The relationship between luminance (i.e., the photometric intensity of light) and its perception (i.e., sensations of lightness or brightness) has long been a puzzle. In addition to the mystery of why these perceptual qualities do not scale with luminance in any simple way, "illusions" such as simultaneous brightness contrast, Mach bands, Craik-O'Brien-Cornsweet edge effects, and the Chubb-Sperling-Solomon illusion have all generated much interest but no generally accepted explanation. The authors review evidence that the full range of this perceptual phenomenology can be rationalized in terms of an empirical theory of vision. The implication of these observations is that perceptions of lightness and brightness are generated according to the probability distributions of the possible sources of luminance values in stimuli that are inevitably ambiguous.A fundamental problem in vision (and perception generally) was recognized at the beginning of the 18th century by George Berkeley (1709Berkeley ( /1975, who pointed out that the sources underlying visual stimuli are unknowable in any direct sense. The light that falls on the eye from any region of a scene conflates the contributions of reflectance, illumination, and transmittance (as well as a host of subsidiary factors that affect these parameters). As a result, the physical provenance of light reaching the eye-and therefore the significance of the stimulus for visually guided behavior-is profoundly uncertain. This fundamental fact presents a biological quandary. Successful behavior in a complex and potentially hostile environment clearly depends on responding appropriately to the physical sources of visual stimuli rather than to the stimuli as such. If, however, the retinal images generated by light cannot uniquely define the underlying reality the observer must deal with, how then does the visual system produce behavior that is generally successful?The purpose of this review is thus to consider evidence, much of it derived from our own experiments over the last few years, about the way the uncertain relationship between the physical world and the perceptual world is resolved by the nervous system (see Purves & Lotto, 2003). The gist of this body of work is that what one sees at any moment appears to be fully determined by the probability distributions of the possible sources of the stimulus rather than the physical qualities of the stimulus (which are ambiguous) or the properties of the objects and conditions that generated the stimulus (which cannot be known directly). Although this framework has several important precedents (see below), it differs from most mainstream neurobiological thinking in recent decades and suggests other ways of conceptualizing the purposes served by the known physiology of visual system circuitry. The context here for exploring the merits of this conception of vision is sensations of lightness and brightness, which are arguably the most fundamental qualities of human visual experience.
Luminance, Brightness, and LightnessLuminance is ...