Visual Performance for Motorists - Road Lighting: Fundamentals, Technology and Application - page 29

(2.0 for the example of Fig. 3.17 ) as the metric for the quality of the road-lighting
installation (Lecocq 1991 ; AFE 2002 ).
In 2006, the roadway lighting committee of IESNA decided to withdraw STV as
a design metric, but to retain it as a comparison or fine-tuning tool for road-lighting
installations that have been designed on the basis of luminance criteria but that can
benefit from further optimization. The concept of total revealing power, discussed in
a previous section (Sect. 3.2.2), offers an alternative for this. With it, unlike as with
small target visibility, different reflectance factors are taken into consideration.

3.3.3 Relative Visual Performance

As has been shown, the starting point for supra-visibility measures is the threshold
contrast of visibility. Measurements of threshold contrast, even when determined
under laboratory conditions, show a large spread, even when determined repeatedly
with the same observer. Apparently, the border between vision and non-vision is not
such a sharp one. Rea and Ouellette ( 1988 ) therefore developed a supra-threshold
measure, called relative visual performance, or RVP, based on reaction times for
detecting objects. Both the contrast thresholds and the level of supra-visibility are
derived from these reaction-time experiments. The authors claim that their contrast
thresholds are therefore more representative than those derived using conventional
methods.
The RVP model was originally developed for indoor lighting applications from
experiments measuring speed and accuracy of reading-tasks (Smith and Rea 1982 ).
The reaction-speed experiments with objects have extended the RVP model for use
in road-lighting applications. RVP, like visibility index, can be calculated from the
contrast and size of the object, the luminance of the background, the glare and the
age of the observer. Appendix B gives the formulas for the calculation of RVP from
the photometric values of road lighting installations for different object sizes and
reflectances, the latter being required for the calculation of the contrast (Rea and
Ouellette 1991 ). The reaction performance tests on which the model is based were
carried out with both positive-contrast and negative-contrast objects. The conclusion
from these tests was that the difference in contrast threshold for positive and neg-
ative contrasts is negligible. This is contrary to the results of Adrian, used for the
calculation of visibility level and small-target visibility. It should be noted, therefore,
that the reaction performance tests for the lower lighting levels were only carried out
with positive-contrast objects.
Visibility level is not linearly proportional to visibility. For example, a visibility
level, VL, of 10 does not correspond to a situation that is twice as visible as one with
a visibility level of 5. Often the level of visibility becomes saturated at VL levels
higher than 10 to 15. This saturation of visibility is not directly apparent from the VL
values. By contrast, RVP values are proportional to the speed and accuracy of the
visual processing. Figure 3.18 shows the results of RVP calculations for one object
size (7 min of visual angle).