Abstract
Ensemble color percepts are biased towards higher chroma (e.g., Kimura, 2018), but it is unknown whether this bias also affects perceived hue when chroma and hue are correlated. We studied whether hue is processed independently of chroma and lightness by varying their respective correlations in color mosaic stimuli and measuring perceptual biases. Stimulus colors were defined in LCh-coordinates of the CIELUV color space. Preliminary experiments determined observers’ just noticeable differences (JND) for lightness, chroma and hue. The test stimulus was a 6-by-6 grid of 0.5-degree square elements of uniform color, displayed against a gray background with chromatic noise. In different experimental conditions, the hue, lightness and chroma values of the stimulus elements were either fixed or variable in proportion to the observer’s JNDs. Variation in one dimension could be independent of, or perfectly correlated with, one of the other dimensions. The comparison stimulus was the same size as the test, but without color variation. In Experiment 1, the test and comparison stimuli were presented sequentially for 500ms. Depending on the condition, the observer compared them in terms of hue, chroma, or lightness. The comparison stimulus value was varied over trials, and the perceived color of the test was estimated by fitting psychometric functions to the proportion of bluer/more chromatic/lighter responses. In Experiment 2, observers adjusted the comparison stimulus to match the color of the test in all three dimensions simultaneously. Observers showed a bias towards higher chroma, but not towards the correlated hue. In contrast, there was no bias in lightness estimates, but a bias towards the hue correlated with higher lightness. The different dimensions of color do not seem to be averaged jointly, and chroma and lightness variations have distinct effects on the color ensemble percept. References: Kimura, E. (2018). Averaging colors of multicolor mosaics. JOSA A, 35(4), B43-B54.