What to observe
Observe the movement of blue and yellow “feet”. The feet seem to step alternately, like tiny feet going tip-tap-tip-tap… This is more pronounced if you do not look directly on the feet, but between them. In reality their movement is always simultaneous. I find this phenomenon particularly cute.
What to do
When deselecting the ‘Hi contrast’ checkbox, the grating contrast is reduced, demonstrating that the feet are not stepping out of phase.
The stepper “bars/foot” sets the number of bars per foot. This is preset to 4, the effect occurs also at all other even values. For odd values, e.g. 5, the motion also becomes odd: instead of stepping alternately, the feet seem to move like worms, extending and shortening. [This option added 2012.]
The sliders at top and bottom control the luminance of their respective foot. When you make the yellow foot brighter than the light bars of the grating, it does not lag behind any more. When you make it very dark, it moves together with the blue foot. The sliders (added 2010) allow you to experiment with the explanations below, I still prefer the “level 2” explanantion.
The pop-up menu below “Hi contrast” allows you to select other colour combinations, namely red-green and gray shades. For red-green, the contrast of the grating is reduced a little, making the dark lighter, but it still needs a very dark red (=brown) to equate luminance.
There are more buttons with (hopefully) obvious functions.
Stuart Anstis first demonstrated this illusion in 2003.
Deselect the “Color” checkbox. Now it becomes obvious that the edges of the light ‘foot’ merge with the light bars, and are only visible when they traverse the dark bars. So half of the time there really is no motion cue, and perception goes into default, i.e. no motion. For the dark foot the same holds, only at alternate times.
With reduced contrast of the grating, equiluminance of edges and grating is no longer present, so the effect disappears.
The “worm-like” movement for an odd number of bars per foot is automatically explained as well.
Level-1-explanation plus: The edge information is only contained in colour, and the (magnocellular) motion system cannot see it.
The effect still persists when there is a slight contrast between foot and grating (can be achieved with the sliders). Anstis (2004) attributes this to slowing down of motion under conditions of reduced contrast (Thompson 1982), and goes much deeper into the subject.
Just when I thought this illusion can be easily understood, there appears this complicated paper Howe et al. (2007), seemingly disproving my above thoughts and suggesting a very intricate explanation. Recently, I included this paper in a seminar and we tried to follow its evidence and reasoning, and couldn't. In consequence, my favoured explanation regresses to levels 1+2 above.
The demonstration above was inspired by Stuart’s version. Thanks to Wolfgang Wesemann for first drawing my attention to this phenomenon, and Wolfgang Beyer for bright ideas.
Anstis SM (2003) Moving objects appear to slow down at low contrasts. Neural Netw 16:933–938
Anstis SM (2004) Factors affecting footsteps: contrast can change the apparent speed, amplitude and direction of motion. Vision Res 44:2171–2178
Thompson P (1982) Perceived rate of movement depends on contrast. Vision Res 22:377–380
Howe PDL, Thompson PG, Anstis SM, Sagreiya H, Livingstone MJ (2006) Explaining the footsteps, belly dancer, Wenceslas, and kickback illusions. Journal of Vision 6, 1396–1405