Abstract
A 3-D spiraling cylinder can be perceived when viewing a random-dot cinematogram (RDC) in which dots are assigned new directions of motion from a distribution spanning 90 deg because local motions consistent with an actual transparent 3-D spiraling cylinder are present in each frame of the display (Williams & Phillips, unpublished). We examined the effect of local motion duration on the strength of this 3-D percept.
Observers viewed RDCs comprised of 150 dots, each assigned a direction from a uniform distribution of directions spanning 90 deg (frame rate=100 Hz). Each trial, observers reported the perceived depth using a scale ranging from 0 (no depth) to 9. Individual dots changed their direction of motion after the number of frames specified by the duty cycle had elapsed. Exp 1 varied stimulus duration for duty cycles of 1 (direction change each frame) and another equal to the stimulus duration (no direction change). Exp 2 tested many duty cycles with duration fixed at 100 frames. Exp 3 varied direction distribution sampling resolution for duty cycles of 1 and 100 (duration = 100 frames). For each condition in a trial block (10 trials/condition), we computed a mean depth rating. Ten mean depth ratings per condition were collected for each subject in Exp 1 and 2, and 5 mean depth ratings per condition in Exp 3.
Overall, depth ratings increased with duration but perceived depth was lower for a duty cycle of 1. Perceived depth systematically increased with duty cycle, reaching asymptote at 10–13 frames. Sampling resolution of the direction distribution did not affect perceived depth. Our results suggest that local motion signal strength is critical to perceiving depth in these displays. The longer a dot travels in a consistent direction, the greater the strength of the local motion signal. The perceived-depth asymptote at duty cycles of 10–13 frames (100–130 msec) reflects the ‘strongest’ local signal, consistent with motion detector cell integration times.
Supported by NSF Grant IBN-9983563 and the Ohio Board of Regents.