Infrared (FLIR) footage: what it shows and what it does not.
Most of the modern video and the FBI stills are infrared (FLIR), not radar — 67 of 77 analysed clips. Interpreting them requires understanding what the grey values represent and what they do not.
White-hot vs. black-hot
A thermal camera maps temperature differences to brightness. But the operator can flip the polarity at will: in white-hot, warmer objects appear bright; in black-hot, the same object appears dark. So whether a UAP looks white or black is a display setting, not a property of the object. Both conventions appear across this corpus — the FBI infrared sequence happens to be black-hot (dark objects), while most DoD clips are white-hot. The crosshair stays white in both because the HUD is drawn separately.
Brightness is not temperature
These are non-radiometric targeting pods with automatic gain control — the display constantly re-stretches contrast. Converting a grey value to degrees would also require the object’s emissivity, the atmospheric path, and the range. None of that is available (range is typically redacted). Relative thermal contrast (hotter or cooler than local background) is measurable; absolute temperature is not derivable from this footage.
Size, and why “detail” varies
Apparent size can be measured in pixels, but converting to feet needs range — which we don’t have. A target that fills more pixels isn’t a higher-resolution image; it’s simply closer or larger in angle. Most targets span only a few pixels, at or below the sensor’s resolving limit. Multi-frame methods — sub-pixel registration, super-resolution, lucky-imaging — reduce noise and modestly improve sampling, and on the rare close or clearly-tracked objects this does surface real structure (a finite round disc at closest approach, an angular extended form, an apparent rotation — see the Objects analysis). They never exceed the optics’ resolving limit or invent detail; AI upscaling, which synthesises detail not present in the source, is not used.
Reading a point source
what known objects look like in IREvery clip in the release is UAP material, so the useful question is not “what else is in the frame” but what the tracked object itself can and cannot tell us. The reference below is for reading any IR point source; the note states what is actually measurable.
A missing exhaust plume in low-resolution IR at unknown range is not evidence of absence — a distant jet seen tail-on or a small drone can also read as a point. The defensible statement is narrow: the dominant population in this release is compact, high-contrast point sources without conventional propulsion or airframe signatures at the available resolution. Pixel brightness, apparent size and relative contrast are measurable; absolute temperature, true size and range are not.
Rather than catalogue the scenery, the analysis focuses on the tracked object: each is hand-located across its clip, then isolated, super-resolved, stabilised, and played back — which on the clearest clips resolves real structure and motion (a round disc, an angular form, an apparent rotation). See the Objects analysis.
See the video evidence and the methodology for how records are extracted.