Strong two-dimensional visual representations of nature do not necessarily mimic the physical world, but rather the retinal images that conflate the contributions of illumination, object reflectance/absorption, and atmospheric transmittance. While it is true that aspects of the physical world can be accurately measured–observers may be quick to notice that the measurements are often at odds with the perceptions they elicit. This is not to say that the study of the natural world is inconsequential to the aspiring realist–on the contrary–it is essential. Real world measurements allow for an effective exploration of perception and thus allow visual artists to communicate such phenomena successfully on the canvas.

Let’s look at one problem that seems to arise quite often regarding basic form study and perception:

The cylinder is one of the most basic geometric solids that is often systematically studied in drawing or painting programs. The cylinder is a bit challenging for the uninitiated draftsman as both the shape and form appear to change significantly as does its orientation to the viewer.

A particularly challenging exercise involving this solid can be found in our own Language of Drawing program. It is known as the Cylinder Wheel exercise. The goal of the exercise is to draft eight strong representations of cylinders in varied orientations radiating from, or orbiting around, a central point. While perspective is often (or should be) taken into account during this exercise–subtle simplifications or biases toward the assumed nature of the cylinder can quickly demonstrate a disparity between the physical world and perception.Take a look at this cylinder (above). Do you notice a difference in size between elliptical base a and base b? Which looks larger to you? It is a subtle effect. The size of a and b are identical but many would see the more ‘distant’ base (b) as slightly larger in this context. Even though the boundaries lines of the cylinder body are parallel (not converging)–the implied depth from other monocular cues (i.e. occlusion of the “distant” base, equal elliptical representation of bases, etc.) triggers this particular perceptual result. While this effect has annoyed many of my students over the years, understanding why it occurs grants great insight into the mechanics of visual perception.

To understand the problem better we can look to a simple configuration of lines known as the Ponzo Illusion.

Standard Ponzo Illusion

The seemingly simple line configuration presented by Italian psychologist Mario Ponzo in 1911 is an effective demonstration of perception at odds with the physical world. The standard Ponzo illusion is configured so that a horizontal line or other figure that is nearer to the interior apex of two converging lines has a tendency to be perceived as greater in length or size as opposed to an identical line or other figure within the converging lines but more distant from the apex. If the standard Ponzo figure is interpreted as a distance or linear perspective cue abstract then an observer will interpret the “inducing lines” of the Ponzo configuration as parallel lines which are in fact converging into the distance in accordance with the effects of linear perspective. In this context, it would be appropriate to assume that two similar objects at different distances can provide equal-sized retinal images only if the more distant object is larger than the nearer.

Variation on Ponzo Illusion

Variations on the illusion demonstrate similar effects. In the above variation we can see that the circle on the right appears larger that the one on the left.  As with the standard illusion, both shapes are identical in size.

While some experiments in the past have manipulated Ponzo line configurations and other geometric “illusions” to downplay the contributions of linear perspective (e.g., Coren & Girgus 1978; Yamagami 1978), many tests were performed that confirmed the impact of depth cues in influencing Ponzo effect judgments (e.g., Gogel, 1975; Kilbride & Leibowitz, 1975; Leibowitz, Brislin, Perlmutter, & Hennessy, 1969; Miller, 1997; Newman & Newman, 1974; Patterson & Fox, 1983; Schiller & Wiener, 1962). I submit that the alterations to the Ponzo configuration that purport to confound intuitive explanations involving linear perspective do not refute the contributions of perspective cues—but seem to reinforce the connection by demonstrating a significant diminishment of the effect as distance and perspective cues are further abstracted.

An additional bolster to the idea of Ponzo’s effect magnitude being reliant on contextual distance/perspective cues can be found with cross-cultural experiments regarding the illusion in Uganda. (Leibowitz & Pick 1972). Reactions to the geometric configurations varied between study groups who were accustomed to “industrialized” environments and groups living in more natural, rural environments. Students from the university responded to the illusion very similarly to U.S. university students, while the rural villagers saw no illusion at all.

The Ponzo effect wonderfully illustrates how even the most subtle contextual cues can greatly influence the perception of stimuli. Even though the elliptical bases are drawn identical and the cylindrical body lines are parallel—the cylinder seems to swell as it recedes. The artist can compensate for this by altering the representation to include depth cues that are aligned with the aforementioned ones (slightly converging lines and a slight size reduction for base b). Again, even though the physical measurements do offer important information regarding a real-world object attributes (i.e. a cylinder with equal circular bases, parallel contour lines, etc…), the representation should be tempered with an understanding of how such an object my be perceived by a biological vision system.

Consider this: Here are 3 cylinders with very subtle cues influencing shape. Which one do you think best represents a cylinder with equal bases in space?

REFERENCES:

Coren, S., & Girgus, J. S. (1978a). Seeing is deceiving: The psychology of visual illusions. Hillsdale, NJ: Erlbaum.

Gogel, W.C. (1975). Depth Adjacency and the Ponzo illusion. Perception & Psychophysics, 17, 12S-132.

Leibowitz, H.W., & Kilbride, P.L. (1975) Factors affecting the magnitude of the Ponzo Perspective Illusion among the Baganda. Perception & Psychophysics Vol. 17 (6),543-548.

Leibowitz, H.W., Brislin, R., Perlmutter, L., & Hennessy, R. (1969). Ponzo perspective illusion as a Manifestation of Space Perception. Science, 166, 1174-1176.

Leibowitz H.W., & Pick H. A. Jr. (1972). Cross-cultural and educational aspects of the Ponzo illusion. Perception & Psychophysics, 12, 430–432.

Miller R. J.(1997). Pictorial depth cue orientation influences the magnitude of perceived depth. Visual Arts Research, 23, 97–124.

Newman C. V., & Newman B. M. (1974). The Ponzo illusion in pictures with and without suggested depth. American Journal of Psychology, 87, 511–516.

Patterson R. , & Fox R. (1983). Depth separation and the Ponzo illusion. Perception & Psychophysics, 34, 25–28.

Schiller P., & Wiener M. (1962). Binocular and stereoscopic viewing of geometric illusions. Perceptual & Motor Skills, 15, 739–747.

Yamagami, A. (1978). Two kinds of apparent size distortion in the Ponzo illusion. Japanese Journal of Psychology, 49, 273–279.