In this thesis three studies are presented. The first study (Chapter 1) investigates the properties of 2 nd -order spatial frequency channels. Here, the stimulus is a 2 nd -order sine wave grating (or a sum of sine waves) formed by a pair of vertically and horizontally oriented texture images. We find modulation contrast sensitivity to be nearly flat over a five-octave range of spatial frequency, but band pass when stated as efficiency (relative to an idealized observer confronted with the same task). Increment threshold shows the improved performance with a sub-threshold pedestal seen in the "dipper effect," but the typical Weber's law behavior at higher pedestal contrasts is not observed at the highest pedestal modulation contrasts achievable with our stimuli. Subthreshold summation experiments indicate that 2 nd -order filters have a moderate bandwidth. The second study (Chapter 2) involves identification of 2 nd -order letters. Using a critical-band masking paradigm we investigate the channels underlying identification of 2 nd -order letters. We find that 2 nd -order letters are identified with a single, approximately 1.5 octave wide channel. This channel is fixed across different noise conditions indicating the inability of observers to look off-frequency. When the experiment is repeated at different viewing distances, we find that the channel frequency used to identify the letters changes proportionally with the letter frequency. In other words, scale invariance holds for 2 nd -order letter identification. The third study is on depth cue combination. Through a slant adjustment task we investigate how two pictorial cues, linear perspective and texture, are combined allowing for the possibility of correlation between the two cues. Eight observers participated in this study. We find all observers' performance to be consistent with a weighted linear combination of the two cues. Further, we find three observers to be optimal, with uncorrelated cues, three other observers to be optimal with positive correlation among the two cues. The performance of the remaining two observers were consistent with suboptimal combination.
