In order to observe the major result, just keep fixating the red point. After a while, you should see the two gratings sliding over each other instead of a plaid moving upward. Then, the perception of plaid stimuli switches back and forth between the coherent and the transparent interpretation -- it is bi-stable.
This interactive demo allows you
also to observe that the probability of seeing the transparent or the
percept is mostly affected by changing the directions of the
the more different the directions are, the more easy it is to see
When the angle between the two directions is only 60 degrees (+/- 30
from the global direction), it might seem impossible to see the
percept. In fact, if you keep fixating, you should be able to see the
slidings even in this condition - but it can take more than 1 minute,
then the transparent percept duration is short.
By changing the speed of the stimulus, you should experience that slowly moving plaids tend to slide less easily.
In this demo, the luminance of the intersections is compatible with physical transparency. By clicking on "Not transparent form cues", you can verify that transparent motion is less likely. If the stimulus is presented for only short durations, then transparency can be experienced only when the intersections are transparent, as has been documented by Stoner and Albright (1) (see the "Stoner story" demo). However, if you keep fixating the red point (you have to go back to the interactive demo), you should be able to see the two gratings clearly sliding over each over. The effect of form cues on integration and segmentation is therefore qualitatively similar for plaids and rotating disks.
Albright, T.D., and Ramachandran, V.S. Transparency and coherence in
motion perception. Nature 344: 153-5, 1990.
Stoner, G.R., and Albright, T.D. The interpretation of visual motion: evidence for surface segmentation mechanisms. Vision Res. 36: 1291-310, 1996.