To its credit, the Discovery Channel did perform another, and apparently solid, test to the flare hypothesis. The network submitted Krzyston's footage to Dr. Leonid Rudin at the Pasadena image-processing firm Cognitech. Rudin was also given a daytime shot from Krzyston's yard showing the distant Sierra Estrella, which is invisible in the nighttime video. Rudin matched the day and night shots frame by frame, lining them up on a distant ridge. The result: an animation loop showing that the flares are not only above the Estrella, but blink out as they reach the top of the mountains, precisely as distant flares would.
In a "10-Files" episode, KSAZ Channel 10, however, questioned the Cognitech analysis. Krzyston insists to Channel 10 that the objects were hovering below the Estrella ridgeline and couldn't have fallen behind the mountains. Channel 10 suggested cryptically that Cognitech purposely faked its test — "Has the footage been altered? And by whom and why? The mystery continues" — and showed its own test, which a Channel 10 production man claimed took "not long at all," proving that the 10 p.m. lights in Krzyston's video were well below the Estrella ridgeline.
New Times asked Scowen to perform the test himself, using two frames grabbed from Krzyston's original video and a 35 mm daytime photo taken from Krzyston's yard by UFO researcher Dick Motzer. After a half-hour of careful scaling, positioning, and rotation with imaging software, Scowen found a good match for the ridge visible in both shots. His results: The flares are just above the Estrella ridgeline or right at it, just as Rudin at Cognitech had found.
Afterward, Scowen was shown the "10-Files" episode and its claim that Channel 10 matched the frames quickly. He wonders how they could have checked several parameters in only a short time. "You have to make sure that the zoom is set the same way. If it's a standard camcorder, there's no numeric readout of the zoom. . . . Did the guy at Channel 10 match the scale? My guess is that he just laid the two pictures on top of each other."
Rod Haberer, producer of the "10-Files" piece, says that he's "comfortable with what we put on the air." But when he's asked what software the station used to match and scale the daytime and nighttime shots, he admits that they didn't use a computer at all. Channel 10 simply laid one image from Krzyston's video atop another in a digital editing machine.
Scowen says it doesn't surprise him. "We're used to dealing with this with the lay public. People do the minimum until they get the answer they want. In science you have to go back and check and recheck to make sure you're correct. I think Cognitech did a great job," Scowen says.
Rudin says his firm took its job seriously when the Discovery Channel asked it to match the images. "I testify in a court of law routinely; I'm a diplomate of several forensic societies," Rudin says. "Basically, you're talking to the guys who do this for a living."
Told that an astrophysics professor found the Cognitech experiment more convincing, Haberer suggested that his station had merely presented a different point of view, as if the question of a flare falling either behind or in front of a mountain had more than one answer.