Citation

Gerbino W, Stultiens CI, Troost JM, de Weert CM. J. Exp. Psychol. Hum. Percept. Perform. 1990; 16(1): 3-20.

Copyright

(Copyright © 1990, American Psychological Association)

DOI

unavailable

PMID

2137521

Abstract

Perceived transparency was studied as a constancy problem. In the episcotister (E-) model of scission, luminances are partitioned into layer and background components; four luminances determine values of two layer parameters that specify constancy of a transparent layer on different backgrounds. The E-model was tested in an experiment in which 12 Ss matched 24 pairs of four-luminance patterns by adjusting two luminances of the comparison pattern. Both the standard and the comparison were perceived as a transparent layer on a checkerboard. The E-model predicts matches when layer values are identical in the two patterns. One parameter was constant, constraining the adjustment along the second dimension. Obtained values corresponded well with E-predictions. Alternative models based on local luminance or average contrast ratios accounted for less variability. Results indicate that transparency models should utilize luminance, not reflectance, as the independent variable.

Language: en