Visual Performance: A Comparison in Terms of Detection of Presence and Discrimination of Detail

A great deal of information has been obtained by many investigators on the threshold relationships between contrast and luminance. These all follow the same general pattern: smaller contrasts can be seen with higher luminances. However, differences in experimental procedures, criteria, test targets and observers make direct comparison difficult. Such differences have led some to express concern that the basic relationships for one type of test object may not be applicable to others. They feel this may indicate inappropriate lighting recommendations for many practical tasks which bear little resemblance to the object used for the standard visual performance curve used by the IES.

In his 1959 paper, Blackwell did point out that errors would be introduced into the lighting specification method due to non-parallelism of threshold curves obtained with various visual tasks. He presented data illustrating the differences obtained for the detection threshold with crosses, rectangles and circular disc targets. However, he suggested that differences in the form of the curves can be minimized by evaluating a task at a luminance level as close as possible to the one specified by his system Furthermore, the curve for the four-minute disc, which was selected as the standard, was roughly of intermediate form with respect to the others. Since all of his data are in terms of the detection of the presence of the target, we still have the question of possible differences when discrimination of detail is involved.

For present purposes our primary concern is with the shape of the contrast-luminance relationship when the criteria of detection of presence and discrimination of detail are employed. That is, are the differences in the shapes of the curves caused by the criteria and types of test objects, or by other factors? Thus, the objective of the present investigation was to compare threshold data for a disc target with similar data for two objects - a parallel bar and a Landolt ring - which involve discrimination of detail. It is obvious that many other types of targets could be used, but these should indication what might be expected. Furthermore, these two objects also lend themselves to observations using the criterion of detection of presence, and thus permit an additional comparison.