Apparent motion
A Socratic walk-through of apparent motion — reasoned out one step at a time, not lectured.
The question we started with
THE QUESTION #Why does a rapid series of still pictures look like continuous movement?
A film is a sequence of still photographs, each one motionless, shown one after another. Nothing on the screen ever moves across it. Yet you do not see a rapid slideshow — you see a horse gallop.
The explanation almost everyone has heard is persistence of vision: the eye holds each image briefly, so the images overlap and blend into continuity. It is a satisfying story, it appears in a great many books about cinema, and as an account of motion it is wrong. Seeing why it fails is the fastest route to what is actually happening.
Reasoning it through
REASONING #Take the persistence story at its word. Suppose an afterimage lingers, so frame two arrives while frame one is still fading. What would you expect to see? A superimposition — the horse in two positions at once, smeared. That is blur, not movement. Adding more frames faster makes the smear denser, not more mobile. A blend can no more produce a direction than a double exposure can.
Notice too what persistence does explain, because it is a real effect with a real job. If you interrupt a light often enough it stops looking like a flicker and starts looking steady — flicker fusion. At 24 frames per second the shutter blackouts flutter visibly, and projectors fixed that with a two- or three-bladed shutter flashing each frame two or three times, raising the interruption rate above the fusion threshold without shooting more film. The trick works entirely on the flicker and does nothing to the perceived movement, which is the cleanest demonstration that the two are separate problems.
So if blending is not the answer, what is? Consider the simplest case — no film at all. Flash a dot on the left, extinguish it, then flash a dot on the right. Two lights, each stationary, never on together. What you see depends almost entirely on the gap between them. Make the gap long, say a quarter of a second, and you see exactly what happened: one light, then another. Make it very short and you see them as simultaneous. In between lies a window — around a twentieth of a second for typical displays, though it shifts with separation and brightness, relationships Adolf Korte set out in 1915 — where you see something that did not occur at all. You see a dot travel.
Max Wertheimer studied this in 1912, and the paper is usually taken as the founding document of Gestalt psychology. Two distinct percepts came out of it, and they are routinely confused.
Beta movement is the one at the optimal interval: you see an object, and you see it move from the first position to the second. It has identity and a path — this is what a film delivers.
The phi phenomenon is different. At shorter intervals, observers report movement without an object moving — a sense of something passing between the positions, often described as shadowy or objectless, while both dots are also seen. Wertheimer named this pure phi precisely because the motion had come apart from any moving thing. Popular writing collapses the two and calls all apparent motion "the phi phenomenon", which reverses the more common case: cinema is beta. Robert Steinman and colleagues made the point at length in 2000, in a paper titled, unambiguously, "Phi is not beta".
That the two can be separated matters. If motion were simply an inference drawn from noticing an object first in one place and then another, phi could not exist, because in phi there is no object to track. Motion is registered in its own right.
And that is what the physiology shows. The visual system contains neurons tuned to direction — cells in primary visual cortex, and beyond them area MT — responding selectively to movement across a particular part of the field in a particular direction, and two flashes appropriately spaced drive those detectors much as real movement does. The system must also decide which blob in frame two corresponds to which in frame one, the correspondence problem, which it settles with strong assumptions about objects persisting and taking short paths. So apparent motion is not a failure of the eye. It is the ordinary machinery of motion perception, working as designed, on input that happens to be discontinuous.
The analogy
THE ANALOGY #Think of a smoke detector. It does not perceive fire; it responds to particles in a chamber, and it has been built so that the pattern fire produces sets it off. Burnt toast produces enough of the same pattern, and the alarm sounds — not because the detector malfunctioned, but because it is doing precisely what it was built to do with the input it received.
the alarm has one output and no interpretation, whereas the visual system builds a rich, structured percept — an object, a path, a direction — and can even be induced to report motion with no object attached, which no threshold device could do.
Clarifying the model
THE MODEL #Persistence of vision is not fictional. Temporal integration in the visual system is real, and it is why flicker fuses and why a sparkler drawn through the air leaves a trail. What is false is the inference from it to motion. Smearing successive frames together would, if anything, work against clean motion perception; the story survives mainly because it was written into film textbooks early and repeated. And beta and phi are not two names for one thing — the distinction is the evidence that motion perception does not reduce to object tracking.
The refinement is that "frame rate" is really two requirements wearing one number. One is that the interruptions be frequent enough to fuse, which is about flicker. The other is that successive images be close enough in time and space for the motion system to link them, which is about correspondence — and when a film pans too fast that second requirement fails while the first is still met, so the image judders even though nothing flickers.
A picture of it
THE PICTURE #How to readthe physical stimulus never changes — two stationary flashes, one after the other — so only the gap between them changes, and the diagram runs left to right as that gap grows. Each box is a percept the system occupies exclusively at that gap, and the inner line names what an observer actually reports. The return transitions are the same journey run backwards, since the ordering is driven by the interval alone. Note that pure phi sits between simultaneity and beta rather than being another word for it.
What became clearer
WHAT CLEARED #Film does not work by images lingering and blending. It works because the visual system has dedicated machinery for detecting movement, and that machinery responds to two appropriately spaced flashes the way it responds to a moving object — constructing a path that was never on the screen. Persistence of vision belongs to a different question, flicker, which cinema solved separately with a shutter.
Where to go next
ONWARD #- The correspondence problem: how the visual system decides which element of one frame is which element of the next, and what happens when it guesses wrong.
- Why higher frame rates change the perceived texture of a film, and why audiences often dislike the result.
Key terms
TERMS #| Term | What it means |
|---|---|
| Beta movement | apparent motion in which an object is seen travelling between two successively presented positions. |
| Phi phenomenon | Wertheimer's pure movement percept at short intervals, in which motion is experienced without any object being seen to move. |
| Flicker fusion | the point at which an interrupted light stops appearing to flicker and looks steady. |
| Korte's laws | the empirical relationships between interval, separation and intensity that determine which apparent-motion percept occurs. |
Every term the collection defines is gathered in the glossary.