Academy · How it works

The photo finish is not a photograph.

The image that decides races was never “taken” — it was grown, one hair-thin slice at a time. Once you see how, you’ll never look at one the same way again.

The photo finish you’ve seen on television — horses straining at the wire, runners mid-lean — looks like a photograph. It isn’t one.

There was no moment when that scene existed in front of a camera. No shutter clicked while all those competitors stood in those positions. A photo finish image is built in a way so different from ordinary photography that it’s fair to say the camera at the finish line doesn’t really take pictures at all.

Here’s what it actually does — and why the strange truth is exactly what makes the result trustworthy.

A camera that sees almost nothing

A Full HD image is 1920 × 1080 pixels — over two million picture points capturing a whole scene at once. Now imagine throwing away almost all of it: keep the 1080-pixel height, but only two columns of width.

That sliver is what a photo finish camera sees. With pixels 3.45 micrometres across, a two-pixel strip is 6.9 micrometres wide on the sensor — about one-tenth the width of a human hair. The lens focuses exactly one thing onto that hair-thin strip: the finish line. Not the track before it, not the run-out after it. The line.

One such exposure is nearly useless — a few coloured specks, maybe the edge of a shoe. So how can a camera that sees almost nothing decide a race?

Try it with your hands

Hold both hands up in front of your eyes, palms toward you, leaving a paper-thin vertical gap between them. No peeking around the edges — you may only see through the slit.

Now have someone move a pen across the gap. You see its tip … then a bit of the middle … then the tail as it leaves. You never see the pen. You see a sequence of slivers, each one showing whatever was in the gap at that instant.

That is precisely the camera’s view of the finish line. The difference: the camera remembers every sliver — capturing thousands per second — and its software places them side by side, each new slice next to the previous one.

Watch what happens as a horse crosses. First slice: the tip of a nose. Next: a bit more nose. Then head, neck, saddle, rider, tail — slice by slice, a complete horse assembles in the image. The camera never saw that horse. It saw a thousand slivers of horse, each at a slightly different moment, and the picture you’re looking at is all of those moments standing shoulder to shoulder.

The part most viewers never learn

Now the consequence that changes everything about how you read these images.

In a normal photograph, left-to-right is space: a gap between two runners shows the distance between them. In a photo finish image, every slice was recorded at the same place — the finish line. What changes from slice to slice isn’t position. It’s time.

Horizontal distance in a photo finish image is time. A gap between two competitors doesn’t show how far apart they were on the track; it shows how much time passed between their arrivals at the line. The image is a timeline you can look at.

And this is why the image can decide a race: every slice carries its own precise timestamp. When the official moves the vertical cursor onto a horse’s nose or a runner’s torso, the software isn’t estimating anything — it’s reading out the timestamp that already belongs to that slice. The picture isn’t just showing the result. The picture is the measurement.

(That movable line on the screen, by the way, is not the finish line. It can’t be — the entire image was recorded at the finish line. The cursor just selects which recorded moment to read.)

Why everyone looks a bit strange

Photo finish images have a famous side effect: stretched horses, compressed sprinters, bicycle wheels like odd ovals.

Nothing is broken. A slow-moving competitor spends longer crossing the line, so more slices contain them — they come out looking stretched. A fast competitor crosses in fewer slices and looks compressed. The image was never trying to be a portrait; it’s a record of how long each part of each competitor occupied the finish line. The distortion isn’t an error in the evidence. In a real sense, it is the evidence.

The one thing the camera must be told

For all its precision, the camera has a humble blind spot: it doesn’t know where the official finish line is. It measures, with perfect accuracy, whatever hair-thin strip of the world it was aimed at.

Aim it slightly before or after the true line, and it will time the wrong place — flawlessly. That’s why aligning the camera to the finish line isn’t part of getting a nice picture. It’s part of the measurement itself, and it’s done with great care before racing starts.

Final challenge: do you really understand the photo finish?

1. Why can the operator move the vertical line from one competitor to another?

It’s a timing cursor. The physical finish line can’t move, and the whole image was recorded there anyway. The cursor picks one of the thousands of recorded slices — placed on the body part that counts under the sport’s rules — and the software reads that slice’s timestamp. Moving the cursor doesn’t change the result; it chooses which moment to measure.

2. Where is the real finish line in the image?

Everywhere — the whole image is the finish line. This is the answer that rewires how you read the image. The camera never looked anywhere else; the picture gets its width from time passing, not from a wider view of the track. Every column of pixels is the same place at a different moment.

3. Two competitors appear far apart in the image. Were they far apart on the track?

Not necessarily — the gap is time. Both were photographed at the same place. The space between them shows how much recording time passed between their arrivals. Reading a photo finish like a normal photograph is the single most common misunderstanding about these images.

4. Why can one competitor look stretched and another compressed?

Different crossing times, different widths. A slower crossing fills more slices; a faster one fills fewer. The image records duration-at-the-line, not body shape — which is why odd proportions are a signature of honest photo finish evidence, not a fault in it.

5. Why does a time appear when the operator clicks in the image?

The time was recorded with the slice. Each slice is stamped as it’s captured, so every column of the image already knows when it happened. Clicking doesn’t calculate anything; it looks up a timestamp that has existed since the moment of capture. The image is a readable timeline.

6. What changes when the camera records more slices per second?

More slices, finer timeline. A competitor taking the same time to cross is built from more pieces, and less time passes between neighbouring slices. How many slices per second a race actually needs is its own interesting story — see the line rate article.

7. The camera was accidentally aimed slightly past the official finish line. Can moving the cursor fix the results?

No — no software can move the recorded line. The camera timed the place it was pointed at — perfectly, and irreversibly. The cursor selects when, never where. That’s why camera alignment is treated as part of the measurement itself, checked with care before racing begins.

The explanation worth remembering

A normal photograph shows many places at one moment. A photo finish shows one place across many moments.

So when someone asks why the operator is “allowed to move the line,” you can answer in two sentences: the finish line doesn’t move — the whole image was recorded there. The line on the screen is a timing cursor, choosing which recorded moment to read.

Next in the Academy: RAW, explained without photography school →