Simulation bench. Pick a scenario, then change how far away you think the light was. Speed, acceleration and size all scale with that one guess.
Video bench. Analyse the built-in clip or import your own .mp4. The ring follows the brightest moving feature, and the readouts below use its path against the background.
Sky simulator
Scenario
Choose a video. It is analysed as soon as it loads, and the ring follows the brightest moving feature. If it picks the wrong one, pause, click the right one and press Track from click.
Frame width depends on the phone and zoom. As a rough guide for an iPhone held sideways: about 69° at 1×, 38° at 2×, 17° at 5×. Held upright it is narrower, about 43° at 1×. Handheld shake is cancelled using the background where there is detail to lock onto, such as trees, wires and hills. Against a plain sky it cannot be, so a steady camera works best.
Angular speed of the lead object, last 12 s of playback. Playback speed changes how fast you watch, not the numbers.
What the witness reports
What was it?
Pick a distance first, then check your guess.
Photo bench. Measure a photo of a light in your browser, then answer the questions a photo cannot show. The two submitted frames are on the second tab.
Identify an image
Add a photo of a light or object in the sky. The page measures it in your browser and ranks ordinary explanations by how well they fit. Nothing is uploaded. Answer the questions a photo cannot show, such as how it moved and whether it made a sound.
What the photo cannot show
Only the time of day, and daylight or several lights, are filled in from the image. They are the same answers the triage tool below uses, so change them there if they look wrong.
Measured in your browser
Likely explanations
A rule-based ranking from the measurements and your answers. A still photo cannot show movement, sound or duration, so those start at Not sure. Change Evidence to Photo or video only if nobody saw it by eye.
Check next
Submitted photos
Two frames supplied for this site, one at night and one by day. Each entry says what the frame itself can and cannot tell us.
Photo 1 · Night, purple light by a tree
What the frame shows
- ObjectOne compact light, about 20 pixels across, with a white core and a purple edge. It is not resolved into a shape.
- ShapeThe squarish outline comes from compression blocks and pixel interpolation. A light this small does not record its real shape.
- ColourPurple fringes around a blown-out white core are common when a small bright light overloads a phone sensor. Strobes that alternate red and blue can also blend to purple.
- SceneA tree silhouette on the right gives a horizon reference, so the light is in open sky and not on the tree. Its height above the horizon cannot be read without a wider frame.
- Distance and sizeNot recoverable. With no known landmark, speed and size could be anything, as in the simulator above.
- WitnessThe light flashed many different colours. Near the end of the video it showed a disc-like triangle, with no sound and no recognisable drone body. This frame caught one purple moment, and the video was not available to check. At night and at a distance only a drone’s or aircraft’s lights show, not its body, and drones are usually inaudible beyond a few hundred metres. Three separate lights can also be joined into a triangle by the eye and by video compression. You said it was fairly close, which weakens the distant-drone explanation. A star is unlikely given the shape. It also kept stopping and starting for about ten seconds, then left very fast. That rules out a fixed-wing aircraft, which cannot stop in the air, and a star or balloon. A helicopter can hover but would be loud. A drone can stop and start, so the open question is the speed. A close object looks faster than a far one at the same speed, and consumer drones reach about 20 m/s while racing drones pass 40 m/s. The photographer estimated about 65 m/s at full speed (about 235 km/h, Mach 0.19), judged from where it was. That is above a typical consumer drone and above most racing drones. Purpose-built speed-record drones have gone faster. It is a visual estimate from a guessed distance, so treat it as rough: at a different distance the same motion gives a different speed, as in the simulator above. The simulator also has a reconstruction of this sighting, built from the photographer’s account.
Photo 2 · Daytime, pale oval in clear sky
What the frame shows
- ObjectOne small, pale yellow-white oval, about twice as wide as it is tall, with a darker underside and a slightly ragged lower edge.
- ShapeThe disc outline is soft all round. Enlarging a distant speck this far blurs it into an oval whatever it was, so the outline is not evidence of a disc.
- LightThe top is brightest and the underside is shaded. That fits a solid body lit by the sun from above, such as a sunlit aircraft, balloon or bird, more than a self-lit object.
- ScenePlain blue sky, no horizon, cloud or landmark. Nothing gives distance, size or height.
- WitnessThe photographer also saw it moving by eye, not only through the iPhone. That rules out a smudge, reflection or insect on the lens. The soft edges still come from the zoom.
Most likely, from these frames alone
Photo 1: unidentifiable. It was fairly close and silent, flashed many colours and showed a disc-like triangle with no recognisable drone body. It stopped and started for about ten seconds, then left very fast. Photo 2: a sunlit aircraft, a balloon or a bird, because it was seen moving by eye. A single still cannot separate these. Open the triage tool above with the original details, then try these checks.
- Original files with metadata: time, exposure and lens, so the bearing and zoom can be worked out.
- A wider shot, or video, to show whether it moved against the tree and the stars.
- Whether it blinked or changed colour, and whether a sound was heard.
- For photo 1: the video itself, or frames from the end that show the shape; whether the colours repeated in a pattern; whether the light moved against the tree, and whether the triangle’s corners were separate lights or one solid outline.
- For photo 2: which way it moved, whether the path was straight and steady or wandering, whether it flapped or rocked, how long it was in view and how big it looked against a fingertip held at arm’s length.
- Sky charts and flight-tracker archives for that place and minute.
Triage a sighting
Answer what the witness saw. The tool ranks ordinary explanations by how well they fit. The weights are hand-set for teaching, so treat the result as a place to start checking, not a verdict.
Check next
Case files
Well-known cases in date order, tagged by where the public record leaves them. Explained means an official or well-supported finding. Disputed means credible explanations exist but people still disagree. Unresolved means the case is officially open or unexplained. Programme marks government efforts to study the subject.
How sightings are classified
The astronomer J. Allen Hynek, a consultant to the US Air Force's Project Blue Book, laid out a scheme in his 1972 book The UFO Experience. It sorts reports by what was seen, not by what it was.
NLNocturnal light
Lights in the night sky with no visible body. The most common kind of report.
DDDaylight disc
Objects seen by day, often described as discs, ovals or eggs.
RVRadar-visual
A radar return that matches something seen by eye. Two sensors help, though radar has false returns of its own.
CE1Close encounter, first kind
A sighting within about 150 m (500 ft). That distance is a rule of thumb.
CE2Close encounter, second kind
Physical effects reported: scorched ground, stalled engines, interference.
CE3Close encounter, third kind
Occupants reported. Later writers added a fourth kind for abduction claims. Hynek's own scheme stopped at three.
Field kit: what to log
- Bearing and elevation. A fist at arm's length spans about 10 degrees and a thumb about 2.
- Start and end time, with the time zone.
- Whether it stayed fixed against the stars. Stars turn slowly together, so a light that holds still against them is often a planet.
- Weather, cloud and moon phase.
- Any sound, and how many witnesses stood where.
- Original, unedited video or photos with metadata, and whether a second camera saw it.
The Drake equation
Frank Drake wrote this in 1961 to organise one question: how many civilisations in the Milky Way could we detect? It says nothing about whether any have visited. Move the sliders and see how a few unknown terms swing the answer across many orders of magnitude.
Night shift
You have the observation deck until dawn. Spot lights in the sky, lock on, read the instruments and file what each one is. At dawn, say whether there was a UFO over the station. It runs in 3D in your browser, so it needs WebGL and takes a few seconds to load its models.
Best on a desktop or a recent phone. Older browsers should use legacy mode.
Sky chart
See what was really overhead. Pick a place and a time, then enter the bearing and height of a light if you have them. The chart shows the Moon, the five bright planets and the brightest stars, and which of them sits closest to your light. Positions are approximate, good to within about a degree, which is enough to rule Venus or Jupiter in or out.
You are looking up: north at the top, east on the left. The centre is straight overhead and the rim is the horizon.
What lines up
Enter a bearing and height to compare your light with the sky.
Above the horizon now
Radar ops
A practice scope. Contacts cross your sector. Click one to read its track: speed, height, transponder and how it moves. Then say what it is. Most are ordinary. A few behave in ways nothing ordinary should. Time runs fast so a crossing takes under a minute.
Field calculators
The sums a sighting report usually needs: how big, how fast and how far. Each one shows its working, and each result can go in the log book. Everything runs in your browser.
Angular size to real size
The full Moon is about 0.5°. A thumb at arm's length is about 2°. An object's true size cannot be known without its distance.
Speed from a crossing
This assumes it crossed at right angles to your line of sight. If it came towards you the real speed is higher, if it was moving away from the line it is lower.
Triangulation from two observers
Both observers must be looking at the same light at the same moment. Angles near 0° or 180° give poor results, and the two angles must add to less than 180°.
How far if it were a…
A single apparent size fits many objects at different distances. This table is why a size without a distance proves nothing.
Unit converter
Arm's length gauge
Hands vary. Measure yours once against something of known size, such as the Moon or a building you can measure the distance to.
Witness interview
A structured interview form, in the order an investigator would ask. Fill in what the witness says, in their words. The report below builds as you go, with follow-up questions based on the answers. Nothing leaves your browser. Names stay in the report only if you type them.
Follow-up questions
Report
Reference library
Programmes, cases, terms and method, in short entries you can search. It keeps to what is documented and says so where a matter is disputed. For the rest, read the sources.
Observer record
Your standing in the lab, worked out from your log book. Every simulator run, track, measurement, interview and night shift earns points. Rank and badges follow the log book, so loading a .sav file restores them.
Points by activity
Badges
Craft gallery
Models of the three shapes that turn up most in reports, with notes on what people describe, what they usually turn out to be, and what would settle a case. The models are artwork, not evidence, and no photograph of a real object is shown.
What reports describe
A large dark triangle or boomerang, usually seen at night. A light at each corner, sometimes one more in the middle. Most witnesses say it was low, slow and quiet. Many describe a dark area that blocked the stars behind it.
What it usually turns out to be
Aircraft lights seen from below or at an angle can make a triangle. So can three or more aircraft or drones flying together, and lit kites or balloons. A stealth bomber has a triangular outline and flies over parts of the US and the UK.
What would settle it
Times to the minute. Bearings from two places. Whether the lights kept the same pattern. Any sound at all. Flight data for the time and area, and whether the dark area really hid stars.
Reports worth reading
The Belgian wave, 1989 to 1990, when many triangles were reported over several months. Phoenix, 13 March 1997, when thousands saw a line of lights. Later that night flares dropped by military aircraft explained a second set of lights. The first is still argued over.
What reports describe
A smooth white capsule, roughly 12 to 14 metres long, with no wings, no tail, no visible exhaust and no markings. The name comes from pilots who compared it to the sweet. The best known reports are from the Navy in November 2004, and an infrared video from that period.
What it usually turns out to be
Aircraft or contrails seen nose-on, balloons, birds and distant objects whose size and range are guesses. Infrared cameras show glare and rotating artefacts. Range and speed worked out from one camera are easy to get wrong, because one view gives angles, not distance.
What would settle it
Range from a second sensor. The camera's own metadata, field of view and zoom. Radar tracks stored with times. The full video, not a clip, and the pilots' notes written at the time.
Reports worth reading
The Navy videos known as FLIR1, Gimbal and GoFast. The Department of Defense confirmed in 2020 that they are genuine Navy footage. That confirms where they came from. It does not say what the objects were.
What reports describe
A round or oval object with a flat profile. The word saucer comes from a 1947 newspaper phrase. The pilot at the centre of that report described how the objects moved, not what shape they were.
What it usually turns out to be
Lens flare and ghost reflections show up as round blurs. A small object close to the camera looks large. Lenticular clouds look like stacked discs. Weather balloons lit by the sun, bright Venus, kites and thrown objects all appear in files.
What would settle it
The original image file with its metadata. Other frames from the same place. Whether anything else in the shot carries the same blur. A sky chart for the time, to rule Venus or Jupiter in or out.
Try it here
The Photo bench measures an image and ranks ordinary explanations. The Sky chart shows whether a planet was where your light was.
Log book
Simulator runs, video tracks, photo measurements, triage results, Drake settings and night-shift case files are logged here. Entries stay in this browser. Save a .sav file to keep a copy or move your book to another device, and load it back any time.
If your browser blocks the download
Copy this text into a plain text file and name it with a .sav ending. Loading it back works the same.