Abstract
Six linear chromatic mechanisms are sufficient to account for the pattern of threshold elevations produced by chromatic noise masking in the LM plane of cone space (Shepard et al, 2015). We assume the mechanisms are "labeled lines" (Graham 1989), implying here that the colors of physically-different stimuli that are detected by a single mechanism should all be the same, but that the colors of two stimuli detected by different mechanisms may differ. In the present study, three observers' detection thresholds, from multiple chromatic noise conditions, were fit with the six mechanism model. Observers were then presented the threshold-level tests (Gaussian blobs, σ=1°), in the presence of the same chromatic noises that were used in the detection experiment. Observers used a second display, positioned off to one side, to select a matching chromaticity on a HSV color wheel. According to the model, tests that lie along one mechanism threshold line should produce very similar color matches, while tests that lie on two different mechanism lines should be matched with appreciably different colors. In other words, the observers' color matches apply a label on each line, providing an additional test of the detection model and insight into the subjective experience resulting from these mechanisms. Cluster analysis shows that measured color matches fall into no more than 6 clusters in (u', v') space (across all the noise conditions), and these clusters correspond to the six mechanisms in the model. Furthermore, observed matches corroborate the qualitative differences between each observers' detection model fits. Most importantly, where the detection model determines that a given test angle is detected by different mechanisms in different noise conditions (due to differential masking), the hue of that test angle changes in a consistent way. These color matches verify the six-mechanism model, and quantify the hue signaled by each mechanism.
Meeting abstract presented at VSS 2016