Abstract
If a visual property serves as a “preattentive” basic feature, then that property will support efficient visual search. When the target in a search task has that feature while the distractors do not, reaction times (RTs) will be largely independent of the number of items present. Several surface properties (shading, shininess) have been identified as basic features in this sense. Do the surface properties of transparency and opacity support efficient visual search? In Exp. 1, we found that search for transparent targets among opaque distractors (and vice versa) in a static display was very inefficient (> 30 ms/item). In Exp. 2, stimuli were horizontal green-tinted bars that moved over a random, achromatic texture of disks of different luminance levels. Opaque surfaces were generated by having the tinted random texture move with the bar. Transparent surfaces were generated by having the texture of the bar change over time such that the background texture appeared to remain stationary while a green filter moved over it. Static examples of transparent and opaque bars would be indistinguishable. Search for the opaque target among transparent distractors was quite efficient (5.3 ms/item); search for a transparent target among opaque distractors less so (16 ms/item). One confounding factor is the different pattern of motion of the texture within opaque and transparent bars. To determine if Os could use this motion cue, we presented the same horizontal bar search stimuli on a blank background. Mean RTs were much slower in these conditions. Now the moving (formerly “transparent”) texture was easier to find amongst the unmoving (“opaque”) textures (17 ms/item) than vice versa (71 ms/item). It is possible that cues other than opacity/transparency contribute to detection of these items on a textured background. Nevertheless, these data suggest that opacity may serve as a basic visual feature.