How people see and are psychologically impacted by lighting has been a subject of much study and discussion for years. Describing light as "lumen output" and measuring it as "foot candles" on a work plane have been the traditional ways of describing and defining how much light is required to perform a variety of tasks. However, that is being re-examined based on results of studies on visual performance and the psychological impacts of lighting. Additionally, the "color rendering index" (CRI) and correlated color temperature (CCT) describe the quality of the light (relating to how true colors appear compared to under a noon north sky on a clear day). As lighting technology evolves into various types and colors, simply measuring the lumens proves not to be fully adequate in predicting how well people can see. An excellent example is the low-sodium lamp which produces many lumens, but only two colors (yellow and gray); the ability to make out details—beyond shapes of objects—is lost under this light source. Different light sources produce light in different spectral ranges and there is a wide variety of spectral output available in fluorescent lamps.
Vision itself is affected by many factors, from light intensity, distribution, color, and contrast, as well as reflections, glare, air quality, motion of subjects and viewers, and more. Our eyes use different parts to see in bright light and low light conditions. The eye contains cones and rods which were thought to work in opposite conditions. Cones provide color vision and fine detail (photopic) in bright light and rods take over in dim light (scotopic). In bright light our pupils contract allowing more detail to be perceived, while depth of field and perceived brightness also increase. In low light our eyes dilate to allow more light in.
Light meters and recommended light levels for tasks have traditionally been calibrated for daytime viewing, and general interior lighting, based on the photopic response. However, studies are indicating that the scotopic vision is more involved in interior lighting than thought, and affects pupil size. At recent conferences, some presenters encouraged designers to specify the photopic/scotopic (P/S) ratio of lamps when selecting them in order to get better design, efficiency, and better vision for occupants.
Sam Berman—formerly with the Lighting Systems Research Group at Lawrence Berkeley Laboratory and a major supporter of the importance of the P/S ratio in lighting selection—developed a conversion factor that applies the P/S ratio to lumen output of various light sources, and then expresses the effective lumens the eye will perceive for vision based on the size of the pupil and the effect on vision (see Table 1 below). Some lamps, like low-pressure sodium, lose most of their output using this method, while others like high-quality fluorescent lamps gain substantially.
Induction lamps are basically equivalent to high-quality fluorescent lamps with a CRI of 80 and a color temperature of 4100K (T-8 in the table below). Berman’s table suggests that, while the T-8 4100 lamp has rated lumens of 90 per watt, the pupil (effective) lumens are actually 145 per watt. If contrast and distribution are controlled, this suggests that fewer watts are needed to provide good vision than rated lumen output would suggest, meaning energy savings will result.

Table 1. Conversion factors for lumens to pupil lumens
Correction factors applied to conventional values of lumens per watt yield a value for pupil lumens per watt, which is a measure of how effectively the eye sees the light that is emitted. The pupil is more receptive to light at the blue end of the spectrum./p>

Light source	Conventional lumens per watt	Correction factor (P/S ratio)	Pupil lumens per watt
Low-pressure sodium	165	0.38	63
5,000-K T5 fluorescent	104	1.83	190
4,100-K T8 fluorescent	90	1.62	145
Clear metal halide	85	1.49	126
5000-K LVD Lamp	80	1.62	129
5,000-K pure triphosphor fluorescent	70	1.58	111
3,500-K triphosphor fluorescent	69	1.24	85
50-watt high-pressure sodium	65	0.76	49
2,900-K warm white fluorescent	65	0.98	64
Daylight fluorescent	55	1.72	95
35-watt high-pressure sodium	55	0.57	31
5,000-I 90 CRI fluorescent	46	1.7	78
Vitalite fluorescent	46	1.71	79
Deluxe mercury vapor	40	0.86	34
Standard incandescent	15	1.26	19
Tungsten halogen	22	1.32	29

Summary

Recent studies seem to favor white light (as from induction lamps) for viewing moving objects in low-light conditions, such as spotting a pedestrian, animal, or other moving object off to the side of the roadway at night. Some cities opt to use white light rather than the yellowish light of high-pressure sodium (even though the price is higher) in hopes of reducing accidents. The improved color rendering of white light in retail areas and places where people congregate after dark make it a popular choice for street lighting in downtown areas.
Induction lamps produce high quality white light. More lamps and fixtures are becoming available, but they do carry a premium price and still have limited applications. The long life of these lamps can substantially reduce maintenance costs due to relamping.
White light is proving to have advantages for visual performance. Current codes and standards are based on measurements that do not address the impact of pupil lumens, and pupil lumens can be quite different from traditionally measured lumen output of lamps. Studies on the relevance of light spectrum and the mechanics of vision are ongoing, and codes and standards may reflect that in the future.