Abstract

The effects of edge blur on minimally distinct border (MDB) judgments were investigated using a computerized color video system. A single observer made MDB matches of the red, green, and blue primaries and fixed-ratio binary mixtures of the primaries to a 30-townsend (TD) standard white. The method of adjustment was used. The monitor was 260 cm from the observer and viewed through a Powell achromatizing lens. The stimuli were 1° 30′h × 40′V and were presented with a 4-s on duty/0.25-s off duty cycle. The edge separating test and standard stimuli were formed either by appropriate step functions for each of the three monitor guns or by functions computed by spatial convolution of the appropriate sharp edges with Gaussian blur functions having standard deviations of 1, 2, 4, or 8 min of arc. Within the range of spatial and chromatic parameters explored, the spectral sensitivity, precision, and additivity of the MDB matches were all essentially unaffected by the introduction of edge blur. The results suggest that the sharpness of a border is less important than has been assumed in determining the characteristics of MDB matches.

© 1987 Optical Society of America