The phantom contrast: An experimental (phenomenological-style) research.

A linear gradual luminance gradient can largely modify the lightness of a target region in the direction of a contrast induction (Agostini & Galmonte, 1997; 2002). In the present work we investigated, by using a phenomenological method similar to that used by Kanizsa (1954) to measure his Paradox, a version of the above cited display where the luminance range has been drastically reduced, making the gradient almost unnoticeable. A grey target placed at the centre of this “phantom” gradient appears quite different in lightness from an identical target surrounded by a homogenous luminance surface which value is equal to that of the last luminance of the gradient surrounding the target. The backgrounds appear quite similar, but the target placed at the centre of the gradient undergoes a “phantom” contrast effect. This perceptual paradox suggests that lightness perception is affected also by smooth changes in luminance, even when they are almost under threshold. The observed effect can not be simply the result of the local contrast between the target and its closest surrounding luminance, since narrowing the spatial distribution of an identical reduced range gradual luminance gradient the perceived effect is the opposite. This suggests that the visual system computes surface colours taking in to account the global spatial distribution of luminance gradients.