In his classic science-fiction story “The New Accelerator,” published in 1901, H. G. Wells describes a drug that speeds up a person’s metabolism by a factor of 1,000. For the two protagonists who valiantly test the potion, the world appears strangely slowed down, almost frozen in movement. The story got one of us (Schattschneider) thinking: If we could slow down time, could we see single photons fly through space? Could we observe relativistic phenomena? In particular, could we ever glimpse a strange prediction called the Terrell-Penrose effect?
The Terrell-Penrose effect would make objects moving at nearly the speed of light look oddly rotated. The notion seems to go against another prediction of Einstein’s special theory of relativity known as Lorentz contraction, which holds that as things go faster they will shrink. Although the Terrell-Penrose effect had been tested in thought experiments and simulated on computers, it had never been demonstrated in real life.
The prospect of real-world testing lay dormant until recently, when one of Schattschneider’s colleagues, quantum scientist Philipp Haslinger of the Vienna University of Technology, mentioned to him an experiment called the SEEC project, which aims to visualize the way light moves across surfaces. He shared a video in which a laser pulse seems to move at a speed of meters per second, only about one billionth of the speed of light. There it was again: the idea of slowing down time—Wells’s New Accelerator, this time in the form of not a magic potion but ultrafast photography.
On supporting science journalism
If you’re enjoying this article, consider supporting our award-winning journalism by subscribing. By purchasing a subscription you are helping to ensure the future of impactful stories about the discoveries and ideas shaping our world today.
But the scenes in the project were stills; the object didn’t move. What if we accelerated the subject being photographed to a speed close to that of the laser? Would we then see Lorentz contraction? Or would we instead see the even stranger Terrell-Penrose effect? Almost immediately we hatched a plan for an experiment. When two of us (Schattschneider and Juffmann) met up in Juffmann’s laboratory, we found that we were both inspired by the Wells story.
Schattschneider teamed with the SEEC project (Haslinger, Juffmann and artist Enar de Dios Rodríguez) and with master’s students Helm and Dominik Hornof to demonstrate the Terrell-Penrose effect in a lab for the first time.
If we could pull it off, we would see an element of relativistic physics that had never been observed. We’d also show that relativity is still offering surprises more than a century after the theory was first introduced.
To understand what exactly the Terrell-Penrose effect is, we first need to consider the Lorentz contraction, one of the more puzzling predictions of special relativity. According to this principle, the length of an object moving at speed v shrinks along the direction of motion when measured by a stationary observer. The compression factor is √1 − v2/c2, where c is the speed of light.
Could a distant observer detect the compression? Austrian physicist Anton Lampa, inspired by Albert Einstein’s concept of using calibrated rods to measure distance, discussed this question in 1924. He found that the different travel time of light from one or another end of the rod to the observer obscured the effect. For Lampa, the visual appearance of the Lorentz contraction was sort of an unwanted side effect to be eliminated. This was probably why his groundbreaking work did not receive the attention it deserved.
Dutch theoretical physicist Hendrik Lorentz, working in the 1920s and 1930s, believed the contraction (which would later be named after him) would be visible. This assumption was not widely questioned until three decades later, when English mathematician Roger Penrose and American physicist James Terrell, working independently, both arrived at a surprising conclusion: The Lorentz contraction is not visible. An object moving at nearly the speed of light wouldn’t appear shortened. Instead it would look rotated. This intriguing result, published in 1959, came to be known as the Terrell-Penrose effect.
The optical illusion occurs because the light an observer sees from an object did not all get reflected off that object simultaneously. Light from the far side had to start its journey a bit earlier than light from the near side. For slowly moving objects, this difference has no effect. But imagine the object is moving incredibly fast.
In the small amount of time it takes light to travel just one meter, the object will have already moved noticeably. The light that reaches your eyes simultaneously from different points originated at different moments in the object’s journey—creating the illusion of rotation and elongation. In the end, however, we don’t see the elongation: interestingly enough, it is exactly compensated by the Lorentz contraction, yielding a purely rotated image of the object.
This striking effect had never been observed because the speeds required are astonishingly high, far beyond what we can achieve with macroscopic objects in a lab. As a result, the Terrell-Penrose effect had long remained a theoretical prediction. But with the technology of the SEEC project, we’ve escaped these limitations. By using ultrafast lasers, high-speed cameras and precision-timing systems, we mimicked relativistic speeds and made the Terrell-Penrose effect visible for the first time. Our results were published in Communications Physics in May 2025.
Our experimental setup relies on a few essential modern technologies. The first is a pulsed laser that emits bursts of light just one picosecond long—that’s 0.001 billionth of a second. Each pulse travels outward like a thin, spherical shell of light. This light scatters off the object we want to image, and the reflected light is collected by the lens of an ultrafast camera.
That camera is the second crucial piece of technology. One of the first attempts to capture motion at high speed was made by English photographer Eadweard Muybridge in 1878. Using a series of fast exposures, he proved that at some point in a gallop, all four of a horse’s hooves leave the ground. His cameras reached shutter speeds of about a millisecond—incredibly fast for the era. Today we’ve achieved exposure times orders of magnitude shorter—down to picoseconds or even femtoseconds. The camera we used has an exposure time of only 0.3 billionth of a second (that’s 300 picoseconds).
It relies on what’s called a gated image amplifier. In this device, an incoming photon hits a photocathode, where it is converted to an electron via the photoelectric effect. If the gating is on, the electron is accelerated toward a microchannel, where many consecutive collisions with the channel walls create a cloud of secondary electrons. These then hit a phosphor screen, which converts them back to photons that are detected by the CCD camera. The overall effect amplifies the light of each original incoming photon into multiple photons at the end point.
These tools are used by the art and science project SEEC Photography. The project visualizes how light moves across objects—a process so fast it is invisible to the human eye.
Human eyes work by creating images on our retinas when light scattered from objects reaches them. When an object is illuminated, regions of it that are farther from us will be imaged later than those that are closer. This time difference is tiny—for a spatial separation of one meter, it amounts to three billionths of a second (0.000000003 second). That delay is imperceptible to humans. But when we use a camera with an exposure time of less than one billionth of a second, we can see the effect.
To record this phenomenon, the SEEC project imaged several scenes, including one of a dog skeleton. The camera captured a series of frames, each one taken at a slightly different time with respect to an incoming laser pulse. In effect, each photograph captured a different “slice” of the skeleton—the area momentarily illuminated by the shell of light. This process enabled the project team to reconstruct light’s movement across the surface as if time were slowed down. One bizarre implication is that the image of the object and its shadow will no longer coexist simultaneously.
To visualize the Terrell-Penrose effect, we just needed to apply this trick to a moving object. We carried out our test at Juffmann’s lab at the University of Vienna. First we arranged the laser, the camera, and the stage where the object would move. We installed SEEC’s intensified camera, which the team had bought on eBay several years ago. To our delight, the device worked flawlessly—although it took some work for the one of us overseeing the setup (Helm) to get used to the control software and an operating system older than she is. We then had to cope with limited lab space: to achieve the path we wanted, we had to steer the pulsed laser out of our lab, across a hallway and into a lecture hall on the other side of it. This setup restricted our time slots to the weekends.
Once we had established the pulsed laser illumination, we placed two objects, a sphere and a cube, on a movable cart at the front of the lecture hall. Hornof built the objects from materials bought at a hardware store. To mimic the Lorentz contraction that would be happening if they were truly moving at relativistic speeds, he intentionally compressed them along the axis of movement. (Without this step we would have seen elongation in addition to the rotation the Terrell-Penrose effect should produce.)
This image from the experiment shows Terrell-Penrose’s rotating effect on a Lorentz-contracted cube that appeared, through clever camera tricks, to be moving at 80 percent of the speed of light.
From “A Snapshot of Relativistic Motion: Visualizing the Terrell-Penrose Effect,” by Dominik Hornof et al., in Communications Physics, Vol. 161; May 1, 2025 (CC BY 4.0)
We began by taking a sequence of 32 ultrafast photographs of both objects while they were stationary. For each photograph, we changed the timing between the laser pulse and the camera’s shutter so that each image captured light from a different “slice” of the object. This created a time-lapse series of light traveling across the object’s surface, exactly as the SEEC project did. We changed the timing by 400 picoseconds in between illuminating each next slice, corresponding to a distance of six centimeters between slices.
When imaging the contracted sphere, we moved it six centimeters between each recording. Effectively, the sphere appeared to travel at a speed of six centimeters per 200 picoseconds, which is 99.9 percent of the speed of light. We repeated this process 32 times and combined the recordings into one snapshot of the object. The result? The sphere, which we had flattened into a circle, appeared rotated and spherical in the snapshot, just as Terrell-Penrose predicts.
The outcome with the cube was similar. In this case, we moved the object five centimeters between each recording, mimicking a speed of 5 cm/200 ps—roughly 80 percent of the speed of light. Again, our resulting snapshot showed the cube rotated, in excellent agreement with the prediction from Terrell-Penrose. We found it interesting that the cube’s vertical edges also appeared curved as hyperbolae—a prediction made back in 1970 by Ramesh Bhandari.
Our result shows that we can study certain relativistic effects in a lab by artificially reducing the speed of light. The Terrell-Penrose effect is confirmed: “Lorentz-contracted” objects appear rotated, not contracted.
Our technique opens the door to testing other relativistic effects. Could we use similar tricks to see time dilation or the strange relativistic displacement of starlight called stellar aberration? Might we be able to enact Einstein’s thought experiment about lightning strikes seen from a moving train, which shattered the idea of absolute time and simultaneity?
Ultimately we transferred Wells’s dream of slowing down time into real life. Our experiment revealed an aspect of physics never seen before, thanks to a serendipitous combination of art, science and science fiction.
