Picture a ball flying past you at 90% of the speed of light, and a camera taking its photograph. What shape is the ball in the photo? Most people say squashed flat along its direction of motion, and so do many books and videos. A camera would show something else: a round ball, seen from a different angle, as if it had turned, with part of its back in view. This simulator calculates it live in your browser, following our black hole and pulsar. Every image here is calculated from the equations of special relativity, not drawn or animated by hand.

Two dice. Left, at an everyday speed, facing the camera with one red pip. Right, moving at 90% of light speed: the same size, but it looks turned so its three-pip side faces the camera
Left: a die at an everyday speed, which looks just as it does at rest. Right: the same die moving to the right at 90% of light speed. It isn’t squashed; it looks turned, and you see the face that points back the way it came (three pips).

What you’re looking at

Why turned, not squashed?

Two things happen at once. A moving object really is shorter along its direction of motion: at 90% of light speed, 44% of its length. That is the part the books get right. But a camera does not see an object at one instant. It records the light arriving at one instant, and light from the far side of the object set off earlier, when the object was somewhere else. Light from the trailing side, which would be hidden if the object stood still, left from farther back along the path and still reaches the camera.

For a small or distant object, the two effects combined give almost exactly the view you would have of it at rest from a different angle. A sphere’s outline stays a perfect circle. The angle grows with speed: 6° at 10% of light speed, 30° at half, 64° at 90%.

Overhead view: a camera at the bottom and a die moving to the right along a path 3 metres away. Light leaves the die's back corner first, then its front face after the die has moved on; both rays reach the camera together. When the photo is taken the die is far to the right.
Seen from above, to scale, at 90% of light speed. Light from the die’s back corner leaves first (1); about 3 billionths of a second later, light leaves its front face (2), after the die has moved out of the first ray’s way. Both arrive together, so the photo shows the back and the front side by side. By then the die is far to the right (3).
A die squashed into a narrow upright slab
What you would measure: the die at 90% of light speed, squashed to 44% of its length, as it would look if light arrived instantly.
The same die, normal width, looking turned so its three-pip side faces the camera
What a camera would record: turned, not squashed.

A mistake that lasted decades

Einstein’s 1905 paper says a moving sphere, “viewed from the stationary system”, is a flattened ellipsoid. He meant measured, but for more than fifty years most people took it to mean seen. George Gamow’s much-loved Mr Tompkins in Wonderland (1940) sends a bank clerk to a dream city where light travels barely faster than a bicycle, and draws its cyclists squashed thin.

Anton Lampa had worked out how a fast rod would really look in 1924, but his paper was largely forgotten. In 1959, independently, Roger Penrose proved that a fast sphere still looks round and James Terrell showed that a small fast object looks turned. The first photographs came in 2025, when a team in Vienna staged the motion with laser flashes and an ultrafast camera, so that light seemed to travel at under 2 metres per second. Their cube and sphere looked turned.

Why nobody has ever seen it

Nothing big moves fast enough. The fastest spacecraft ever built, the Parker Solar Probe, reached 191 km/s around the Sun in 2024. That is 0.064% of light speed, and it would look turned by 0.04°. Particles are another matter: the electrons that drew the picture on an old tube television travelled at about 30% of light speed, and a gold atom’s innermost electrons move at more than half of it. Relativity also reshapes gold’s outer electrons’ orbits, so it absorbs blue light and looks yellow rather than silvery. Relativity reaches your phone too: a GPS satellite’s clock loses about 7 millionths of a second a day because of its speed, and gravity adds more; uncorrected, the clock errors would put your position off by about 10 km a day.

More things the calculation shows

Animation vs. simulation

Nothing in the program says “turn the object”. For each pixel, your graphics card follows the light that reaches the camera back in time to the point on the moving object it came from, using the equations of special relativity, then sets the colour from how the motion has shifted that light. The apparent turn, the curved lines and the colours all come out of that one calculation, and each piece of physics can be switched off to see what it does. With light arriving instantly the die is squashed; with no contraction it is stretched and skewed; only with both does it look turned. (The “Speed up” view is a sequence of such snapshots, each of the die moving steadily at one speed, not a single die accelerating.)

Comparing against known results

Where to view it

It runs on your device’s graphics card and works best on a laptop or desktop. It opens on “Fly past”. Then try “Speed up” to watch more of the back come into view as the speed rises, “Globe speeding up” to watch the dots swap places, “Compare all four” to see the squashed and turned versions side by side, “Where is it really?” for the true position, the “Headlight effect” for the true brightness, or the real speeds list, from the Parker Solar Probe to 99.9% of light speed.

Built with Claude Opus 5.5

I built this with Claude Opus 5.5: the physics design, the GPU code, the test programs that check it, and the write-up. Every image on the page comes from the simulator itself, and the method section gives the equations, the tests and the papers behind them.

I’m sure I’ve gotten some details wrong, so corrections are welcome. The images are free to use, for teaching or anything else, with credit. Try it here.