Most visual illusions are produced using carefully contrived drawings or gadgets to fool the visual system into thinking impossible things. Recently, waiting at a train station, I encountered a real-life Ponzo illusion.
The traditional form of the Ponzo illusion is produced by drawing a pair of receding railway lines. The context suggests different depths in the drawing. An object towards the top of the drawing appears larger than an identical object near the bottom of the drawing. Using a principle of size constancy, the visual system estimates the size of any object as its retinal size multiplied by the assumed distance. Thus, the ‘most distant’ of the two identical yellow lines appears to be longer.
The setting of this new Ponzo illusion is a railway station situated at Vitrolles Airport, Marseille (see photo below). The station has glass panelled shelters on the platforms on each side. The glass panel at the front of each shelter displays two rows of grey rectangles. Apart from their decorative function, one assumes that these rows of rectangles are intended to help prevent people from walking into the glass panel as they move in and around the shelter. The photo below shows the arrangement of the two rows of rectangles on the shelter.
The stimuli for the illusion consist of rectangles that are slightly longer than a credit card, approximately 10.0 cm long x 1.5 cm wide with a separation of about 3.0 cm between successive rectangles. The plate glass window is about 5 mm thick and is marked with rectangles on both sides of the glass in perfect alignment so that a 3-D effect is created indicating a false sense of solidity to these rectangles. This ‘3-D look’ may strengthen the Ponzo effect illustrated below.
The illusion is demonstrated in the photo below. Two people sitting directly in front of the shelter are waiting for a train. The upper set of rectangles appears as a set of columns positioned along the railway lines at a distance of approximately 7 metres in front of the two passengers. In this case, the upper set of rectangles appear to have a height of around 2-3 metres. The lower set of rectangles are perceived at their correct location and size on the plate glass window, behind the two passengers. The lower set are actually physically smaller, owing to the camera angle, but the illusion exaggerates the size difference enormously.
Further illustration of the effect indicates how the brain scales the stimuli to the context. When the rectangles are projected onto the opposite platform they appear huge – almost as high as the lamp post of around 5 metres.
When the rectangles are projected onto the nearby platform, however, they appear proportionately smaller (1.0-1.5 metres).
Owing to the camera angles, the actual size of the rectangles in the upper picture is larger (5-10%) than in the lower picture, but nowhere near the illusory ‘expansion’ that takes place when they are projected by the brain to the opposite platform.
Blocking the Distance Cues
The magnitude of the Ponzo illusion becomes somewhat indeterminate when the distances cues were fortuitously blocked by a passing freight train. In this case the rectangles are ‘drawn into’ the scale of the passing wagons, stretching in size beyond the appearance when the wagons are not there.
The Ponzo illusion can be most easily explained in terms of linear perspective. The rectangles look longer when they are projected to the distance of the opposite platform because the brain automatically interprets them as being further away, so we see them as longer. An object located farther away would have to be larger than a nearby object to produce a retinal image of the same size.
The more visual cues surrounding the two vertical lines, the more powerful the illusion. The passing freight train obliterated some of the distance cues and so the length of the lines was more difficult to assess.
A new post explores the possibility that the illusion described above may be more than another form of the Ponzo Illusion.
One thought on “A New Ponzo Illusion (or something more?)”
You must log in to post a comment.