Effect of Ambient Illuminant on Surface Colour Constancy

Abstract

Generally, we are completely enclosed in a real environment, which may act as an entire view-field or adapting field. But in most studies on colour constancy experiments have been made with spatially restricted stimuli. We built a room with gray (N5) walls inside to measure the effect of ambient illuminant on colour constancy. The room illuminant could change its colour from white (D65) to either blue, orange, green, or purple. The observer sat in this main room and adapted to the illuminant for 5 min before the start of the experiment. The observer was shown a smaller room, which had the same variable-colour illuminant, through an aperture (11 deg × 8 deg) in the left side wall. We set the illuminant for each room independently, but in asymmetric illuminant-colour conditions either was set to white. The observer viewed the central part of a colour CRT monitor, placed behind a small aperture (5 deg × 5 deg) in the front wall, and matched its colour appearance to four OSA colour chips in the smaller room. Under symmetric illuminant conditions, eg blue vs blue, the observers' settings showed a complete match with the physical chromaticities of the colour chips. In asymmetric illuminant conditions, eg white vs blue, matched colours showed systematic deviations from both physical chromaticities and colour constancy. This implies that taking the ambient illuminant as adapting field did not yield perfect colour constancy. We introduce a simple model based on incomplete adaptation to the ambient illuminant and a spatial-interaction mechanism, which accounts for our results.