Abstract
Contour interaction for visual acuity is greater for contrast-modulated letters than for luminance-modulated letters and appears to be more robust to the effects of dioptric blur (Hairol et al; Waugh et al, VSS2010). As dioptric blur can be complicated, we investigated the effects of controlled imposed image blur on contour interaction for contrast-modulated (CM) and luminance-modulated (LM) C stimuli. Stimuli were constructed from dynamic binary noise, added or multiplied to a square-wave profile. Noise was either unscaled, subtending a constant angular size for all Cs, or scaled to have ∼15 noise checks per C. We measured acuity for square Cs, with and without four flanking bars separated by 0–2 letter widths under different levels of image blur. Blur was created by filtering the image with low-pass Gaussians (sds of 32–0.5 c/deg). Acuity thresholds (logMAR) were obtained using a method of constant stimuli and 4AFC paradigm. Visual acuity thresholds for CM Cs were always higher than for LM Cs by 0.3–0.4 logMAR. This difference was magnified when bars were placed next to (0.4–0.6 logMAR) or close to (0.4–0.5 logMAR) the target C. Across blur levels the amplitudes and the extents (in barwidths) of contour interaction were greater for CM than for LM Cs, particularly for unscaled noise. The initial rate of improvement in visual acuity (MAR) with decreasing blur (increasing Gaussian sd) for crowded and uncrowded Cs, was similar for LM and CM Cs (power function exponent of ∼−0.9), although slightly flatter for CM Cs constructed with scaled noise. As with dioptric blur, CM stimuli are more susceptible to the effects of contour interaction than LM stimuli with imposed image blur, especially for unscaled background noise. Differences reduce for scaled noise, suggesting that the effects of blur on carrier noise contribute in part, to the overall visual acuity response.
JPP was funded by a VHS Summer Research Scholarship.