distance in conditions of clear air. But knowing that our visual skills decline as we age, we cannot assume that this kind of sign will be suitable for older people and those with visual impairments. Regardless of our visual skills, we can expect these signs to be less visible than well-designed ones, especially when viewed from distances. We can also expect them to become obscured sooner as smoke accumulates.

Some say that pictograms might offer a more reliable alternative to text, especially to those who speak a foreign language. B.L. Collins and N.D. Lerner, of the U.S. National Institute of Science and Technology, studied a wide range of pictograms and observed that none is completely immune to misinterpretation.⁵

Recently, there has been considerable interest in the United States in determining the optimum configuration for arrows. Collins concludes that the chevron style of arrow is the most visible and that it should be at least 60 mm (21/4 in.) high to be visible from 30 m (100 ft.).

Sign Placement: High vs. Low

F.R.S. Clark warns, however, that smoke density is not always lowest near the floor.⁶ He explains that as fire gases cool, smoke sinks and eventually fills the space. He also reports that smoke can be forced evenly through rooms by forced ventilation systems and that it can be pushed downwards by the action of water sprinklers. In these cases, low signs may offer little advantage. In addition, they might suffer from increased risk of wear and tear as well as blockage by people and movable objects at floor level.

While there is merit to low signs, they are usually recommended as supplements rather than alternatives to conventionally placed overhead ones. Supplemental low signs are now required in some jurisdictions, such as California.

Reliability of Exit Signs

The results indicate that the more frequent the exit sign maintenance checks, the greater the probability that any given sign will pass the inspection. For example, without a routine inspection program, there is no more than a 45% chance that at least 90% of the signs will be working properly. However, if inspections are performed at least once a month, the chances increase to 75% that at least 90% of the signs will pass.

There was also an indication that long-life sources improve reliability. For example, the probability was less than 62% that at least 90% of the incandescent exit signs would pass the next maintenance check; with long-life L.E.D. sources, the chances increased to 77%. Fluorescent sources fell somewhere in between. Reliability was lowest (56% probability that 90% of the signs will pass) when many different types of sign were in use. This highlights the complexities of maintaining incompatible products. Likewise, reliability decreased as the number of signs in the installation increased.

To architects, this means that identical signs should be used throughout the project, wherever possible, especially in large installations. This not only simplifies maintenance, but it increases the chances that the signs will operate when needed. As a bonus, it assists occupants by providing continuity throughout their routes.

The research also suggests that architects should encourage their clients to keep a suitable maintenance schedule for the signs, and indeed all key components of the building, to ensure that the building continues to function as safely, efficiently and effectively as originally conceived.

Impact on Design

The solution is to consider signage from the very beginning of the project. This helps ensure that sufficient resources remain available for achieving the necessary balance between local code requirements, client needs and occupant safety. It also helps ensure that the evacuation system integrates effectively with all aspects of the building design (see Figure 5 and Figure 6).

Figure 5 and Figure 6 (close-up). An exit sign, partially concealed within an alcove formed by a new wall featuring the bulletin board. The photograph was taken in a jurisdiction that permits the placement of signs to the left of doors, as in this example. The narrow, obstructed, ill-defined exit route is a problem. On the other hand, the high brightness of the sign would boost visibility in smoke. However, light from the fluorescent luminaire directly above the sign would greatly reduce the visibility of the sign in smoke. Likewise, the large white luminous background of the sign would tend to bleach out the legend, in smoke. The close-up view of the sign reveals that the lettering height and the spacing between characters were compromised at the expense of visibility, especially when viewed from a distance, by a visually impaired person, or through smoke.

Some might feel that the architect's responsibilities end when the signs conform to the codes. But the reality is that the codes have not completely caught up with the latest research results on emergency egress. Architects should ideally consult this information and reach beyond the codes to provide the safest possible building interiors within their means.

After all, what greater professional responsibility does an architect have than to ensure the safety of building occupants?

References

1. Bryan, J.L. Implications for codes and behavior models from analysis of behavior response patterns as selected from Project People and Project People II Study Programs. Report to the U.S. Department of Commerce and the U.S. Department of Health and Human Services, No. BS-GCR-83-425, 216 pp., March 1983.

2. Rea, M.S., Clark F.R.S., and M.J. Ouellette. Photometric and psychophysical measurements of exit signs through smoke. Division of Building Research, National Research Council of Canada, DBR Paper No. 1291, (NRCC 24627) 42 pp., June 1985.

3. Wood, P.G. The behaviour of people in fires. Fire Research Note, No. 953, Building Research Establishment, Fire Research Station, Borehamwood, U.K., 113 pp., November 1972.

4. Ouellette, M.J. Exit signs in smoke: design parameters for greater visibility. Lighting Research & Technology, Vol. 20, No.4, pp.155-160, 1988.

5. Collins B.L. and N.D. Lerner. An evaluation of exit symbol visibility. National Bureau of Standards. Report of the U.S. Department of Commerce, No. NBS 83-2675, 52 pp., March 1983.

6. Clark, F.R.S. Strategies for improving visibility in fires. Institute for Research in Construction, National Research Council of Canada, Canadian Building Digest 246, January 1988, 4pp.

7. Monroe, L. New directions for signage. Buildings, Vol. 83, No. 3, pp.72-74, March 1989.