Radar vs camera launch monitors, and what each sees

What a Doppler unit is actually looking at

A Doppler unit sits behind you and reads its own signal coming back off whatever moves. What it obtains first-hand is speed along the line to the sensor, sampled over and over while the ball is in the air. TrackMan's account of its radar says the ball is tracked continuously from impact through the full flight: measured, not estimated.

That claim depends on where the unit stands. Outdoors, with clear air in front, the ball flies its whole arc through the beam and the distances come from a tracked path. Indoors, with a screen six metres away, the flight it follows is over in a few yards and the carry it prints has been computed rather than watched.

A camera gets one moment and no flight

Foresight describes its own method as exact photometric measurements taken with a clear, direct view of the impact window. Four cameras watch a defined patch of air, and out of it come ball speed, launch angle, azimuth, total spin and spin-tilt axis, plus club speed, impact point, attack angle, club path, impact loft and closure rate.

What no camera holds is the flight. TrackMan describes optical systems as a fraction of a second at impact and then algorithms predicting the rest, which is a radar company's sales argument and a fair account at once. The split is mostly marketing anyway: a Mevo+ pairs Doppler radar with synchronised high-speed imaging, and Rapsodo's MLM2PRO runs dual optical cameras alongside radar processing.

Where the sensor has to stand

5.5 m / 18 ftTrackMan 4 screen-to-unit distance indoors

TrackMan's own room guide puts figures on the difference. A TrackMan 4, sitting behind the ball, wants 5.5 metres between screen and unit, and reducing it, the guide warns, introduces inconsistencies on faster shots. The ceiling-mounted iO, which that guide says focuses on the strike and the earliest part of ball flight, asks 3 metres from screen to ball and nothing behind.

Smaller radar boxes repeat the demand. FlightScope wants a Mevo+ eight feet from the tee outdoors with unrestricted ball flight, indoors seven to nine feet with thirteen feet of flight in front. Rapsodo asks 6.5 to 8.5 feet behind the ball plus eight to the net: fourteen feet of room for a shot that never leaves the building.

Spin is the number that gives the game away

42 of 118Driver shots discarded because spin was estimated, not measured

Indoors, radar frequently cannot see enough of the ball to read spin, so the manufacturers change the ball. FlightScope requires metallic dots or a Titleist RCT ball before it calls indoor spin on the Mevo+ accurate. Rapsodo is plainer: balls carrying its RPT marking are required before the MLM2PRO shows spin at all.

The units know which they have done, and say so. Alex Bliss and Ben Langdown ran a Mevo+ beside a TrackMan 4 indoors: one golfer, 410 shots, foil marker in a dimple as FlightScope requires. Spin arrives labelled measured or estimated, and they threw out every estimated shot: 42 driver swings of 118, 31 of 118 with a 7-iron, 41 of 174 with a pitching wedge.

Even the label is no guarantee. Five of the eighty surviving 7-iron shots came through tagged measured and read between 11,000 and 13,500 rpm, roughly 6,000 above the TrackMan on the same strike, which the authors put down to the ball's seam modulating the return at twice the true rate.

What the published agreement figures show

513 rpmShot-to-shot scatter between two monitors, driver spin

Ball speed and club speed are where the technologies agree. Brennan and colleagues tracked 29 youth golfers, ten drives and ten 6-irons each, on a Mevo+ and a TrackMan 4 at once, and reported correlations of 0.92 or better for both.

Spin was the worst of it: a mean bias reaching 1,238 rpm and a 95% interval from minus 2,628 to 5,103, a band close to 8,000 rpm wide around a driver figure near 3,000. That study kept every shot, estimates included. Bliss and Langdown, keeping only measured ones, got the driver mean difference to 55 rpm, shots still scattering 513 rpm either side.

Small angles are the other casualty. Across the same driver shots the two boxes differed by 40% on club path, 44% on attack angle and 47% on face to path as mean absolute percentage error, and with a 7-iron they disagreed about clubhead speed by 4.44 mph. One golfer hit all of it.

Reading a spec sheet that will not separate them

Neither manufacturer separates the two kinds of number. Foresight lists ball launch data and club data as one flat set. FlightScope's Mevo+ page lists seventeen full-swing parameters, carry, apex height, roll and total distance among them, a few lines from the note asking for thirteen feet of indoor ball flight. Roll distance from thirteen feet of flight is a model's opinion in the same typeface as ball speed.

One question sorts them: what physically reached the sensor before this number existed? Ball speed, launch angle, spin and impact point sit close to the observation. Carry, total, apex and landing angle are model output. The TrackMan in that indoor study ran normalised, zero altitude and no wind, and its 249-yard driver carry describes a ball that met a screen six metres away. A phone sits outside all of it: Pebble Swing reads joint angles and the backswing-to-downswing tempo ratio from the frames, and no ball number.