Luminance profiles that produce interesting "border effects" (Kingdom & Moulden, 1988) have frequently been used to investigate the perception of surface lightness (perceived reflectance). There are some particularly great investigations of how our perceptual interpretation of these profiles relates to the geometry and illumination of the scene, such as Knill & Kersten (1991), Bergström (1977), Purves, Shimpi, & Lotto (1999), and van den Brink & Keemink (1976). Here, I would like to explore the less-considered topic of how the surfaces on which the luminance profile resides are defined.
A common version of such a luminance profile is shown in the demonstration below. Going from top to bottom, the luminance profile of the central surface is constant up until a short decline just prior to the abrupt increase at the central border, which is followed by a short decline before reaching a constant state again. Perceptually, however, we tend to see a single darker surface above a single lighter surface when the amplitude of the luminance deflections are low. Try increasing the "Gradient amplitude" slider to be able to see the complete luminance profile.
The demonstration does not work on Internet Explorer and also not seem to work on many browsers on mobile devices. It also seems to have some substantial glitches on specific combinations of Chrome, Windows, and video card—disabling hardware acceleration in the browser settings seems to fix it. Apologies if you don't see anything below, or if it has weird artefacts or flicker.
I'm interested in exploring the role of the surface boundary indicator in this perceptual effect. In the above "Traditional gradient" demonstration, the surfaces are delineated from the background by a change in luminance relative to the black surroundings. But does there have to be a luminance boundary in order for the surfaces to have apparent lightness differences?
We will look at a form of surface delineation that is based on relative motion. First, set the "Preset" in the demonstration to "Static noise". You will see that the image has been replaced with a field of noise. However, the central surfaces are still present—there is just no visual signal available to segment them from the background. The surfaces becomes visible when the background begins moving; try setting the background in motion by clicking on the demonstration or selecting the "Motion-defined surfaces" option in the "Preset" field. You should then see the central surface pop out from the background. If you turn the motion off again, you are likely to see the central surfaces quickly disappear and merge with the background.
Now we can probe the question of interest—can such motion-defined surfaces be the target of whatever process causes the apparent lightness differences with the traditional luminance-defined surfaces? If you select the "MDS with gradient" option in the "Preset" field, the luminance profile will be added to the noise pattern.
Do you experience the perceptual effect of a darker surface above a lighter surface when the outer surface boundaries are defined by motion? I am pretty sure that I do, but I have looked at these demonstrations a lot and it is difficult to trust my judgement. We have tried a few times to measure the appearance of demonstrations like the above in naïve observers, but we have actually found it difficult to get a reliable quantification of the traditional effect. I think it is interesting either way—that is, it ought to be instructive of the underling mechanisms whether motion-defined surfaces can or cannot support lightness effects.
Thanks to Sophia Garlick Bock and Lindsay Peterson, who have been involved in developing and testing various incarnations of this demonstration.
- Bergström, S. S. (1977) Common and relative components of reflected light as information about the illumination, colour, and three-dimensional form of objects. Scandinavian Journal of Psychology, 18, 180–186.
- Kingdom, F. & Moulden, B. (1988) Border effects on brightness: A review of findings, models and issues. Spatial Vision, 3(4), 225–262.
- Knill, D.C. & Kersten, D.J. (1991) Apparent surface curvature affects lightness perception. Nature, 351(6323), 228–230.
- Purves, D., Shimpi, A., & Lotto, R. B. (1999) An empirical explanation of the Cornsweet effect. Journal of Neuroscience, 19(19), 8542–8551.
- van den Brink, G. & Keemink, C. (1976) Luminance gradients and edge effects. Vision Research, 16(2), 155–159.