Abstract
When a full range (black to white) Mondrian is covered by a veiling luminance equal to the luminance of a white patch, observers see only light gray patches (consistent with the range compression) and no veil. But an abstract full range 3D still-life covered by a veil is seen veridically and the veil is perceived (Gilchrist & Jacobsen, 1983). A small object-filled room, viewed close-up to fill the whole visual field, in which every surface is painted white is perceived as white while the same room painted black is perceived as middle gray, a perceptual difference due to the high amplitude of luminance variations in the black room, compared with the low amplitude in the white room (Gilchrist & Jacobsen, 1984). So what happens when a black room is viewed through a veiling luminance? Will observers perceive a black room and see the veil, as predicted by the 3D display? Or will they perceive a white room, and no veil, due to the reduction in luminance amplitude? Nine observers viewed a 40 cm cubic room filled with various objects, all painted matte black and made matches using a Munsell chart. A rear-illuminated sheet of translucent acrylic was reflected in a sheet of clear glass through which the observer viewed the room. The observers perceived a white room, with an average Munsell match of 9.3, and in general did not perceive a veil. In a no-veil control condition (n = 7) the room was seen as gray (6.9). Because painting a black room white does not transform the image in the same way as covering a black room with a veil, except in the gross sense of luminance amplitude, this implies that information on the exact shape of the luminance gradients is not used by the visual system.
NSF (BCS - 1027093) NIH (BM 60826-02).