Discriminating smooth from sampled motion: chromatic and luminance stimuli

Abstract

It was reported by Mullen and Boulton [ Vision Res.32, 483 ( 1992)] that there is a lack of smooth-motion perception for chromatic stimuli, and we were interested in quantifying this observation to try to see how it relates to the underlying processing of moving chromatic stimuli. We suggest that, if a stimulus looks as though it is not moving smoothly, a greater degree of quantization will be necessary before observers are able to discriminate between fine and coarse temporal sampling of the moving pattern. We used a two-alternative forced-choice discrimination paradigm to measure observers’ ability to discriminate finely sampled (smooth) from coarsely sampled (jerky) motion in luminance (achromatic) and chromatic (red–green) gratings. Stimuli were presented for 0.5 s, moved at 2 or 4 Hz, and had a spatial frequency of 1, 5, or 10 cycles/deg. At low contrasts (0.5 log unit above detection threshold), sampled motion was more discriminable in luminance than in chromatic gratings. As contrast increased, performance improved in both chromatic and luminance gratings. As spatial frequency increased, performance improved for chromatic but fell for luminance gratings. Analysis of the data suggests that the ability to discriminate fine from coarsely sampled motion in both luminance and chromatic gratings can be predicted from the temporal Fourier spectra in terms of the “window of visibility” [ J. Opt. Soc. Am. A3, 300 ( 1986)]. We conclude that the chromatic window of visibility is smaller than the luminance window at given spatial and temporal frequencies and gives a measure of the temporal resolution of the chromatic motion system. However, our results imply that the window of visibility cannot explain the percept of smooth motion. Under conditions in which there are no visible artifacts in the chromatic stimulus, observers seem still to perceive the motion as jerky, and we consider other possible explanations for this phenomenon.