Abstract

Metelli (1974, 1985) proposed a model of the intensity relations in perceptual transparency based on a physical model of transparency involving additive color mixture. Beck et al. (1984) reported four findings which suggest that a change in point of view is required. First, they showed that the relevant variable for the perception of transparency is the psychological variable of lightness and not the physical variables of luminance or reflectance. Second, they showed that physical models of transparency based on either additive or subtractive color mixture entail four constraints. However, only violations of constraints (i) and (ii) adversely affect judgments of perceptual transparency; violations of constraints (iii) and (iv) do not. Nonveridical perceptions of transparency occur when constraints (iii) and (iv) are violated. Third, there are simple and strong violations of constraints (i) and (ii). Transparency judgments, though reduced in number, still occur for simple violations of constraints (i) and (ii). Transparency judgments occur infrequently for strong violations of constraints (i) and (ii). Physical models of transparency fail to differentiate between simple and strong violations. Fourth, it is necessary .to distinguish between perceptual transparency occurring with and without constancy. Constancy occurs when the color seen through the transparent surface is seen to be the same as the color of the surface not overlaid by the transparent surface.

© 1987 Optical Society of America