Visual sensitivity to spatially sampled modulation in human observers

Thresholds were measured for detecting spatial luminance modulation in regular lattices of visually discrete dots. Thresholds for modulation of a lattice are generally higher than the corresponding threshold for modulation of a continuous field, and the size of the threshold elevation, which depends on the spacing of the lattice elements, can be as large as a one log unit. The largest threshold elevations are seen when the sample spacing is 12 min arc or greater. These results are similar to those observed by Burr, Ross and Morrone [Vision Research, 25, 717-727 (1985)], who proposed an explanation based on a compressive point nonlinearity. Although their explanation is not consistent with the present data, the results may be explained in terms of nonlinear saturation of a spatially opponent stage early in the visual pathway. Theories based on response compression cannot explain the further observation that the threshold elevations due to spatial sampling are also dependent on modulation frequency: the greatest elevations occur with higher modulation frequencies. The idea that this is due to masking of the modulation frequency by the spatial frequencies in the sampling lattice is considered.