Abstract
In spite of 20 years of research, there is little consensus as to the number or sensitivity profiles of the higher-order channels that mediate chromatic sensitivity in human color vision. We have studied the problem in a chromatic noise masking experiment using a combined field sensitivity and field additivity paradigm. Our goal was to evaluate the suggestion of D'Zmura & Knoblauch that channels seem narrowly tuned because the subject uses different channels to detect a test stimulus under different masking conditions (“off-axis looking”). We used an orange/blue isoluminant test grating, and dynamic isoluminant chromatic noise maskers, with variable azimuth but constant (gray) average chromaticity and 5% luminance jitter within a rescaled DKL color space. We measured the field sensitivity of the mechanism(s) that detected our test grating, and found a profile that was most sensitive to the red/green (“RED”) noise masker and much less sensitive to the orange/blue (“ORANGE”) and yellow/purple (“YELLOW”) maskers. This suggested that our test grating was detected via a narrowly tuned “red” mechanism. We compared detection thresholds for our orange/blue test grating presented with a strong RED masker, a strong YELLOW masker, and both maskers combined. The threshold-elevating power of the YELLOW masker was 15 times greater when combined with the RED masker than when it was presented alone. This suggested that the orange/blue grating is detected via a “red” channel when it is presented alone or with the YELLOW masker, but shifts to a “yellow” or “orange” channel when the RED masker is used. The color appearance of the orange/blue test grating changed in agreement with this interpretation: it looked strikingly red/green in the presence of the YELLOW masker alone, but looked orangeish/bluish in the presence of the RED masker. We conclude that the narrow tuning in our single-masker conditions is probably due to off-axis looking, just as D'Zmura & Knoblauch suggested.