Some art historians may have considered the matter closed. But a fresh clash has surfaced over the painter David Hockney's three-year-old theory that early Renaissance painters used cameralike devices to paint with perfect perspective.

Welcome to the chandelier debate.

In a paper being presented today at the International Conference on Pattern Recognition in Cambridge, England, a Microsoft researcher and a Stanford University computer scientist set out to refute the controversial theory proferred by Mr. Hockney.

Three years ago Mr. Hockney went on "60 Minutes," published a book and began an academic symposium in New York to suggest that as early as 1420, such painters as Jan van Eyck began to use optical devices like the camera obscura or concave mirrors to project scenes onto their canvases for tracing. That, he argued, was the reason for a sudden burst of realism in European painting.

In rebutting Mr. Hockney's theory, Antonio Criminisi, a Microsoft researcher in Cambridge, and David Stork, a consulting professor of electrical engineering at Stanford, seize on one of Mr. Hockney's chief examples, the chandelier in van Eyck's 1434 painting "Portrait of Arnolfini and His Wife."

Mr. Hockney had argued that the chandelier was too perfect to have been produced by hand and eye alone and that van Eyck probably used a concave mirror to project an image onto the canvas.

The problem, as Mr. Stork and Mr. Criminisi see it, is that the chandelier in the Arnolfini portrait is hardly painted in perfect perspective. Relying on digital image registration techniques, Mr. Stork and Mr. Criminisi applied projective geometry to one of the arms of the Arnolfini chandelier to see what the others should look like given the various angles that the painter would have seen them from. Then, as Mr. Criminisi described it during a phone interview yesterday, each painted arm was compared with its ideal perspectival projection; they were not identical.

If the two arms had been painted in perfect perspective they would align well; they did not. Therefore the Arnolfini chandelier is not painted in perfect perspective. If van Eyck had used a concave mirror to project an image of a chandelier onto his canvas and had then traced the image with a pencil, later covering over the evidence with paint, his chandelier should be really accurate. Why cheat if you can't get good results?

Of course, the paper's authors acknowledge that the chandelier itself, van Eyck's model, may have been bent out of shape or imperfectly made to begin with. So they hunted down other examples of period chandeliers like the one in the Arnolfini wedding portrait.

After looking at digital pictures of real chandeliers that had been photographed from "a vantage position similar to that of the Arnolfini chandelier," the scientists decided that the flaws in the van Eyck chandelier were more striking than anything that an inferior craftsman could have turned out.

"We are finding it hard to explain the mistakes," Mr. Criminisi said yesterday by telephone, "and therefore we're left with the only possible explanation -- that they are genuine mistakes and that the artist painted by eye."

The scientists' final test of Mr. Hockney's theory was to have a realist painter render a chandelier as accurately as he could merely by eyeballing it. They enlisted a British realist painter, Nicholas Williams, who painted a chandelier that wasn't perfect but was more accurate than van Eyck's.

"This experiment confirms that realistic-looking structures can be painted merely by eye, without the help of optical tools of any sort," the scientists wrote.

Yet the moral is not what you might expect. "A real master of realism," Mr. Criminisi said, "can paint without real accuracy."

David Hockney, reached by phone in London, was buying none of it. "I don't take it seriously," he said of Mr. Stork and Mr. Criminisi's paper. "They don't grasp that there is bound to be a history of an optical projection. When did it begin?"

Mr. Hockney also pooh-poohed the scientists' conclusions about van Eyck's inaccuracies. If van Eyck had used a projected image, "it couldn't have lasted very long -- only as long as the sun is there," shining on the mirror, he said.

Thus he would not have had much time to sketch the projection well. As for the scientists who created the tempest over the chandelier, Mr. Hockney said, "I gave up on them a long time ago."

Jonathan Crary, a professor of art history at Columbia University and the author of "Techniques of the Observer," was leery of both sides. "I have yet to meet an art historian who was upset by Hockney's theory," he said, adding that he also has yet to meet a scientist who has anything interesting to say about art history. "Scientists are preoccupied with imperfection. They are so out of touch with what making art is about."