Why no better UFO images — despite billions of phone cameras?

This spine shows: the answer lies in the technology, not the subject. A phone camera is a small miracle for selfies and holiday snaps — and precisely for that reason poor at capturing a small, distant, fast-moving point of light in the sky. Understanding what such an image is missing also explains why “more cameras” does not automatically mean “better evidence.”

AI illustration

A phone’s image sensor is roughly the size of a fingernail and gathers correspondingly little light. Its fixed optics have a short focal length — great for everyday wide angles, poor for the distant: an object in the sky often covers only a few pixels. Optical zoom would require real, longer glass; the pinch-to-“zoom” is usually just digital zoom that crops and upscales — without gaining a single new piece of image information. Why the images therefore stay structurally blurry is also laid out in an overview by New Space Economy.

On a phone almost everything beyond the main lens is at least partly digital zoom: the image is cropped and upscaled back to full size. No new detail can arise from this — only a smoother or estimated version of what were already just a few pixels. That is exactly why a “zoomed-in” UFO often looks mushy rather than sharper.

Most phones use a “rolling shutter”: the sensor does not expose the whole frame at once but reads it out line by line, top to bottom. If an object moves fast within those fractions of a second, it is already elsewhere by the lower lines — straight edges appear slanted, shapes “wobble” (the jello effect). Planes, birds or lightning thus become seemingly exotic forms. Skeptics have resolved exactly such cases step by step, for instance a “beam of light” over a Mayan pyramid as lightning plus rolling shutter.

A phone photo today is not a plain sensor readout but a computed result: multiple frames are combined, noise smoothed, edges sharpened, resolution upscaled by AI. For known subjects — faces, text, landscapes — this works astonishingly well. But a small, unknown point of light fits no learned pattern: it is either smoothed away as noise or “beautified” with plausible-looking yet invented structure. The result looks sharper but often holds less real information. How far this goes was shown by Samsung’s AI-assisted moon photography.

In a narrow sense film is “more honest”: the grain is a physical trace of light, and no software interprets detail in or out after the fact. That is why some old images are harder to manipulate digitally. But it does not mean “everything was better back then”: film is often lower in resolution, offers no instant review, and chemical development brings its own error sources. More honest in detail — not automatically superior.

In 2023 a widely noted experiment showed the principle: a deliberately blurred moon photo was rephotographed with a phone — and suddenly the moon had crisp craters that were no longer present in the blurred original. The device recognised “moon” and overlaid trained texture. For small objects in the sky this is delicate: the camera may add what it “expects” to see, not what was actually there — documented in Samsung’s own description of AI detail enhancement.

The pile of videos is growing — declassified military footage, countless phone clips, dashcams, bodycams. Except: most of them are low-resolution, short and without context. A shaky video shared a thousand times is no better evidence than a single one — it is simply present more often. Evidential weight comes not from repetition but from data quality: scale, calibration, multiple sensors. That is exactly why the robust single piece of evidence stays rare even as the volume of footage explodes.

On 27 April 2020 the US Department of Defense officially confirmed three previously leaked recordings (FLIR, GIMBAL, GOFAST) and made them available through the Navy. Documented is: the videos are genuine military footage, not a fake. Not documented is what they show — they are short, low-resolution infrared clips whose interpretation remains disputed. A perfect example that “authentic” is not the same as “conclusive”; the official release is documented via the Navy.

A single sharp image is never enough — what matters is whether size, distance and speed can be derived from it, and whether the camera is known and calibrated. This is exactly what scientific approaches aim at: fixed stations with camera, radar and infrared, calibrated optics and stored raw data. Who is trying to build such instruments worldwide is covered in Who actually studies UAP?.

The honest answer to the guiding question is: that the images stay poor is exactly what the physics predicts — it argues neither for nor against anything extraordinary. The path to better evidence therefore runs not through even more phones but through better instruments. And the question of what an unexplained object could even be leads on to Must a UFO be extraterrestrial?.