Abstract
A sinusoidal luminance grating can be perceived as a surface with a series of depth corrugations. Chromatic information can alter the perceived depth of the corrugations: Kingdom's ‘colour-shading effect’. He found that a chromatic sinusoid, orthogonal to a luminance sinusoid, enhances the perceived depth of the luminance-defined corrugation, whilst aligned chromatic sinusoids suppress the depth. One reason put forward for this is that, in the natural world, pure luminance change is likely due to shadow or shading; the co-occurrence of luminance and chromatic change is likely due to reflectance change. In our study, depth perception from shading was explored with a simple combination of chromatic and luminance sinusoids. The stimuli consisted of a luminance and a chromatic sinusoid (each 0.75 cpd), either having the same or orthogonal orientation. Observers were asked to match the amplitude of depth perceived in the luminance corrugation with that of a circular depth pedestal (diameter 2 deg), in a random dot stereogram. As the colour-shading effect predicts, we found for the orthogonal condition, that depth did increase as either chromatic or luminance contrast was increased in the stimuli, although there were large variations in the strength of the effect across our pool of 14 na&ıuml;ve participants. When the chromatic and luminance sinusoids were aligned, some of our observers reported the expected decrease in perceived depth as chromatic contrast was increased. However, around one third of our participants reported instead an increase in the perceived depth. Not all of our participants behave as if the co-occurrence of a luminance and chromatic edge is due to a reflectance change (i.e. a change in the material colour). This challenges the basic logic behind the colour-shading effect. Simpler arguments, based on masking between chromatic and luminance channels could provide an alternative explanation for the perceived depth variation.
EPSRC UK, (EP/G038708/1).