Spatial integration and effective spectral density of one-dimensional noise masks.

Abstract

When masking one-dimensional gratings, the strongest masking effect is achieved by using one-dimensional spatial noise, which can be regarded as a special case of two-dimensional noise where the noise check height is equal to the grating height. The extent of spatial integration in the human visual system is limited, however. Hence, our aim was to investigate whether the effective height of noise checks of one-dimensional noise is similarly limited. We measured detection thresholds for vertical sinusoidal gratings with added spatial noise. The width of the noise checks remained constant, but their height increased until equal to the height of the image window which made noise one-dimensional. The contrast energy thresholds increased in direct proportion to increasing noise check height and the spectral density of noise, calculated by taking into account both the height and the width of the noise checks. The increase levelled off, however, after the critical noise check height (nyc). The critical noise check height in grating cycles changed as a function of spatial frequency (f) as nyc = 4.7 [1 + (1.4/f)2]-0.5. According to our results the effective height of noise checks was thus limited in accordance with studies on spatial integration, showing scale invariance above 1.4 c/deg.