How to film yourself running for a reliable video analysis
A video analysis is only as good as the video. A phone, a treadmill and a few seconds are enough, as long as you follow a few simple rules. I'll also tell you what this method can show, and its limits.
Treadmill or outdoors?
On a treadmill you stay in the same place in the frame, which simplifies everything. The real question is whether your treadmill stride looks like your outdoor stride. A meta-analysis of 33 studies (494 runners) concludes that most spatiotemporal and kinematic measures are comparable. Differences concentrate at foot impact, with a larger foot-to-ground angle and more knee flexion on a treadmill.[1] Keep that in mind if your foot strike is what interests you.
Side-on, not from behind
I recommend filming side-on, because it's the view that teaches the most about technique: foot placement relative to the pelvis, trunk, bounce, cadence. The view from behind shows the foot and ankle, useful for pronation, but I think it isn't enough, as I explain in the posture analysis guide.
The settings of a good video
These are practical Twinrun recommendations, not thresholds from studies.
- Put the phone beside the treadmill, perpendicular to your run, not at an angle.
- Place it at hip height: too high or too low and angles get distorted.
- Stand back two or three metres so your whole body is in the frame, head to feet, with some margin.
- Hold it still, on a support or with both hands, without following the runner.
- Avoid a window or strong lamp right behind you.
- Wear shorts and a fitted top: knees, hips and ankles must be visible.
Run at your usual easy-run pace, after a few minutes of warm-up: the first strides on a treadmill don't always look like your real running. Then film ten to fifteen seconds.
The mistakes I see most
Feet cut off at the bottom of the frame, which stops you seeing foot strike. A three-quarter view, which distorts knee and trunk angles. Holding the treadmill rails, which changes your whole posture. Baggy clothes that hide the joints. Or looking at the phone while running, which turns the head and the trunk with it.
What 2D video can measure
In a study where 20 runners were filmed in 2D and 3D at once on a treadmill, agreement was good for foot angle at contact and knee flexion range, weaker for peak knee flexion, with differences that varied from runner to runner.[2] A side-on video is therefore a good way to spot differences and trends, not a laboratory instrument. That's exactly what I ask of it: read accuracy or trends.
What next?
Open Twinrun on your phone and pick your video. For the free comparison it's sent nowhere. To learn what the differences mean, read how to improve your running form, and for the different ways to get analyzed, running gait analysis.