How-to · Formulas · Error sources
Measure your vertical jump with an iPhone
Jump height is how far your center of mass rises from take-off to the highest point. With an iPhone you measure it from flight time in a slow-motion video. Powerleap corrects for touching down with bent knees and shows the ± of every measurement – 1.2–1.9 cm at 240 fps in our check.
In short
- iPhone upright at hip height, level, from the side, about 2.5 m away.
- Whole body in the picture, feet and floor visible, room above your head.
- Swing your arms, land with straight legs.
- Height from flight time T: h = g · T² / 8 – but only if you land the way you took off.
Step by step
Set up the iPhone
Stand the iPhone upright at hip height – tripod, chair or leaned against a bag – and keep it level, not tilted up or down. The app says “Level the phone” from 8° of tilt and “Raise the phone – hip height” if it stands too low.
From the side, about 2.5 m away
Film from the side at about 2.5 meters. Your whole body including feet and floor must be in the picture, with room above your head for your arms. From about 4 m the take-off came up to 2 frames early in our check – whether because of the distance or the jumping technique, 3 vs. 3 jumps by one athlete cannot tell yet.
Light and frame
Use plenty of light: at 240 frames per second your feet blur otherwise, and the take-off depends on them. Be alone in the picture.
Stand still until the beep
Open “Jump”, stand in place and keep still until the beep and “Ready”.
Jump and land with straight legs
Swing your arms and jump as high as you can. Land on the balls of your feet with straight legs and stand still for a moment afterwards.
Check the result
The app announces the height and shows its ±. If it is unsure, it keeps the jump for review; then you set take-off and landing yourself, frame by frame.
The formulas
| Method | Formula | Valid when … |
|---|---|---|
| Flight time | h = g · T² / 8 | you land in the same posture you took off in. Example: 600 ms of flight time give 44.1 cm. |
| Rise time | h = g · t² / 2 | the highest point is clearly in the picture; t is the time from take-off to the peak. Example: 300 ms give 44.1 cm. |
| Flight time with landing correction | T = √(2h/g) + √(2(h+Δ)/g) | your center of mass is Δ lower at touchdown than at take-off; the app solves the formula for h. |
g = 9.81 m/s² (the app uses 9.80665). T = flight time from the first frame in the air to the first frame with ground contact. Calculator for both directions: vertical jump calculator.
Why flight time overstates a bent-knee landing
If you touch down with knees bent – arms down, flat-footed – your center of mass is lower at touchdown than at take-off. The ground is the same – this refers to your body. Your center of mass falls for longer and the flight time gets longer. Rule of thumb from the formula: the plain flight-time formula overstates the height by about half of how much lower your center of mass is at touchdown. A lab study with 27 people found the center of mass 5.3 cm lower at touchdown on average and the height 2.5 cm too high (Yamashita et al. 2020).
All 24 of our check jumps landed on the balls of the feet first, mostly with the arms lowered or the hips bent. With the same hand marks, the plain flight-time formula gave 0.4 to 7.2 cm more than the calculation with landing correction (table).
What the landing correction does
Powerleap measures Δ directly from your posture in the take-off and landing frames (center of mass from the joint positions) and solves the corrected formula for the height. The height then barely depends on the landing mark – ±3 frames change it by at most 0.44 cm – but more on the take-off mark: about 1.2 cm per frame. The correction is backed for Δ from 0 to 12 cm. Beyond that, or with knees below 150°, the app keeps the jump for review; from 15 cm it counts as a bent-knee landing and counts neither as a record nor for readiness.
Check on 24 real jumps
Ground jumps by three athletes, 240 frames per second, iPhone 17 Pro. Jumps 1–9 (28 Oct 2026): indoors, athlete 1, camera raised, on the floor and low. Jumps 10–18 (28 Oct 2026): outdoors, athlete 2 in loose trousers and athlete 3 in slim trousers, camera low at about 0.7 m and tilted slightly up. Jumps 19–24 (30 Oct 2026): gym, athlete 1, camera at hip height (about 1.1 m) and level, jumps 19–21 from about 4 m (4.0–4.2 m), 22–24 from about 2.5 m. Take-off and landing are marked by hand in the video (±1 frame). From jump 10 on, the hand-set take-off mark tends to be 1 frame late; a range of 2 frames applies there, and the calculation is given for both. Evaluated with version 1.0.1.
The reference is the app’s own calculation with the hand marks. So this checks how well the automatic detection finds take-off and landing – not the formula itself.
| Jump | Athlete | Camera | App height | With hand marks | Without landing correction (hand marks) | Take-off / landing | Status |
|---|---|---|---|---|---|---|---|
| 1 | 1 | raised | 45.9 ± 1.5 cm | 45.7 cm | 46.7 cm | 0 / −1 frames | safe |
| 2 | 1 | raised | 43.4 ± 1.5 cm | 43.4 cm | 46.0 cm | 0 / 0 frames | safe |
| 3 | 1 | raised | 42.7 ± 1.5 cm | 42.7 cm | 46.0 cm | 0 / 0 frames | safe |
| 4 | 1 | floor | 41.5 ± 1.8 cm | 41.5 cm | 47.3 cm | 0 / −1 frames | safe |
| 5 | 1 | low | 40.5 ± 1.8 cm | 40.5 cm | 45.4 cm | 0 / 0 frames | safe |
| 6 | 1 | floor | 39.8 ± 1.9 cm | 39.8 cm | 45.4 cm | 0 / 0 frames | safe |
| 7 | 1 | raised | 47.8 ± 1.5 cm | 47.9 cm | 48.6 cm | 0 / +1 frames | safe |
| 8 | 1 | raised | 46.9 ± 1.7 cm | 46.9 cm | 47.3 cm | 0 / +1 frames | check |
| 9 | 1 | raised | 44.8 ± 1.6 cm | 46.3 cm | 47.9 cm | +1 / +1 frames | safe |
| 10 | 2 | low | 26.0 ± 1.2 cm | 27.0–28.1 cm | 28.7–29.2 cm | +1 / 0 frames | check |
| 11 | 2 | low | 30.1 ± 1.7 cm | 28.1–29.1 cm | 33.3–33.8 cm | −1 / +2 frames | check |
| 12 | 2 | low | 27.4 ± 1.4 cm | 25.4–26.4 cm | 29.2–29.7 cm | −1 / 0 frames | safe |
| 13 | 3 | low | 41.7 ± 1.7 cm | 41.4–42.9 cm | 46.7–47.3 cm | 0 / −2 frames | safe |
| 14 | 3 | low | 43.1 ± 1.5 cm | 43.1–44.4 cm | 44.2–44.8 cm | 0 / 0 frames | safe |
| 15 | 2 | low | 26.0 ± 1.8 cm | 25.0–26.0 cm | 32.2–32.7 cm | 0 / −1 frames | check |
| 16 | 2 | low | 34.0 ± 1.8 cm | 31.6–32.7 cm | 36.5–37.1 cm | −1 / −1 frames | check |
| 17 | 3 | low | 41.4 ± 1.6 cm | 38.6–40.0 cm | 42.9–43.6 cm | −1 / +1 frames | safe |
| 18 | 3 | low | 39.2 ± 1.6 cm | 40.5–41.8 cm | 42.3–43.0 cm | +1 / −1 frames | check |
| 19 | 1 | hip height, 4.2 m | 46.6 ± 1.5 cm | 42.8–44.0 cm | 46.7–47.3 cm | −2 / +1 frames | check |
| 20 | 1 | hip height, 4.2 m | 47.3 ± 1.6 cm | 43.5–44.9 cm | 46.0–46.6 cm | −2 / +4 frames | check |
| 21 | 1 | hip height, 4.0 m | 45.6 ± 1.8 cm | 43.5–44.6 cm | 44.8–45.4 cm | −1 / +4 frames | check |
| 22 | 1 | hip height, 2.5 m | 45.6 ± 1.5 cm | 45.6–46.9 cm | 48.6–49.2 cm | 0 / −1 frames | safe |
| 23 | 1 | hip height, 2.5 m | 45.0 ± 1.6 cm | 45.0–46.5 cm | 48.6–49.2 cm | 0 / 0 frames | safe |
| 24 | 1 | hip height, 2.5 m | 47.3 ± 1.6 cm | 44.6–45.9 cm | 47.9–48.6 cm | −1 / +1 frames | safe |
- Take-off: within 1 frame of the hand mark in 22 of 24 jumps (from jump 10 on: of the range), 2 frames early in jumps 19 and 20 (4.2 m). Against the hand mark alone it is 16 of 24: in jumps 11, 12, 16, 17, 21 and 24 the automatic mark came 2 frames early, in 19 and 20 3 frames. For the 18 jumps of 28 Oct it was 16 of 18 against the range until the revision in version 1.0.1; for athlete 2 the mark came 4 frames early twice (5 against the hand mark).
- Landing: within 1 frame of the hand mark in 20 of 24; 2 frames off in jumps 11 and 13, 4 frames late in 20 and 21 (from about 4 m, heel still high at touchdown). For the 18 jumps of 28 Oct it was 6 of 18 before, up to 5 frames early. Thanks to the landing correction the landing mark barely affects the height.
- Height: within ±1.5 cm of the calculation with hand marks in 22 of 24 (from jump 10 on: of the range for both take-off frames); jumps 19 and 20 were 2.6 and 2.5 cm above. From 2.5 m (jumps 22–24) all three were within ±1.5 cm. Against the hand mark alone it is 16 of 24: the same eight jumps were 1.9 to 3.9 cm too high. For the 18 jumps of 28 Oct it was 15 of 18 against the range before, including jumps 11 and 16 at 4.5 and 5.2 cm too high.
- Noise test: with artificial noise added to the body points, 10 runs per jump, the take-off was within ±1 frame in 215 of 240 runs, the landing in 186 of 240 and the height within ±1.5 cm in 214 of 240. For the 18 jumps of 28 Oct and the three from 2.5 m the take-off was right in every run, for the three from about 4 m in 5 of 30.
- Status “check” (jumps 8, 10, 11, 15, 16, 18, 19, 20, 21): here the app found a reason for doubt – the rise-time check more than 2 cm below the height, foot and hip pointing to different take-off frames, take-off, landing and flight curve not quite fitting together, or a foot that was briefly hidden before take-off (jumps 19–21) – and asks you to review the jump instead of saving it silently. Jump 20 (2.5 cm too high) was saved silently until the latest revision of version 1.0.1. 15 were saved silently, among them jump 9 (1.5 cm below the manual evaluation), jump 12 (1.0–2.0 cm above), jump 17 (1.4–2.8 cm above) and jump 24 (1.4–2.7 cm above) – for 12, 17 and 24 depending on the take-off frame of the range: “safe” does not mean error-free.
- The column “without landing correction” is the plain flight-time formula h = g · T² / 8 with the hand marks.
Limits: three athletes, one iPhone model, recordings from two days; compared with a manual evaluation of the same video – not yet with a force plate. Recommended setup: hip height, about 2.5 m away. In our check (3 vs. 3 jumps by one athlete on the same day) the two outliers in take-off and height were both filmed from 4.2 m. Whether that is down to the distance or to the jumping technique has not been separated yet.
Error sources
| Error source | What happens | What helps |
|---|---|---|
| Camera low, on the floor or tilted | With a low camera the landing mark came too early until the revision in version 1.0.1 (simulation, camera at 0.75 m: 1–4 frames, now ±1). With the camera on the floor and a landing 10–20 cm closer to it, the height came out 1.5–4 cm too high in the simulation. | Hip height, level, land where you took off if you can. The app warns from 8° of tilt and when the camera is too low. |
| Knees bent on landing | Flight time too long, height too high – by about half of how much lower your center of mass is at touchdown. | Land with straight legs. The landing correction handles Δ up to 12 cm; from 15 cm the jump doesn’t count. |
| Landing on the balls of the feet | With the foot pointed, the ankle is not at standing height at first ground contact – about 5 cm above it on a flat forefoot landing, 12–13 cm with the heel held high. The landing mark can then come too early or too late (simulation: 2–8 frames early or 3–8 late; in our check 4 frames late twice). Thanks to the landing correction this barely affects the height, but it does affect ground contact time in the drop jump. | Nothing to do for the height: all 24 check jumps landed on the balls of the feet, 20 of them with the landing mark within ±1 frame. In the drop jump, check touchdown and take-off in the video. |
| Hip instead of foot for take-off | At lift-off the hip is already 8–19 cm above standing height (foot pointed, upper body forward). A take-off mark based on the hip therefore comes too early – each frame too early adds about 1.2 cm of height. | Keep your feet clearly visible (light, nothing in front). From version 1.0.1 the app sets take-off at the foot; the hip only counts when it is close by. |
| Rear foot hidden | The rear foot often leaves the ground up to 2 frames after the front one (on 30 Oct in 4 of 6 jumps, from 4.2 m as from 2.5 m). From about 4 m, body tracking briefly lost it during the dip: the front leg hides it, and tracking places it at the knee. That is how the app recognizes the case – in all three check jumps from about 4 m, in none of the other 21. Two of them (4.2 m) had the take-off 2 frames early and the height 2.5 and 2.6 cm too high. Whether that is down to the distance or to the jumping technique has not been separated with 3 vs. 3 jumps by one athlete; the hidden foot is the marker, it is not shown to be the cause. | About 2.5 m away, hip height – from 2.5 m all three check jumps were within ±1.5 cm. If the app sees the foot hidden, it asks you to review the jump (“One foot was briefly hidden before take-off …”); then set the first frame with both feet in the air yourself. |
| Loose trousers | Body tracking puts the ankle at the hem, and the point gets noisy. Until the revision in version 1.0.1 the app then followed the hip at take-off: twice 4 frames early, height 4.5 and 5.2 cm too high (jumps 11 and 16). | Tight or short trousers, ankles visible. If foot and hip disagree, the app asks you to review the jump. |
| Too little light | Feet blur, the take-off frame becomes uncertain. | Film in bright light, outdoors or with room lights on. |
| Low frame rate | The ± grows: the take-off mark alone quantizes the rise time to ±½ frame – about ±5 cm at 30 fps, about ±0.6 cm at 240 fps. The app shows the total ± of every measurement (1.2–1.9 cm at 240 fps in our check). | 240 fps slow motion. Below 100 fps the value is labeled “estimate”. |
| Feet out of frame, second person | Body tracking can’t find take-off and landing reliably. | Whole body and floor in the picture, jump alone. |
| Landing on a box | The flight-time formula would be wrong, because you land higher. | Nothing to do: the app detects the box and measures via rise time. |
| Phone moves | Movement in the picture looks like body movement. | Set it down firmly; the app uses the motion sensor to check that the phone stands still. |
Frequently asked questions
How accurately can an iPhone measure jump height?
At 240 frames per second one frame lasts 4.2 ms. A take-off mark that is one frame off shifts the height with landing correction by about 1.2 cm. In our check with 24 jumps by three athletes the height was within ±1.5 cm of the manual evaluation in 22 of 24, against both take-off frames where a 2-frame range applies. The other two, filmed from 4.2 m, were 2.5 and 2.6 cm too high; from 2.5 m all three were within ±1.5 cm. Against the hand mark alone it was 16 of 24, at most 3.9 cm off. Studies on video apps that use flight time without landing correction are listed in the comparison.
Which frame rate do I need?
The higher the better: the take-off mark alone quantizes the rise time to ±½ frame. For a jump of around 45 cm that is about ±5 cm at 30 frames per second and about ±0.6 cm at 240 frames per second. Below 100 frames per second Powerleap therefore labels the value “estimate”. The app shows the total ± of every measurement – 1.2–1.9 cm at 240 fps in our check.
Can I use a slow-motion video from the Photos app?
Yes: Profile → “Manual slow motion” → “Import from Photos”. When sharing, choose “Options” → “All Photos Data” so the original arrives at full frame rate.
What does the app measure when I jump onto a box?
With a box landing the flight-time formula is wrong, because you land higher than you took off. Powerleap detects the box itself and calculates the height from the rise time to the highest point.
Source
- Yamashita D, Murata M, Inaba Y (2020). Effect of Landing Posture on Jump Height Calculated from Flight Time. Applied Sciences 10(3):776. doi.org/10.3390/app10030776