Tripod Height: Why Chest Height Distorts Every Angle You Measure
There is one setting on your tripod that matters more than the camera, the app, or the subscription tier you’re paying for, and it’s the little twist-lock on the centre column. Get the lens to hip height and your shaft plane numbers become comparable from one week to the next. Leave it at chest height, which is where almost everyone leaves it because that’s where the legs naturally extend to, and you are measuring a different swing every time you film.
That sounds dramatic. It isn’t. It’s just projective geometry doing what it always does, and golf swing analysis tools being far more sensitive to it than their marketing suggests.
What actually happens when the lens sits too high
A camera turns a three-dimensional swing into a flat grid of pixels. Every angle an AI tool reports is measured on that flat grid, not in the real world. When the lens is above the plane of the object you’re measuring, lines that are genuinely parallel in space converge in the image, and lines at a fixed real-world angle get squashed toward the horizontal.
Here’s the version with numbers. Down-the-line view, camera 8 feet from the ball, lens at 54 inches (roughly mid-chest for a 5’10” golfer). Your hand path at P6, the moment the club is parallel to the ground on the way down, sits around 34 inches off the turf. The lens is looking down at that point by about 12 degrees. That downward tilt compresses the vertical component of everything below the lens axis and leaves the horizontal component alone, which means any angle you measure gets pulled flatter.
Run the same swing with the lens at 36 inches, roughly belt buckle on that same golfer, and the viewing angle to the P6 hand position drops to under 2 degrees. The projection is very nearly honest.
The shaft plane difference between those two setups, same swing, same club, is typically 3 to 5 degrees. I have measured 6 on a driver. That is the entire difference between “you’re slightly under plane” and “you’re steep,” delivered by a tripod leg.
The session-to-session problem is worse than the absolute error
Here’s the thing that actually ruins practice. A consistent error you could live with. If your tripod always reported shaft plane 4 degrees flat, you’d learn to mentally add 4 and get on with it. The trouble is that chest height is not a repeatable measurement, because it depends on where the tripod legs happened to land.
Tripod legs on a mat sit differently to legs on turf. Legs on a slope extend unevenly. You nudge one to avoid a sprinkler head. Each of those moves the lens by two or three inches, and two or three inches of lens height at 8 feet is roughly one degree of viewing angle, which is roughly one degree of apparent plane change.
Four sessions across a month at a typical range bay, all filmed at “about chest height,” with the same golfer making genuinely the same swing:
Session Lens height Reported shaft plane (P6, 7i) Reported spine angle at address
1 53 in 58.1° 32.4°
2 56 in 56.4° 30.9°
3 51 in 59.0° 33.1°
4 57 in 56.0° 30.5°
Spread 3.0° 2.6°Now the same golfer, same four sessions, tripod set to 36 inches using a floor mark:
Session Lens height Reported shaft plane (P6, 7i) Reported spine angle at address
1 36 in 61.6° 35.8°
2 36 in 61.3° 35.6°
3 37 in 61.9° 36.1°
4 36 in 61.4° 35.7°
Spread 0.6° 0.5°The second table is a measurement system. The first is noise with a decimal point on it. And the second table’s numbers are also closer to the truth: spine angle at address for this golfer, measured properly, is about 36 degrees. Chest height was reporting 31.
Notice what the first table does to a practice plan. You work on a flatter takeaway for three weeks. Session 3 says you got steeper. You didn’t. Your tripod just sat lower that day because the bay mat was thicker.
Where hip height came from, and why not eye level
The common advice is “hand height” or “waist height,” and those are all pointing at the same thing: put the lens near the vertical midpoint of the swing’s motion. The club travels from the turf up to above your head, but the parts you’re actually measuring, hands, shaft at P6, lead arm, hip line, spine, all live in a band between roughly 24 and 48 inches off the ground. Hip height splits that band. Errors above the lens and below it partially cancel rather than compounding in one direction.
Eye level, which feels natural because it’s how you watch the swing in person, is the worst of both worlds. It’s high, it’s unrepeatable (you crouch, you stand, your friend filming is 6’3”), and it puts the entire measured band below the lens axis so every error runs the same way.
For face-on framing the same logic applies but the sensitivity is different. Face-on is mostly about lateral motion, hip slide, head position, and those are less distorted by lens height. You still want hip height, because spine tilt away from the target reads shallow from above, but if you only fix one camera, fix the down-the-line one. That’s where the plane numbers live, and plane numbers are what people change their swing over.
There’s a broader setup checklist that covers distance, frame rate, lighting and alignment, and the tripod height point sits inside it: filming a swing an AI can actually read walks through the rest. Height is the one I’d fix first because it’s the one that silently corrupts data rather than obviously failing.
What the actual tools do with a bad camera height
Worth knowing which products care and which don’t, because it changes what you should trust.
Sportsbox 3D Golf is the interesting case. It reconstructs a 3D skeleton from a single 2D video, which in principle means it could correct for camera height. In practice its calibration step asks you to frame the golfer a particular way, and its own guidance is specific about camera position for a reason. Feed it a chest-height down-the-line video and the 3D reconstruction still works, but the pelvis and thorax rotation values drift. Their 6-degree-of-freedom outputs are genuinely useful; they are also the numbers most sensitive to the model guessing depth wrong, and depth guessing gets worse as the viewing angle steepens.
HackMotion doesn’t use video at all, which is exactly why it’s a useful reference. It’s a wrist sensor, so its flexion and extension numbers are immune to your tripod entirely. If you have one, you have a ground truth you can check video-derived wrist angles against. When your video app says your lead wrist is 10 degrees extended at the top and HackMotion says 22, your camera setup is the first suspect.
Uneekor and Foresight units measure the club with cameras or lasers at fixed known geometry, which is the whole point of a launch monitor. Their club path and face-to-path numbers are not affected by your phone tripod. But this is where people get tangled: they compare a launch monitor’s club path (a real 3D measurement, in degrees, of where the head travelled) against a video app’s shaft plane (a 2D projection of where the shaft pointed) and wonder why the numbers disagree. Different quantities. Chest-height framing widens the apparent disagreement and sends people chasing a problem that doesn’t exist.
V1 Golf and Hudl Technique, the workhorse manual-drawing apps that a lot of UK coaches still use, are pure 2D. Every line you draw is a line on pixels. They are honest about this, and they’re perfectly good tools, but they have zero ability to compensate. Whatever your lens height does to the image is exactly what they measure.
Swing AI and the various phone apps that auto-overlay a plane line are the ones to be most careful with, because they present an automatically drawn line with an air of authority. The line is drawn on the image. The image is wrong. The line is confidently wrong.
Setting it, and making it repeatable
Measure your actual hip height once with a tape, at address, in golf shoes. For most UK adult male club golfers that lands between 34 and 40 inches. Women’s club golfers typically 31 to 37. Write the number on a bit of tape stuck to your tripod’s centre column.
Then mark the tripod. Most tripods have numbered markings on the leg sections; if yours doesn’t, a strip of electrical tape at the right extension takes ten seconds and survives being chucked in the boot. Manfrotto Pixi Evo, Joby GorillaPod 3K with a short stand, the Ultrapod II strapped to a range bay divider: none of these are expensive, all of them hold a repeatable height.
Set the lens level, not tilted. A tripod at 36 inches with the head tilted up 10 degrees to “get the whole swing in” has undone most of the benefit. If the club is cutting off at the top, move the camera back, don’t tilt it. From 8 feet with a phone in 0.5x or standard wide, you have plenty of frame.
One more repeatability trick that costs nothing: put the tripod feet on tape marks too. Height fixes the vertical error. Distance and lateral position fix the rest. Three bits of tape on the mat and your session-to-session spread drops to inside a degree, which is finally tight enough that a real swing change shows up above the noise.
What to practise once the numbers are trustworthy
The point of all this is that you now get to believe your own data. Some things worth doing with it.
Film the same club, same target line, every session. A 7-iron. Not driver, because driver plane varies with tee height and ball position more than your swing does, and not wedges, because the shorter shaft steepens everything and the differences get compressed.
Track one number for six weeks. Shaft plane at P6 is the best single candidate for most 6 to 24 handicaps, because it correlates with strike quality and it responds to the drills people actually do. Write it down. A trend across six sessions means something now; before, it didn’t.
When your reading changes by more than 1.5 degrees between sessions, check the tripod before you check your swing. That’s a diagnostic rule, not a superstition. With a fixed-height setup the measurement noise floor is around 0.6 degrees, so anything over roughly 1.5 is either real or a setup error, and setup errors are far more common and far cheaper to rule out.
If you’re working with a coach who films you on their own gear, ask what height they shoot from and match it at home. Two different heights means two incompatible datasets, and you’ll spend lessons arguing about numbers instead of hitting balls.
The tape on the tripod leg is worth more than the next tier of subscription. Most people have it backwards.