How this instrument works
Two tones close in pitch refuse to blend into anything steady; they wobble. Add their waves point by point and wherever crests coincide sound is loud, wherever a crest meets a trough it nearly cancels, and that alignment drifts because one wave completes cycles marginally faster. A trigonometric identity pins it down: summing two equal-amplitude cosines yields a carrier at average pitch, (f₁ + f₂) ⁄ 2, wrapped inside an envelope oscillating at half of their difference. Loudness peaks twice per envelope cycle, once at each bulge, so what anyone actually counts is that full difference — three per second for 440 against 443.
Joseph Sauveur turned this wobble into a measuring instrument. Mute until roughly age seven and by repute unable to carry a tune, he nevertheless coined our word acoustics, and around 1701 he reported to Paris a way of fixing absolute pitch for what may have been history's first time. Take two organ pipes whose interval, and therefore whose ratio, is already known; count throbs per second between them; solve those two facts simultaneously for both pipes. No clock fast enough to time a single vibration was needed, because a slow phenomenon had been made to reveal fast ones. Radio mixers, laser vibrometers, and optical frequency combs still run on exactly that bargain.
Subtraction this plain still carries assumptions. Amplitudes want to be comparable, or nulls stay shallow and pulsing degrades into faint ripple riding a loud note. Pitches must sit close relative to their mean: past roughly 15 Hz apart, swelling smears into roughness, and past 20 or 30 Hz most listeners stop hearing one wavering note and hear two. Absolute value discards direction as well, so 440 against 443 and 443 against 440 both return 3 Hz and neither says which source runs sharp. Finally, this describes pure tones. Feed real instrument notes in and you get a whole family of throbs, one per overlapping pair of harmonics, which is why a badly tuned piano sounds busy rather than merely incorrect.
- Put your reference pitch into First frequency — tuning fork, phone app, or whichever string you trust. Hertz, kilohertz, and megahertz all sit on that unit menu.
- Put whatever is being checked into Second frequency. Order changes nothing here, since an absolute difference is taken.
- Read Beat frequency. Invert it for spacing between swells: 3 Hz means one every third of a second.
- To learn which source runs sharp, deliberately slacken one and listen. A slowing count means you moved toward agreement; a quickening count means away.
- Zero is your target. Once that count drops under about one per second, time ten swells against a stopwatch rather than trusting a guess.
Worked example — a string 3 Hz sharp of concert A
Your reference tone holds concert A at 440 Hz, and one guitar string sounds almost right without quite settling. A meter puts that string at 443 Hz. Enter 440 into First frequency and 443 into Second frequency; Beat frequency returns |440 − 443| = 3 Hz. Three swells every second, one roughly every 0.33 s — slow enough to count on your fingers, obvious enough to hear across a noisy room.
Those three pulses are the entire method. Nobody hears 3 Hz as a pitch; what reaches you is a single note breathing at 441.5 Hz, its midpoint. Ease that peg down and if swells stretch toward one per second you are closing in; carry on until the note stands still. Ears resolve throbs far more finely than they resolve pitch, which is the quiet trick here: a listener who could never name 440 apart from 441 in isolation will notice one lazy pulse per second at once, discriminating to roughly 0.2% for free. Nor is 443 necessarily an error — several European orchestras tune deliberately above 440.
Questions
Why is the throb rate the full difference rather than half of it?
Because loudness peaks twice per envelope cycle. Summing two cosines produces an amplitude envelope oscillating at |f₁ − f₂| ⁄ 2, and every cycle of that envelope carries one positive bulge and one negative bulge. Hearing responds to magnitude, not sign, so both bulges land as a swell. Algebra says half; your ear says twice that; both agree once you stop confusing an envelope with what it does to perceived volume.
Which of my two sources is running sharp?
This number cannot tell you — absolute value throws that away, which is why 440 against 443 and 443 against 440 both read 3 Hz. Settle it by experiment instead: shift one source in a direction you know and listen. Slacken a string, and a slowing count means that string was the sharp one, so keep going. A quickening count means it was already flat and wants tightening.
How far apart can two tones sit and still beat?
Audibly, around 15 Hz for most listeners, though that edge is soft and personal. Under about 6 to 10 Hz you count discrete swells comfortably. From there to roughly 15 Hz they merge into a rough, rattling texture. Push wider and hearing separates that pair into two distinct pitches; arithmetic here still returns a valid number, but nobody perceives it. Electronics face no such ceiling — mixers beat megahertz against megahertz routinely and keep whatever difference falls out.
What pitch do I actually perceive while beating happens?
Their average, (f₁ + f₂) ⁄ 2, which comes to 441.5 Hz in our worked example. You hear one note sitting at that midpoint whose volume rises and falls, not two notes trading places. This is precisely why beating makes such a sensitive tuning aid: hearing stops trying to compare two pitches and starts counting a slow rhythm, and people track rhythm to within a few percent untrained.
Does any of this apply beyond sound?
Superposition cares nothing for medium. Radio receivers mix an arriving carrier against a local oscillator and keep their difference, called an intermediate frequency, because filters behave far better at one fixed low frequency than at whatever a station broadcasts. Laser laboratories beat two optical sources onto a photodiode and read a radio-band difference, which is how frequency combs tie optical clocks to microwave standards. Flight crews once synchronised propellers purely by listening for throbbing to stop.
What does a beat frequency of zero really mean?
That both frequencies agree to within however long you are willing to wait. Exactly zero demands identical sources; in practice 0.1 Hz gives one swell every ten seconds, needing patience and a note that sustains. Tuners exploit this deliberately, since a slower count means finer resolution — so accuracy at the very end is limited by how long a plucked string keeps ringing, never by this arithmetic.