Fig. 1. Schematic of stimulus. Stimuli always contained a center standard pattern surrounded by eight other patterns of asterisks, some of which matched the standard (called targets) and some of which differed from the standard (called nontargets).Second, the effects on eye-movement parameters of the size of the matrix from which patterns were generated was assessed. Patterns were generated from 4 by 4, 6 by 6, 8 by 8, and 10 by 10 matrices. The larger the matrix size, the larger the average number of elements in a pattern. The results of this study showed clearly-perhaps surprisingly-that pattern discrimination was not affected by these variations in matrix size.Third, the effect on eye-movement parameters of the variations in the number of elements in a pattern of a particular matrix size was evaluated. The results of the study showed that variations in the number of elements in target patterns of a particular matrix size affected eye movement parameters differently than did variations in the number of elements in nontarget patterns of the same matrix size. A target pattern was nearly always fixated foveally, whereas the probability of looking directly at a nontarget pattern depended upon its "similarity" to the standard pattern. Duration of fixation on a pattern, however, was determined both by variations in target patterns and by variations in nontarget patterns.Fourth, an attempt was made to identify the features or variations in patterns, in addition to those specifically manipulated in this experiment, that most affect human pattern discrimination. There presently exists no universal psychophysics of pattern perception. That is, although the three physical variables of intensity, wavelength, and purity account well for the perception of homogeneous visual fields, the physical variables that account for the perception of patterned visual fields have not been satisfactorily identified. This is true even though 1~200

The effects of two variables on the regeneration rates required to prevent flicker on CRT displays were investigated in this computer-automuted experiment. The two variables were the sequential order in which the display was scanned ( e g , horizontal, vertical, or pseudo-random scanning) and the persistence of the phosphor on the display. Twenty-one different scan orders and two phosphors (P-12 and P-38) were used. Results showed that flickerless regeneration rates depended principally upon phosphor persistence; scan order had only a minor influence. The main effect of random and pseudorandom scanning was to reduce the disturbing effects of display flicker when it did occur, rather than to reduce significantly the regeneration rate at which flicker did not occur.The cathode ray tube (CRT) is used as a display device in several systems that receive information conveyed over lines rather than information that is broadcasted, e.g., picture telephone and computer terminals. The required bandwidths of such systems (or rates at which information is conveyed from its source to the display) are closely related to system costs. A reduction in bandwidth can lead to a reduction in display system costs. Accordingly, bandwidth reduction is often advantageous, especially if it does not lead to deterioration in system performance. Display bandwidth is proportional to the product of two variables: the resolution of the display multiplied by the regeneration rate of the display. Consequently, a reduction in either the regeneration rate, which is presently 40-60 frames per second (fps) (Adams Associates, 1967), or resolution of a display are two ways to accomplish a reduction in display bandwidth. A reduction in the flicker-free regeneration rate of a display is especially attractive because unlike a reduction in display resolution, it does not detract from the usefulness of the display. "Flicker-free'' means that the viewer does not observe the temporal luminance changes on an otherwise homogeneous luminance field caused by the fading and subsequent regeneration of that field on the display.This study evaluated the effect of two variables on the flicker-free regeneration rate of a CRT display. One variable was the order in which the display was scanned. "Scan order" refers to the temporal sequence in which spots on the face of the screen are illuminated, Scan order was originally considered in the design of broadcast television (Engstrom, 1935). At that time two-line horizontal interlace (odd raster lines are scanned, then even raster lines scanned) was chosen over progressive horizontal scan (consecutive raster lines scanned sequentially) to provide a flicker-free picture. Since then, other types of scan orders in addition to line scans have been discussed, e.g., pseudorandom or dot-line interlace (defined below). A general characteristic of these latter scan orders is that consecutively scanned points, instead of being very close to each other, are relatively widely separated across the display screen. The superiority of these...