By Amy Silverman
By Olivia LaVecchia
By Monica Alonzo and Stephen Lemons
By Chris Parker
By Michael Lacey
By Weston Phippen
With a lens less than an inch across, the typical camcorder has a rather myopic view of the world. Any light source more than a mile or so away simply cannot be resolved with any detail. Distant lights--streetlights, flares, alien headlights, even--become "point sources." Like the stars in the night sky, there's no detail to be made out in them.
The narrow lens of a camcorder focuses the light of a point source onto an electronic chip, which gets excited, so to speak, and releases a pattern of electrons, called pixels, that is translated into an analog signal which is put on videotape. What eventually comes out is your television's attempt to describe how the electronic chip reacted when it was struck by the light of a distant bonfire, for example.
The actual light from that bonfire is long gone, however, and has nothing physically to do with the electronic signal on your videotape.
Which is a shame. Astronomers have long known that you can learn amazing things from that original source of light.
Unable to reach the stars for tests, scientists figured out how to perform experiments on the light coming from them instead. Using prisms or gratings, astronomers separate that light into its constituent colors, called a spectrum, which allows them to determine a star's chemical make-up. This process is called spectral analysis.
Trying to do spectral analysis on the image produced by a camcorder, however, would be like testing a portrait of Abraham Lincoln for his DNA. The man and his image are two very separate things.
Still, Jim Dilettoso claims to perform just that kind of magic.
On a computer monitor, he brings up an image of Comet Hale-Bopp. The comet has a line segment cutting across it and, in another window, a corresponding graph with red, blue and green lines measuring the brightness of the slice.
He shows similar frames with similar line segments cutting through streetlights, the known flares captured by Channel 12, and the 10 p.m. lights of March 13.
Each results in a different graph.
It's rather obvious that the graphs are simply measurements of pixel brightness in the cross-sections he's taken.
But Dilettoso claims that the graphs show much more. To him, they represent the frequencies of light making up each of the images. He claims he's doing spectral analysis, measuring the actual properties of the light sources themselves, and can show intrinsic differences between video images of streetlights, flares, and whatever caused the 10 p.m. lights.
Because the graph of a known flare is different than one of the 10 p.m. lights, Dilettoso concludes that they cannot be the same kinds of objects.
In fact, Dilettoso claims that the graphs of the 10 p.m. Phoenix Lights show that they are like no known light produced by mankind.
The fallacy in Dilettoso's analysis is easily demonstrated. When he's asked to compare the graph of one known flare to another one in the same frame, he gladly does so. But he admits that the two flares will produce different graphs.
In fact, Dilettoso admits, when he looks at different slices of the same flare image, he never gets the same graph twice. And when he produces some of those graphs on demand, many of them look identical to the graphs of the 10 p.m. lights.
When he's asked to produce an average graph for a flare, or anything that he could show as a model that he uses to distinguish flares from other sources, he can't, saying that he knows a flare's graph when he sees it.
It's an evasive answer which hints at the truth: Dilettoso is only measuring the way distant lights happen to excite the electronic chip in camcorders (which is affected by atmospheric conditions, camera movement and other factors), and not any real properties of the sources of lights themselves.
Met with skepticism, Dilettoso reacts by claiming that his methods have been lauded by experts.
"Dr. Richard Powell at the University of Arizona believes that my techniques are not merely valid but advanced to the degree where there was nothing more that they could add," he says.
Powell, the UofA's director of optical sciences, confirms that he spoke with Dilettoso. "He called here and I talked to him, and I could not, for the life of me, understand him," Powell says.
"I don't know how you take a photograph or a videotape after the fact and analyze it and get that information out. We didn't say that his method was valid, we said we didn't have any other way that was any better," Powell says.
Hearing that Powell denies calling his techniques "advanced," Dilettoso claims that Media Cybernetics, the company which sells Image Pro Plus, told him that the software package would do the kind of spectral analysis he does.
Jeff Knipe of Media Cybernetics disagrees. "All he's simply doing is drawing a line profile through that point of light and looking at the histogram of the red, green and blue. And that's really the extent of Image Pro. . . . Spectroscopy is a different field."
New Times took audio and videotapes of Dilettoso describing his image processing to Dr. Paul Scowen, the visiting professor of astronomy at ASU. Scowen left Great Britain in 1987 and received his Ph.D. in Astronomy at Rice University in 1993; he now uses the Hubble Space Telescope to study star formation.