Abstract
Visual discomfort is affected by deviation from the ‘natural’ 1/f relationship between luminance and spatial frequency. However less is known about the contribution of color. One metric that does appear to contribute to discomfort is the difference between neighboring colors (defined in terms of CIE UCS chromaticity separation) when examined in grating patterns and abstract art. For three independent studies (totaling 166 participants), the contribution of chromaticity separation on ratings of discomfort was examined when semantic content was included in the images. We used machine learning algorithms to merge images of objects (animals, chairs, cars etc.) with ‘textures’ that had been used in prior studies of visual discomfort (stripy vs non-stripy objects), trypophobia (visually-induced fear of holes), and natural scenes (125 images in total). The spatial and chromatic properties of the textures were preserved but were unrecognizable in the merged images. Both greater chromaticity separation and greater deviation from 1/f correlated with greater discomfort, and those who were visually sensitive were more likely to report greater discomfort. A remaining question is whether there is an advantage in being sensitive to chromaticity separation? Large chromaticity separations are uncommon in nature, particularly for spatially proximal features, with one exception being animal markings. However, preliminary exploration suggests that the coloration of poisonous/venomous animals does not predict the likelihood of someone avoiding the animal. Altogether, chromaticity separation appears to be a simple metric that can predict visual discomfort under different situations, and has implications for how color is used in the modern environment.