Metal halide (MH) lighting systems are gaining in acceptance relative to highpressure sodium (HPS) lighting systems for many night time applications. The present paper describes a series of studies carried out to address some of the 'quality' issues associated with MH and HPS lighting systems. HPS and MH installations were compared in terms of perceptions of brightness and safety, acceptability for social interaction, facial recognition and eyewitness identification. At equal photopic light levels, a street scene illuminated at night by an MH lighting system was reliably seen as brighter and safer than the same scene illuminated by an HPS system. In terms of acceptability for social interaction, facial recognition and many aspects of eyewitness identification, the measured differences between lighting systems were not as clear.

A field study was conducted to extend fundamental findings about visual performance at mesopic light levels to a driving context. Subjects participating in this field study drove a vehicle along a lighted street while performing a high-order decision-making task. Subjects identified the direction of an off-axis target, toward or away from the street, and braked or accelerated, accordingly. Two sets of light sources were compared: a set of ceramic metal halide light sources and a set of high-pressure sodium light sources. The same study was also performed during the day-time. The results demonstrated that both braking and acceleration response times decreased monotonically as unified luminance increased, suggesting that unified luminance is a suitable rectifying variable for characterizing light levels for different light sources with respect to a complex visual task.

This experiment tests the hypothesis that light sources that produce smaller pupil sizes ensure better achromatic visual task performance at the same photopic illuminance. Two groups of subjects, one in the age range 18-28 years and the other in the range 61-78 years, performed a Landolt ring task for eight different gap sizes, two different illuminances, and two lamp scotopic/photopic ratios. For both age groups, pupil size was determined by both illuminance and lamp spectrum. The speed and accuracy of performance of the task was determined by the gap size, and to a much lesser extent, by the illuminance. Lamp spectrum had no effect on the performance of the task.

No abstract

Recent laboratory studies showed that off-axis detec tion is better for metal halide (MH) lamps than for high pressure sodium (HPS) lamps at mesopic light levels and established psychophysical procedures to determine a set of mesopic luminous efficiency functions based on reac tion times. 1 ' 2 The mesopic luminous efficiency functions showed that MH lamps are 15, 54, and 89 percent more efficient than HPS lamps at photopic luminances of 0.3, 0.1, and 0.03 cd/m 2 , respectively. Other laboratory stud ies using simulated driving contexts verified the validity of the fundamental findings, but found the difference in off-axis detection between MH and HPS lamps to be much larger than would be predicted simply by the lumi nous efficiency functions. 3 " 5 Other factors, which were not investigated in the He et al. study, such as a target size and location, interact with spectral power distributions (SPD). In the real world, there are many variables that influence peripheral detection under mesopic condi tions such as oncoming headlight glare, forward head lights, and uniformity of illuminance distribution. Therefore, before applying the fundamental findings for practical roadway lighting it is important to investigate these variables in a real roadway setting.Two field studies-Forward Headlight Study and Oncoming Headlight Glare Study-were carried out to extend the fun damental findings to actual driving situations. The Forward Headlight Study was designed to understand how forward headlight illumination assists driver detection of off-axis targets on a roadway illuminated by fixed pole lighting. In this first study, subjects were asked to drive a car along a fixed track and detect the off-axis targets. The Oncoming Headlight Glare Study investigated how oncoming headlights impair off-axis target detection for drivers. In this second study, subjects sat in a parked car on the track while anoth er car was parked in the opposite lane with and without headlights (i.e., glare to the subjects). ed illuminance on the parking lot and no extraneous light sources were visible to subjects while performing the experiments. In the parking lot, 35 traffic cones demarcated a 300 m-circular track (Figure 1). Thirty cones were used to indicate a 3 m wide driving lane 80 m in length straight along the north end of the parking lot.The study used five temporary streedight luminaires. Each luminaire consisted of two independent units on a single 12 ft pole-one used a high-pressure sodium (HPS) lamp and the other used a metal halide (MH) lamp. A cylinder on the top of each pole held the luminaires; by rotating the cylinder on its axis the positions of the lumi naires could be easily and quickly changed during the study. Both luminaires had attached metal mesh filters to reduce illuminance level to 5-6 Ix, average, on the pave ment so that the illuminance levels met IESNA recom mendations. 6 The details of the luminaires and lamps are given in Table 1. Figure 2 shows the spectral power distri butions of light emitted by the luminaires with ...