Showing papers by "Stuart Anstis published in 2002"

Was El Greco Astigmatic

Abstract: Why did El Greco (1541–1614) paint such elongated human figures? It has been suggested that he suffered from astigmatism. This is an optical defect of the frontal surface of the eyeball, which if over-corrected by a spectacle lens, could have optically stretched his retinal images horizontally, causing him to paint tall, thin objects that looked normal to him. A horizontal optical stretch would have restored his elongated portraits to normal proportions. Similarly, Holbein’s (1497–1543) squat, broad portraits might be attributed to a vertical astigmatism. A logical objection is that an astigmatic defect would stretch the sitter and the portrait horizontally by the same amount, so that the two distortions would cancel each other out. El Greco should thus have produced geometrically accurate portraits. There are also historical objections: El Greco sketched his figures on canvas in pencil in normal proportions and elongated them only when he painted them over; and he painted angels taller and thinner than mortals, suggesting a deliberate mannerism. Instead of relying on logic or history, I converted normally sighted observers into “artificial El Grecos” using a specially constructed cylindrical-lens telescope to stretch their visual images horizontally by 30%. Five subjects looked through this telescope in turn, with the other eye patched. Upon attempting to draw a freehand square from memory, they drew a tall, thin rectangle elongated vertically by 35%. When they were shown an actual square and asked to copy it, they drew a perfect, square copy—although both original and copy looked like squat, wide rectangles to them. So freehand portraits “from memory” showed an El Greco effect, but copy portraits “from life” did not. I conclude that subjects drew freehand shapes that appeared as square retinal images and drew simple facsimiles when asked to copy a shape. Logically, then, El Greco could have painted accurate life studies and distorted portraits from memory. To simulate El Greco’s supposed lifelong astigmatism, a volunteer was persuaded to wear the El Greco telescope over one eye for 2 days, with her other eye covered. At night she was blindfolded. Four times a day, she drew both a copy of a square and also a freehand square. Her copied squares were perfectly square. On the first day, her freehand squares were 50% too tall, but they became squarer every day because she rapidly adapted to the optical distortion, and after two days she drew as though she had normal vision (see Fig. 1). I conclude that even if El Greco were astigmatic, he would have adapted to it, and his figures, whether drawn from memory or from life, would have had normal proportions. His elongations were an artistic expression, not a visual symptom.

Visual adaptation to a negative, brightness-reversed world: some preliminary observations

Abstract: There have been many studies of visual adaptation to spatial rearrangements, starting with Stratton's (1897) classic studies on adaptation to an upside-down world. These have been reviewed by Rock (1966) and Howard (1982). Luminance information is crucial to such visual tasks as extracting shape from shading and recognising faces. If a picture of bumps and hollows is turned upside down, or reversed in brightness, the perceived depth reverses (Ramachandran 1988). Cavanagh and Leclerc (1989) have shown that shadows are treated as such only if they are darker than unshadowed regions. Extraction of depth by shape from shading seems to be an early process which precedes perceptual grouping (Ramachandran 1988a, b) and pop-out in visual search tasks (Enns 1990). Is shape from shading affected by perceptual experience? Hershenberg (1971*?) showed that chicks reared with grains lit from below preferred to peck at photographs of grains lit from below versus lit from above. If humans adapt to reversed luminance, will they "unlearn" that light comes from above, or that light is brighter than darkness? In recognising faces, why is it so hard to recognise photographic negatives of famous faces? The Fourier power spectra are identical to those of positives, and the phase spectra nearly so. The difficulty arises only with 3-D lith (black/white) or half tone (gray scale) portraits, not with outline drawings. Probably, luminance reversal disrupts shape from shading which is a crucial step in recognising the 3-D shape of faces. In order to learn whether humans could perform these tasks when normal brightness relations were disturbed, we have studied visual adaptation to a world that was reversed not in position but in brightness. Adaptation procedure. We examined the effects of long exposure to a visual world in which brightnesses were reversed by means of a closed circuit video link. During passive adaptation the observer watched TV in negative. During active adaptation he walked about, or sat and viewed his own hands, conversed and interacted with others, while viewing a negative monitor fed from a TV camera. Perceptual phenomena studied before, during and after this adaptation included the perception of highlights and shadows and perception of depth in convex and concave face masks which were lit from above or from below. The observer was also confronted with negative TV images and asked to extract 3-D shape from shading, to recognise facial emotions such as anger, surprise or happiness, and to identify celebrities from their negative faces. It was observed that....*

The Boogie-Woogie Illusion

Abstract: A grid of vertical and horizontal lines, each composed of light and dark squares, is moved rigidly at 45° to the vertical on a gray surround. When the luminance of the background is set midway between the luminances of the light and dark squares, the squares appear to race along the lines even though they are actually ‘painted’ on the lines. The effect arises from the unequal apparent speeds of the lines and their textures. The light and dark squares along the lines define a first-order pattern whose apparent speed, parallel or along the line, is close to veridical. The lines themselves have no overall luminance difference from the background so that they are defined by a second-order difference. As reported elsewhere, apparent speed is reduced for second-order motion so that the motion perpendicular to the line is perceived as slower than the motion along the line even though they are physically equal. The imbalance creates the impression that the small squares are moving along the lines rather than movi...