In The Dark About The Lumen
The present definition of the lumen is based on experiments performed in the 1920s and may be inaccurate for nighttime tasks such as driving, writes guest author Mark Rea. A redefined "nighttime lumen" could force us to reconsider the efficacies of light sources for street lighting, leading away from high-pressure sodium lamps towards sources such as metal-halide lamps.
In The Dark About The Lumen
Light sources can be compared using several criteria including price, life, beam distribution, color rendering index (CRI), and correlated color temperature (CCT). One important criterion is efficacy (lumens per watt), or how much light is emitted per unit of electrical power. Government agencies are particularly interested in this criterion. Among the methods used by many agencies to foster the utilization of more energy-efficient light sources are legislation and direct subsidies to building owners, as well as funding for research and development. And efficacy is often the sole criterion upon which the evaluation and promotion of these light sources are based.
Although both the lumen and the watt are physical quantities, the former was defined based on human experiments conducted in the 1920s that were designed to determine the spectral sensitivity of the human visual system to light. A spectral sensitivity function quantifies the efficiency of radiant energy, wavelength by wavelength, in stimulating a visual response.
Data collected in these early experiments were used in defining the lumen, which is based on a spectral sensitivity function for the fovea under high photopic light levels. The fovea is that part of the retina containing only cone photoreceptors and is responsible for central, high-acuity vision. Thus, the efficacy value published for every electric light source manufactured represents the ability of that lamp to produce light for the fovea under high light levels.
For most commercial applications, such as office tasks, this definition of light is adequate and appropriate because most of these tasks require good foveal vision and are performed under relatively high light levels. However, it should be stressed that not every task we perform is dependent on daytime foveal vision (Table 1). Driving an automobile at night, for example, requires peripheral vision under low light levels. If the reader doubts the importance of peripheral vision for driving, try driving an automobile while looking only through a small tube such as a rolled-up piece of paper. The tube allows good foveal vision but excludes peripheral vision. (This demonstration is extremely dangerous, so do it only briefly and only on a clear road!)
NIGHTTIME SPECTRAL SENSITIVITY
As it turns out, the spectral sensitivity of the peripheral retina under low light levels is quite different from that of the fovea under high light levels. Yet, again, the efficacy of every light source, irrespective of its application or the visual task, is based upon the (photopic) lumen. Metal-halide (MH) lamps, for example, are approximately 16% less efficacious than high-pressure sodium (HPS) lamps, based on the (photopic, "normal") lumen. In other words, MH lamps produce approximately 16% fewer lumens per watt than HPS lamps. For photopic foveal tasks then, about 19% more energy is required from a MH lamp than from a HPS lamp to produce the same visual effect.
At very low light levels, when performing peripheral tasks requiring only the rod photoreceptors, the opposite is true. Under these so-called scotopic conditions a MH lamp is 125% more effective than a HPS lamp at producing the same visual effect!Table 2 compares some common light sources in terms of their photopic and scotopic efficacies. (These values have nothing to do with the pupil lumen research being pursued at Lawrence Berkeley National Laboratory in California. The values presented here are independent of pupil size. See IAEEL Newsletter 2/92: Tuning the Light Spectrum To Improve Energy Efficiency
Before everyone runs off to change every HPS street light, however, it is important to point out that there is a gradual shift in spectral sensitivity from daytime foveal vision to nighttime peripheral vision. (Despite what the horizontal lines in Figure 1 might imply, the transition from one level to another is not abrupt, nor are the transition points precisely known.) Below a level of approximately 10 lx on a surface such as a roadway or parking lot, the human visual system will begin to have a mesopic spectral sensitivity. This sensitivity is a hybrid between that of the cones and the rods in the periphery. Under high mesopic conditions MH may be only slightly more efficacious than HPS. As the illuminance drops further, the gap between the effectiveness of HPS and that of MH widens until, eventually, only rods are active and scotopic conditions apply.
FINDING A "NIGHT LIGHT"
One purpose of the research being conducted in our laboratory at the Lighting Research Center (LRC) is to quantify this change in spectral sensitivity under mesopic conditions. We hope that this research will help technical societies and government agencies to recommend and promote the most effective light sources for nighttime applications such as roadway, parking lot and security lighting. HPS will definitely not be the preferred light source. Whether new light sources will be developed remains to be seen. The LRC is certainly in no position to develop a new light source. We are, however, very optimistic that in the very near future a special prototype "night light" that maximizes the energy effectiveness of light for nighttime applications will be produced and demonstrated.
Finally, it is worth noting that achieving a proper roadway lighting design is more complicated than simply retrofitting one light source for another on existing poles. There is much that we do not understand about the visual tasks during driving. It is not clear, for example, how the glare from oncoming headlights affects sensitivity in the peripheral visual field. Does sensitivity remain mesopic or does it become photopic in response to the glare? Is mesopic sensitivity localized in one area of the retina or extended throughout the visual field? These and other important questions confound our ability to design roadway lighting that is energy efficient and responsive to human needs.
For comments see "Letters"
Mark S. Rea
Mark S Rea is the director of the Lighting Research Center
Rensselaer Polytechnic Institute,
Troy, NY, 12180, USA
Fax: +1 518 276 2999
| Light source
| Photopic lm/W
| Scotopic lm/W
|
| Incandescent
| 14.7 (1.00)
| 20.3 (1.00)
|
| High Pressure Sodium
| 127.0 (8.64)
| 80.6 (3.97)
|
| Metal Halide
| 107.0 (7.28)
| 181.0 (8.92)
|
| "Full Spectrum" Fluorescent | 54.5 (3.71) | 108.0 (5.34) |