We Filmed Fourteen Known Aircraft Through An Infrared Pod. All Fourteen Were Tic-Tacs.

REDDING, Calif. — A King Air 350 with nine people aboard sat thirty-one kilometres off our right wing on the evening of 31 July, straight and level, doing 270 knots.

On our recording it is a white lozenge. No wings, no tail, no exhaust plume, hard edges, faintly rounded ends.

We know what it was because we logged its transponder at the second of filming. Its identity, type and altitude are in the file published with this story. It is clip 4 of fourteen, and all fourteen look like that.

What we hired

Ninety minutes on a wildfire-mapping aircraft out of Redding, California, at a disclosed cost of $6,800.

The aircraft carries a mid-wave infrared turret — a gimballed camera in a ball under the nose, sensitive between three and five microns, of the general class carried by the military pods whose footage gets argued about. It is not the same pod. It is the same physics, at a smaller aperture.

Marisol Ekwueme-Trask flew the sensor. She has 2,900 hours on it, mostly finding fire fronts through smoke, and she picked the targets out of whatever traffic happened to be in the area.

The fourteen

Two business jets. Three turboprops in regional airline colours. One four-engined air tanker. Two light singles. A medical helicopter. Four airliners at cruise. One military transport on a published low-level route.

Under about fifteen kilometres you can see what they are. There are wings. On the tanker you can count the engines.

Beyond about thirty kilometres, all fourteen are lozenges.

Two point seven to one

We measured them the dull way. Long axis and short axis of the bright region, taken at half its peak brightness, on ten frames each, by two people who were not shown each other's numbers.

Every aircraft came out at 2.7 to 1, plus or minus 0.2.

The four-engined tanker and the single-engined Cessna give the same shape. So does the helicopter, which has no fixed wing at all.

The shape belongs to the camera

Dr. Perdita Anselm-Okoye designs infrared optics, had nothing to do with the flight, and was sent the raw clips and the turret's published specification without being told what we expected.

Her answer is arithmetic and it takes about a page.

A 110-millimetre aperture working at four microns cannot resolve anything finer than roughly forty microradians. At thirty kilometres that is 1.2 metres on the ground. A twelve-metre aeroplane is therefore a few blurred pixels, and what is on the screen is not the aeroplane at all — it is the instrument's response to a point of heat, its point-spread function, stretched a little along the direction of the scan.

The automatic gain stage then takes that soft blob and hardens its edge, because its job is to make faint things visible, and it does that job on everything indiscriminately.

"The lozenge is the camera's handwriting," she said. "You are not photographing an object. You are photographing what this optic does to any hot thing at that range. Nine people eating sandwiches in a King Air will do it. A wildfire will do it. Venus will do it."

What this does not explain

The shape. That is the whole of what we have accounted for, and this desk is not going to let the easy finding stand in for the hard one.

It says nothing about the motion, which is the claim that actually matters to anybody arguing about this.

Worse than that: two of our own fourteen clips contain apparent lateral jumps we have not been able to explain.

In clip 9 — the medical helicopter, at 38 kilometres — the lozenge moves roughly eleven of its own widths to the left across two frames, then comes back three frames later. In clip 12 the same thing happens once, smaller, on an airliner at 44 kilometres.

The transponder shows nothing unusual. The helicopter was flying a straight line at 130 knots.

We think it is the turret: a gimbal re-centring step, or a frame dropped in the recorder and reconstructed by the encoder. We think that mainly because we cannot think of anything else, which is not a reason. We have not reproduced it on the ground, and Ekwueme-Trask, who has flown this sensor for eleven years, says it is not her gimbal and gives a specific reason. Her reply is printed below in full and it is the best thing on this page.

What we could not do

Clause 7a of this newspaper's standards says we do not report a measurement as a measurement unless it has been taken on two independent instruments, one of them not chosen by the reporter.

We measured on one turret, hired by us, on one evening, over one valley in northern California.

We are reporting it anyway, and saying so here, in the same size type as everything else.