Abstract
We investigated anisotropic suppression of contrast sensitivity by broadband masks (strongest at horizontal: the “horizontal effect”) as a function of spatial frequency to assess whether there was a differential contribution to this anisotropy of content at different spatial frequencies. Using isotropic, 1/f broadband noise as an contrast pedestal, we tested increment thresholds for six 1.0 octave (non-overlapping) bands centered at 0.28, 0.65, 1.5, 3.4, 7.9, and 18.2 cpd, and also for the full band from .2 to 25 cpd. We found that at a number of pedestal contrasts, contrast sensitivity for the first four (lowest) bands was a linear function of frequency (cf. Schofield and Georgeson, 2002 VR), with sensitivity to the 7.9 cpd band nearly enough to account for broadband increment sensitivity the frequencies at which the sensitivity functions intersected full band sensitivity tended to fall around 8 cpd, indicating that a single channel in this vicinity of the frequency range is responsible for detection in the broadband condition. For oriented test bands we found that the horizontal effect typically seen in detection of broadband stimuli (Essock et al 2003 VR) would be anticipated given the sensitivity pattern to the peak frequency bands. Orientation effects were also apparent at other frequencies, perhaps contributing to the horizontal effect of perceived contrast of broadband oriented noise reported elsewhere (Hansen and Essock 2006 VR). Based on sensitivity across a number of background RMS contrasts, we develop a model of perceived broadband contrast structure in isotropic noise to account for earlier findings.