Traditionally, the lighting engineering community has emphasized illuminance, the amount of light reaching a surface, as the primary design goal. The Illuminating Engineering Society (IES) provides tables of illuminances for different types of tasks which lighting engineers consult in designing lighting systems [Kaufman8l]. Illuminance has proven to be a popular metric because it corresponds closely to the amount of energy needed to light a building as well as the initial cost of the lighting system. Perhaps more importantly, illuminance is easy to calculate, especially in simple unobstructed spaces with direct lighting. However, illuminance is not well correlated with visual performance, which is the real reason for installing a lighting system in the first place. Visual performance is a psychophysiological quantity that has been tied to physical quantities such as contrast, size and adaptation level by subject experiments [Cobb28] [Rea86]. These physical quantities can be approximated from illuminance using a host of assumptions about the environment, or derived directly from the distribution of luminance. Luminance is the quantity of light traveling through a point in a certain direction, and it is this quantity that the eye actually "sees". However, the difficulty of calculating luminance for common tasks has made it an unpopular metric. Despite its importance to lighting design, luminance is rarely used because there is a lack of the necessary computational tools. In this paper, we will demonstrate a computer calculation of luminance that has significant advantages for lighting design. As well as providing an immediate evaluation of visual quality for task performance, less quantifiable factors such as aesthetics can be studied in synthetic images produced by the program.