What Kinalyse measures, how far to trust it, and where its limits are. If your question is not here, ask.
Jump height from flight time, contact time and Reactive Strength Index, countermovement depth and duration, barbell trajectory with vertical displacement and peak velocity, sit-to-stand repetitions and phase durations, sprint splits, and joint angles from pose estimation.
From an imported force plate: peak force, rate of force development, net impulse and impulse-momentum jump height.
For jumps and anything timed, 120 fps or better. At 30 fps a single frame is 33 ms, which on a 500 ms flight is a 3.3% error in time and about 6.7% in the height derived from it. Modern phones shoot 120 or 240 fps; it is worth turning on.
Film square to the plane of movement, keep the whole body in frame, and include something of known length for calibration.
For anything measured in metres, yes — the app needs a known distance in the same plane as the movement to convert pixels into real units. Timing-only metrics such as flight time and contact time do not need it.
Calibration error grows with camera distance and with the object being off the calibrated plane, and it is the largest error source in field use.
Yes, frame by frame, and corrections are recorded rather than hidden. A result that a human adjusted is marked as such in the app, the report and the CSV — that is the difference between a black box and an instrument.
Honestly: not yet measured against a laboratory. The app has not been validated against a force plate or a marker-based system, and it says so wherever a number is shown.
What it does show is repeatability — how consistent a measurement was across the trials in one session. Trend charts draw ±1 SD bars from that, and a change smaller than the bars cannot be told from trial variation.
A 2D joint angle is only valid while the segment moves in the image plane. An athlete rotated off-axis produces systematically wrong angles with every joint found and tracking confidence at maximum — nothing is failing, so nothing would otherwise report a failure.
Kinalyse estimates the rotation and tells you what it costs: about 2° on a 45° angle at 20° of rotation, nearly 8° at 40°. You can then decide whether to accept it or re-film.
No, and the app will not help you do it. Kinalyse measures and monitors; it does not diagnose or predict injury. Two-dimensional video from a single camera at unknown rotation cannot support that claim, and presenting it as though it could would be worse than useless.
They measure different things. Flight-time jump height assumes take-off and landing posture are identical — tuck the legs in flight and the flight time lengthens without the centre of mass rising any higher, and the formula reads that as a better jump.
A force plate’s impulse-momentum height measures the rise of the centre of mass and does not care what happens in the air. Compare like with like across sessions rather than mixing the two.
On your Mac, in a protected folder belonging to Kinalyse alone. Settings ▸ Data ▸ Show in Finder opens it. Nothing is uploaded anywhere.
No. Time Machine covers the folder as part of your home directory; if you do not use it, copy the folder periodically. Athlete history is the part of this app that cannot be regenerated.
Yes. Measurements export as CSV, reports as HTML or PDF, and lab data as C3D. Nothing is locked in a format only Kinalyse can read.
Not yet. Kinalyse is a Mac app. An iOS version is being considered but is not close, and it would need to be designed for a phone rather than shrunk onto one.
⌘Z steps back through the last twenty actions, and
it covers more than drawings — setting a calibration, correcting a
tracked point and deleting an annotation are all undoable. The Edit menu
names what it is about to undo.
The full documentation ships inside the app — press ⌘?
or open Help from the sidebar. It works without a connection. For anything else,
email info@kinalyse.com.