Can you measure sprint speed from a phone video?
You can, to about the accuracy of the two marks you put on the ground, which is better than it sounds and worse than a pair of gates.
Coaches have been asking a version of this for years, usually while holding a phone with forty clips on it. The answer used to be no, not really, not without a scale. It is now yes, with one piece of setup and a short list of ways it can lie to you.
Where the number comes from
A video does not know how big anything in it is. An athlete filmed from close up and an athlete filmed from the stands are the same handful of pixels wide, so nothing in the clip says whether they covered eight metres or eighteen. That missing piece is the scale, and marking it is the whole job. Put two things a known distance apart on the line the athlete runs down, two cones ten metres apart is plenty, type that distance in, and tap each of them on the video. From there the app turns the hips' travel across the frame into metres, and the clip's own timing does the rest.
Form check then adds a speed line for sprint, block start and hurdles: the top speed and the average across the clip, in metres per second with miles per hour beside it.
What top speed means here
Top speed is the fastest 0.4 second window in the clip rather than the fastest single frame, because a tracker that wobbles for one frame would otherwise hand you a world record. The average runs from the first tracked frame to the last. The two together say more than either alone: a wide gap between them usually means the clip contains a chunk of acceleration, or the athlete drifted off the line you marked.
The four ways it goes wrong
- Marks that are not on the athlete's line. The scale only applies at the distance from the camera where you set it. Anything nearer the camera than the marks reads fast, anything farther reads slow, so the marks belong where the athlete actually runs.
- A camera looking down the track instead of across it. An athlete running towards the phone barely moves across the frame, which is the one thing this measurement needs. The Film Room warns when the view is head-on, and that warning applies to every angle in the report, not only the speed.
- A clip that is too short. Film the whole run through the marked zone with a little either side, and play it with the skeleton on so there are frames to measure.
- A treadmill. Nobody travels on a treadmill, so there is nothing to read. The app declines rather than printing a zero.
There is a ceiling as well. If the answer comes out above 13 metres per second, the app tells you the scale is wrong rather than printing it, because nobody has run that fast. The quickest speeds ever recorded in a 100 metres are a little over 12.
Against a pair of gates
Gates remain the better instrument. They measure one segment with two phones on a synced clock, and nothing about the number depends on marking a scale correctly. If you are comparing athletes with each other, or checking whether a training block worked, use the gates.
Video speed wins on two things. It works on footage you already have, including clips filmed long before anybody thought about measuring anything, and it reports a peak inside the run rather than an average across a fixed segment. A 60 metre run filmed with the marks in the middle gives you roughly how fast the athlete was when they were quickest, which one timed segment cannot.
What to do with the number
One reading on its own is not worth much. What makes it useful is the same discipline any test needs: the same marks, the same camera position, the same point in the session. Then a change in the number is about the athlete rather than about the setup.
It also helps to know roughly where the numbers sit, or the first one will either alarm you or flatter you. A well trained high school boy at full speed is usually somewhere around 9.5 to 10.5 metres per second at his quickest, a good girl nearer 8.5 to 9.5, and the fastest men in the world sit a little over 12. An athlete reading 11 on your runway has not out-sprinted the field. The marks were set wrong.
Treat this as an estimate with a known error rather than a measurement. Two cones, a tape pulled tight, a phone square to the track and an athlete on the marked line will get you close. Any of those done casually and the number is decoration.
Time your next meet with the phones you already have
Timing, the Film Room and the speed gates are free forever. Publishing results to your own page starts at $59 a month.
Related
Speed training, and how to tell whether it worked
There is no shortage of speed sessions to copy. The hard part is knowing whether the last eight weeks of them did anything.
Read →How to time a flying 30 without buying timing gates
The test is the standard one. The timing is where most squads either spend $400 or give up and use a stopwatch.
Read →Where does your sprinter actually hit top speed?
Most young athletes reach it somewhere between 20 and 30 metres, and they can only hold it for about eight steps, which changes where you should be testing.
Read →Try it: Running form analysis app: skeleton, form check, speed and cadence from a phone clip