TEST CASE 3: Van Eyck's Arnolfini marriage portrait 1434

The Arnolfini Marriage, Jan van Eyck, 1434
Click to Enlarge
Detail showing mirror
Click to Enlarge

This picture includes an elaborate chandelier hung centrally from the ceiling, high above the floor Obviously one would expect to be looking up at it. Instead, we find ourselves looking at it end on, as if level with our eyes. One explanation is that the chandelier was traced from a projected image, projected onto the surface of the painting with a concave mirror. Anything as complicated as a chandelier would be exceedingly difficult to draw by eye in perspective. Copying a projected image would be so much easier.

In principle, a simple test can reveal if the chandelier was drawn freehand or from a projected image. The perspective lines on a freehand sketch should not pass through a single vanishing point. For a projected image, which is coherent, they should. That will only hold if the antique Arnolfini chandelier itself is symmetric and not irregular. So a separate part of the analysis must inquire if that is indeed so. Different parties report different results.


Stork demonstrates that the perspective lines do not pass through a single vanishing point. He further calculates how distorted the actual chandelier would have to be for the perspective lines to pass through a single point. He finds that the chandelier would have to be noticeably distorted, much more distorted than would be feasible. He concludes that the chandelier was not drawn by projecting an image.


Falco, in a different analysis, shows that the chandelier is regular, according to the measured radial and angular positions of the six chandeliers. Such regularity would not be found if the chandelier had been drawn by eye. The regularity does not hold for the decorative trim attached to the six arms of the chandelier So perspective lines drawn through the trim will not reflect the overall symmetry of the chandelier. That might underlie Stork's difficulty in finding a vanishing point,

Falco uses, as a calibration, a modern light fixture with the same symmetry. On a photograph of this, the perspective lines do not converge neatly to the single vanishing point expected for perfect symmetry. Instead, for their less-than-perfect symmetry, they form a distinctive pattern. Falco finds a similar pattern for the perspective lines drawn through the candleholders of the Arnolfini chandelier. This indication of the regularity of the perspective supports the idea of a traced image.


Irrespective of the experts, two reasons suggest that the chandelier was traced. It is difficult to sketch a complicated chandelier but they are simple to trace. Second, the chandelier looks odd. We see a sideways image of it yet we view it from beneath. This sideways image could not have been sketched from beneath.

Falco concludes for two reasons that the chandelier was traced. From computer analysis, he finds the chandelier to be regular and symmetric. A sketched chandelier would not be regular because sketching introduces random error. Hence it was traced. Again, he matches the chandelier's distinctive pattern of perspective lines with that from a model system - the projected image (photograph) of modern lighting fixture of the same symmetry. The match shows the chandelier to be a projected image.

Stork (and Falco) find that that the chandelier's perspective lines do not meet at a single vanishing point as is required for a projected image. (This failure could result from a distorted chandelier but Stork shows, by computer analysis, it is not so.) Stork concludes that the chandelier was drawn freehand. (Falco states that the single-point convergence does not apply to an object as complex as the chandelier.)

At issue are the conditions that should apply in the analysis of perspective for complex objects and our inability to assess computer analyses. Notwithstanding, all of the arguments given here conclude that the chandelier was traced, with one exception. If, following Falco, it can be shown that for an object as complicated as the chandelier, perspective lines should NOT converge to a single vanishing point, all arguments support the conclusion that the chandelier was traced. What conditions apply to the perspective lines of complex objects? This is an issue that needs to be resolved.

Before it can be established incontrovertibly that the chandelier in the Arnolfini Marriage Portrait was traced from a projected image, there remains to be shown what can be achieved in practice. Specifically:

  1. A traced image of the chandelier to simulate what is found in the painting.
  2. A freehand sketch of the chandelier to simulate what is found in the painting.

Copyright 2006 Brandeis University
Site by Raka