 |  |
In visual perception studies, experiments show that people, given a shaded image, have the tendency to interpret the 3-D configuration based on the assumption that light is coming from above. This is illustrated in the following two images which are exact 180-degree rotations of each other.
| | |
The image on the left is typically
perceived as concave depressions surrounded by convex bubbles, while
the image on the right is perceived as convex bubbles surrounded by
concave depressions. In both cases, the percept is driven by the
assumption that the image is being lit from above, regardless of the
rotation of the image.
This preference for top lighting is
frequently exploited by photographers and cinematographers. A sense
of unnaturalness and evil can be conveyed when a scene is illuminated
by bottom lighting.
 |  |
While plenty of evidence points to a preference for top lighting, does this mean that we prefer equally all lighting directions that are above the horizon? Is there a more specific preference, and where
might that be?
To answer this question, we used
stimuli like the ones shown below in which subjects were asked to spot
the odd-man-out concave depression for a variety of different lighting
directions. 0 degrees denote lighting from directly overhead, 30
degrees denote lighting from 30 degrees left of the vertical, and -60
degrees denote lighting from 60 degrees right of the vertical.
Our data gives us the surprising
result that people most often prefer a lighting direction that is top
and LEFT! We have found a correlation between this
top-left lighting preference and handedness; right handers have a
stronger preference for left lighting, while left handers have a
weaker preference for left lighting, sometimes even a preference for
right lighting.
Indeed, if you take a close look at your web browser, you will
immediately discover that, more often than not, 3-D icons such as the
buttons, slider bars, etc. are rendered with top-left lighting! For a more detailed description of these experimental results,
download the postscript-format manuscript, "Where is the Sun?". References: Ramachandran, V. S. (1988). Perception of shape from shading. Nature, 331(6152), 163-166. Braun, J. (1993). Shape-from-shading is independent of visual attention and may be a 'texton'. Spatial Vision, 7(4), 311-322. Sun, J. Y. \&\ Perona, P. (1996b). Where is the Sun? Investigative Ophthalmology \&\ Visual Science, 37(4), 4283.
This preference for top-left lighting in interpreting 3-D shape,
while not having been previously documented, is apparently something
that has been exploited by artists for centuries. In a survey of 225
paintings taken at random from the Louvre, the Prado, and the Norton
Simon, we found that more than 75% of the paintings were lit from the
left. Roman mosaic protraits from close to two thousand years ago
typically used top-left lighting as well.
