Colour constancy describes the ability of our visual system to keep colour percepts stable through illumination changes. This is an outstanding feat given that in the retinal image surface and illuminant properties are conflated. Still, in our everyday lives we are able attribute stable colour-labels to objects to make communication economic and efficient. Past research shows colour constancy to be imperfect, compensating for 40% and 80% of the illumination change. While different constancy determinants are suggested, no carefully controlled study shows perfect constancy. The first study presented here addresses the issue of imperfect constancy by investigating colour constancy in a cue rich environment, using a task that resembles our everyday experience with colours. Participants were asked to recall the colour of unique personal objects in natural environment under four chromatic illuminations. This approach yielded perfect colour constancy. The second study investigated the relation between illumination discrimination and chromatic detection. Recent studies using an illumination discrimination paradigm suggest that colour constancy is optimized for bluish daylight illuminations. Because it is not clear if illumination discrimination is directly related to colour constancy or is instead explained by sensitivity to changes in chromaticity of different hues, thresholds for illumination discrimination and chromatic detection for the same 12 illumination hues were compared. While the reported blue bias could be replicated, thresholds for illumination discrimination and chromatic detection were highly related, indicating that lower sensibility towards bluish hues is not exclusive to illumination discrimination.Accompanying the second study, the third study investigated the distribution of colour constancy for 40 chromatic illuminations of different hue using achromatic adjustments and colour naming. These measurements were compared to several determinants of colour constancy, including the daylight locus, colour categories, illumination discrimination, chromatic detection, relational colour constancy and metameric mismatching. In accordance with the observations in study 2, achromatic adjustments revealed a bias towards bluish daylight illumination. This blue bias and naming consistency explained most of the variance in achromatic adjustments, while illumination discrimination was not directly related to colour constancy. The fourth study examined colour memory biases. Past research shows that colours of objects are remembered as being more saturated than they are perceived. These works often used natural objects that exist in a variety of colour and hue, such as grass or bananas. The approach presented here directly compared perceived and memorized colours for unique objects, used also in the first study, and confirmed the previous findings that on average, objects were remembered more saturated than they were perceived.
Verknüpfung zu Publikationen oder weiteren Datensätzen
