Based on day to day experiences, computer graphics researchers have concluded that abrupt changes in velocity of a visual stimulus are easily detected. Meanwhile, vision scientists have shown that the human visual system does not have acceleration or "change of velocity" detectors. The second view implies that attention to the movement of an object is required to detect changes in its velocity. We report the results of a set of experiments to test the sensitivity of the human visual system to smooth and abrupt changes in the base velocity of objects. These experiments are the temporal equivalent of spatial contrast sensitivity experiments performed using sine wave gratings many years ago.

Using 2AFC and PEST we determined the minimal amplitude of periodic velocity modulation (ie, the threshold of detection) of the base velocity of an object for three periodic waveforms (sine, square, and triangle), frequencies of modulation from 1 to 9 Hz and base velocities between 3 and 9 degrees/second.

Summary: the human visual system is able to detect each of the waveforms equally well with slightly better sensitivity to the square waveform -- implying that velocity discontinuities are largely irrelevant and it is the distance traveled that leads to higher sensitivity; the highest sensitivity is near 3 Hz with decreasing sensitivity and increasing variance away from 3 Hz; sensitivity is highest for lower base velocities and decreases as variance increases with base velocity.

Applications: Quantified limits on the ability of the human visual system to detect changes in velocity can help inform the design of algorithms for compressing motion data (important for computer games) as well as designing the interpolation methods commonly used in computer animation and simulation.

Extensions: experiments are planned to extend these experiments to line segment rotations and rotations of the joints of articulated figures