Abstract
We examine the question whether physically different transparent media can be meaningfully mapped onto one perceptual dimension of perceived transmittance (Robilotto and Zaidi, 2004). We rendered images of variegated checkerboards that are composed of checks of varying reflectances. Part of the checkerboard was covered by a transparent medium that varied in reflectance and transmittance. We used Maximum Likelihood Conjoint Measurement (MLCM) to derive perceptual scales of perceived transmittance for different transparent media. In a single trial we presented two checkerboards next to each other that were identical except for the region covered by the transparent medium. Observers’ task was to judge which of the two media was more transparent. We estimated four scales of perceived transmittance for four different transmittance values. The scale values of each scale decreased monotonically with increasing reflectance of the transparent medium. All observers show a perceptual trade-off between the two physical parameters indicating perceptual equality of transparent media with different transmittance and reflectance. Scales in different conditions were captured surprisingly well by the logarithm of the root-mean-square contrast of luminances within the area of transparency. In a second part of the experiment we presented observers with the same stimuli but reduced to the area of transparency. These stimuli varied in mean luminance and in contrast, but all cues to transparent overlay were eliminated. Observers judged which of the two stimuli had higher perceived contrast. The scales derived for the reduced stimuli were identical to those observed with the full stimuli, indicating that perceptual judgments of perceived transparency and perceived contrast rely on the same mechanism also in more complex images such as the ones used here.
Acknowledgement: MA 5127/3-1 and MA 5127/4-1