Camera height and parallax in golf swing video

What a lens can and cannot see of a shaft

42° reads as 62°Halfway-down shaft, true angle against its down-the-line image

A camera flattens three dimensions into two by discarding the one that points at it. Halfway down, a driver shaft runs backwards, outwards and downwards at once. Take a shaft 1.20 m long sitting at 42 degrees to the ground, with 0.80 m of its length along the axis of a down-the-line lens. Two thirds of the club is invisible. The frame prints about 62 degrees, and the far end of the range still gives 63.

That 20-degree gap between shaft and photograph is not an error you can film your way out of. The shaft is not parallel to the sensor, and one lens cannot recover what points at it. A shaft angle read off video is therefore partly a report on the camera. Two golfers filmed from one spot compare honestly. The same golfer filmed from two spots does not, and the difference looks like a swing change.

Ten centimetres of height, twenty degrees of shaft

≈4.4° per 10 cmApparent tilt of a hip-high shaft, camera near chest height, 3 m back

Camera height moves things only up and down in the frame; sideways positions come from where the camera stands laterally. That stays harmless until two points sit at different distances from the lens. Raise the camera and the near point drops down the frame faster than the far one, so the line joining them tilts.

Take the driver shaft where it lies parallel to the ground in the downswing, roughly hip high, clubhead 1.75 m from a camera three metres behind the ball and grip 2.85 m away. From a lens at 1.00 m it prints 2.6 degrees off horizontal. From 1.40 m it prints 22, about 4.4 degrees per 10 cm of camera height. The same 40 cm moves the shaft at address a fifth of a degree, because both ends of the club sit at nearly the same distance from the lens.

There is no single plane to photograph edge-on

48.3° and 57.9°Clubhead plane and lead-hand plane slopes, 15 skilled golfers, driver

Young-Hoo Kwon's laboratory at Texas Woman's University measured the planes a swing actually forms. Across 15 skilled golfers hitting drivers, the clubhead's functional swing plane sloped at 48.3 degrees, standard deviation 3.3. The lead hand travelled on a plane of 57.9 degrees and the trail elbow on one of 67.4.

Three planes, one swing, and a single down-the-line frame contains all of them. Nesbit's three-dimensional study put the gap between the plane traced by the clubhead and the plane traced by the hands at 9 to 12 degrees. Kwon and Casebolt's review is blunter still: a single planar swing plane does not exist.

Where the camera sits decides which plane a photograph resembles, because a lens renders a plane as a line only when it lies in that plane. Kwon's group measured the clubhead plane's direction at 3.8 degrees closed to the target line, standard deviation 3.7, so a camera squarely on that line already sits outside the plane, differently for every golfer.

Distance does the work the lens gets credit for

Perspective is set by where you stand, not by the glass. Focal length governs how much of the scene fits in the frame, and changes the geometry only because a wider lens tempts you closer. Kinovea's measurement guidance states the rule directly: place the camera as far as possible from the scene and zoom in.

The arithmetic agrees. That same hip-high shaft, filmed from a chest-high lens, prints 22 degrees at three metres and 8.7 at eight. A little more than doubling the distance halved the parallax, with nothing changed about the camera or the golfer.

Apple lists the iPhone 17 Pro's ultra wide at 13 mm with a 120-degree field of view, and its 2x telephoto at 48 mm. The ultra wide is what fits a full swing in from three metres in a narrow bay, which is exactly the setup that stretches depth. Walking back on the longer lens gives a flatter frame.

The tolerance, not the position

The question worth answering is not where to put the camera but how precisely to put it back. From the figures above, holding that shaft angle inside one degree needs the height repeated to about 2 cm and the distance to about 15 cm. Neither survives being judged by eye a month later.

Mark the spot rather than remember it. A taped cross for the tripod feet, the height noted on the leg, the lens locked before recording instead of pinched into place mid-zoom, and a fixed object in the background that shows up in both clips. Puig-Diví and colleagues, validating angle measurement from video, filmed from a lens 0.68 m high and 5 m away; the smallest error always came from the perpendicular view.

One caveat belongs on the page. No published study has measured how far a phone's camera height moves a golfer's shaft angle; the figures here are geometry worked for one stated setup, and your dimensions will shift them. What survives is the ranking: height first, distance second, lens last. Pebble Swing reads joint angles out of the frames, so the camera sits inside every angle it reports.