Two Types of Signals Mediating Surface–Color Judgments

Abstract

Abstract: : Purpose: Observers can readily detect changes in surface color under different illuminants (Foster et al., 2001, Proc. Nat. Acad. Sci. USA, 98, 8151–8156). This ability may be based on changes in spatial cone–excitation ratios, which provide information about the relationships between surface colors rather than surface color itself (Foster, 2003, Trends Cogn. Sci., 7, 439–443). The aim here was to elucidate the possible different signals underlying surface–color judgments. Methods: Simulations of Mondrian–like colored patterns were presented on a color monitor with 15–bit resolution per gun. The patterns consisted of 49 (7 × 7) abutting 0.8–deg–square uniform surfaces with spectral reflectances drawn from the Munsell set. In each trial, two images of a pattern were presented in sequence, each for 1 s with no interval, under two different daylight illuminants: one with correlated color temperature 16000 K, the other 4000 K. The spectral reflectance of the center surface in the second image was, however, changed randomly from trial to trial. Eight observers with normal color–vision performed two tasks in separate sessions: in one, they rated the subjective appearance of the central stimulus, i.e. the quality of the "material" simulation or the sameness of the hue and saturation (Arend et al., 2004, JOV, 4, 321a); in the other task, they decided whether there was an objective illuminant or surface–reflectance change. Results: Although results pooled over observers were similar for the two kinds of surface–judgment tasks (e.g. 0.6–0.8 on a scale with 1.0 representing ideal performance), there were noticeable differences between individuals. With the rating task, some observers assigned closely similar "material" and hue–saturation scores, but with the discrimination task, the same observers made more accurate judgments of surface–reflectance changes. Conclusions: The reliability and consistency of observers’ surface–color judgments depend on task, suggesting two types of signals mediating their decisions: one more subjectively oriented and to do with estimating individual stimulus properties, the other more objectively oriented and to do with the physical invariance of surface–color relationships.

Keywords: color appearance/constancy • color vision • photoreceptors: visual performance