Abstract
Knau and Werner (2002, JOSA A, 19, 208-214) demonstrated that saturation of parafoveally presented colours changes as a function of stimulus size - smaller single wavelength stimuli were found to appear increasingly desaturated, with a stronger effect for older compared to younger observers. Our study successfully replicated their findings using isoluminant colours that isolated the two cone-opponent mechanisms: reddish, greenish, bluish and yellowish. Perceived contrast of a circular patch, varying in size (2˚, 1˚, 0.5˚, 0.33˚, 0.20˚, 0.15˚), presented at 4˚ or 5˚ eccentricity was matched to a 2˚ standard patch at 5.5˚ eccentricity by a group of younger (22-39 years) and older (60-82) observers. Desaturation occurred from 0.33˚ onwards in the younger group and 0.5˚ onwards for the older observers, who showed a more pronounced effect overall. Desaturation was present for all colours and was strongest for bluish stimuli. To better understand these findings, contrast detection thresholds were measured for a group of younger observers using a subset of stimulus sizes. Whilst a clear increase in detection thresholds with a decrease in patch size was observed, there were no asymmetries between mechanisms, contrary to the matching experiment. The desaturation thus cannot be fully explained as a linear consequence of changes to detection thresholds. Finally, the same matching paradigm was used with a group of younger observers on a subset of stimulus sizes, examining if desaturation also occurs if colours contain a suprathreshold luminance signal. This time, desaturation with a reduction in size only occurred for blue. While desaturation of small isoluminant stimuli may be the product of optical factors, desaturation for blue even in the presence of luminance contrast points towards a cortical contribution, in line with exiting literature. Our results imply that colour appearance models based on 2 stimuli cannot fully capture saturation if stimuli are small (≤0.5).