Words acquire authority when they sound instrumented. “Cold orb” sounds much more physical than “dark blob.” That makes it useful shorthand and dangerous evidence.
Thermal imagers do not detect coldness. They detect infrared radiance in a spectral band. A processor then maps that radiance to visible brightness. Depending on palette and settings, high radiance may appear white or black. The resulting picture can be spectacularly useful while still being several steps removed from a thermometer.
Temperature is only one variable.
For an opaque surface, the radiation received by a thermal camera can include radiation emitted by the object itself and radiation reflected from its environment. The proportion depends strongly on emissivity: how efficiently a surface emits thermal radiation compared with an ideal blackbody.
A high-emissivity painted surface and a low-emissivity polished surface can have the same actual temperature and look very different to an infrared instrument. The polished surface can also reflect infrared radiation from the sky, ground, clouds, engines, or other surroundings. This is why industrial thermography requires more than pointing a camera and believing the color bar.
The atmosphere gets a vote.
Radiation between an airborne target and a sensor must cross air. Water vapor, carbon dioxide, aerosols, haze, and other constituents affect transmission. The influence grows with range and varies with spectral band. A distant source can lose contrast or appear altered because the atmosphere is participating in the measurement.
This matters especially when the range itself is uncertain. If we do not know whether a source is hundreds of meters away or many kilometers away, we do not merely have uncertainty about distance. We also have uncertainty about how much atmosphere processed the signal before the detector ever saw it.
Contrast is not absolute temperature.
Military electro-optical systems are designed to make targets visible, not to produce aesthetically faithful photographs. They may adjust gain and level automatically. The same target can therefore appear brighter or darker as the scene changes. A cold sky, warm ground, cloud layer, or different viewing angle can change the entire display mapping.
In AARO's PR120 description, the office notes that the sensor changes modalities and the area of contrast loses distinctiveness against the background. That detail is interesting precisely because it shows the target appearance is coupled to sensor mode.
If one sensor state makes the object conspicuous and another makes it less distinct, the correct response is not “therefore fake” or “therefore cloaking.” The correct response is: identify what changed in the measurement channel.
What would a real “cold object” require?
Suppose, for argument, that the reported object truly had a lower physical temperature than the surrounding air or background. Several possibilities would still remain. It could be a material with little internal heat generation. It could be recently exposed to a colder environment. It could have strong evaporative or expansion cooling. It could have unusual surface properties. Or the apparent contrast could still be partly reflective rather than purely emissive.
An object moving through atmosphere also experiences aerodynamic heating and heat transfer. At modest speeds that may be small. At high speeds it becomes important. If somebody claims a compact object is traveling through dense air at extreme velocity while remaining thermally inconspicuous, that is no longer merely an observation. It is an engineering claim about drag, boundary layers, heat flux, propulsion, and energy disposal. Those claims can be tested.
The interesting question is not “is it cold?”
The better question is: what combination of radiance, surface properties, atmosphere, sensor band, image processing, and range could generate the observed contrast?
That formulation generates experiments. It lets analysts compare candidate objects, simulate atmospheric transmission, inspect known sensor behavior, and ask what would be different if the target were a balloon, drone, bird, debris, cloud feature, distant aircraft, or something genuinely unfamiliar.
Size estimates need their own audit.
The reported approximate four-foot diameter is also worth treating as a separate claim. Physical size cannot be recovered from angular size without distance. If range came from radar, laser rangefinding, stereo geometry, platform telemetry, or another sensor, that is valuable. If it came from a rough visual estimate, its uncertainty may be much larger.
So the public question becomes simple: what measurement produced the four-foot estimate? A useful investigation should preserve the answer instead of allowing a convenient number to harden into folklore.
“Cold orb” is a reported description, not yet a public thermodynamic result. The available footage supports the existence of an infrared contrast feature. It does not, by itself, establish the object's absolute temperature, material, size, or mechanism. Those require calibration, range, metadata, and preferably native sensor data.
What data would move the needle?
A calibrated radiometric feed would be ideal. Failing that, useful information would include sensor band, gain and polarity state, whether automatic gain control was active, field of view, platform and target geometry, range source, atmospheric conditions, and simultaneous radar or other EO detections.
The lesson is broader than UAP. Instrument displays are interpretations built for a task. Treating them as direct physical reality is how technically interesting mysteries turn into arguments about pixels.