Why some notes sound right together

Sit at a piano and press two white keys eight apart. Now press two keys next to each other. You don't need to be told which is which — one of them settles, the other snarls.

Musicians call that consonance and dissonance, and there's a long tradition of waving it away as pure convention. It isn't. Underneath sits some very concrete physics about how objects vibrate, plus a compromise Western music made a few hundred years ago that it has been quietly living with ever since. Your piano is out of tune on purpose. Every one of them is.

Here's the whole thing, from vibrating string to why a choir sounds sweeter than a keyboard.

Pitch is just a rate

Sound is a pressure wave moving through air, and pitch is how many times per second that wave repeats. Faster repetition, higher note. The unit is the hertz — cycles per second — and human hearing runs from somewhere around twenty hertz at the bottom to twenty thousand at the top when you're young.

The top end drops steadily with age, which is why one person in a room can hear a whining television and another genuinely can't. It isn't inattention. The hardware has changed.

What matters musically isn't the individual numbers, though. It's the relationships between them, because the ear is far more interested in ratios than in absolute values. Play a tune high, play it low, and it's obviously the same tune, even though not a single frequency in the two versions is shared. Nothing about the sound is the same. Everything about the proportions is.

Almost nothing vibrates at one frequency

If a note were a single frequency, every instrument playing the same note would sound identical. They plainly don't, and the reason is that a real vibrating object is doing several things at once.

Pluck a string. It swings along its whole length, and that gives you the fundamental — the pitch you'd name. But it's simultaneously vibrating in halves, in thirds, in quarters, in fifths, each of those subdivisions producing a fainter tone of its own. A string vibrating in halves produces exactly twice the fundamental frequency. In thirds, three times. In quarters, four times.

That stack is the harmonic series, and it's whole-number multiples all the way up. It isn't a musical convention, it's what a stretched string does whether anyone is listening or not. Air columns in tubes do much the same.

Which is where the ratios in music come from, rather than the other way round. Twice the frequency is the second harmonic, and we call that interval the octave. Three times the fundamental is an octave plus a fifth. Take that relationship on its own and you've got three to two, the perfect fifth. Four to three gives the fourth. Five to four gives the major third. Every interval Western music treats as basic is sitting there in the first handful of harmonics of a single plucked string.

Timbre is which harmonics you happen to get

What separates a violin from a flute from a human voice is largely which harmonics are present and how loud each of them is. That mixture is timbre, and it's why you can identify an instrument from a fraction of a second of sound.

The clarinet is the cleanest example, because its physics are visibly odd. It's a cylindrical tube closed at the reed end, and that shape strongly favours the odd-numbered harmonics — the first, third, fifth — while suppressing the even ones. You can hear the consequence without knowing any of the theory: a clarinet has a hollow, woody quality that a saxophone or an oboe doesn't.

There's a stranger consequence for the player. Most wind instruments overblow at the octave, jumping to the second harmonic. The clarinet skips it and jumps to the third — an octave and a fifth, a twelfth. That's why clarinet fingering is a nightmare compared with a flute's, and why the instrument needs so many keys to fill the gap.

Your brain will supply a note that isn't there

Here's a trick the ear plays that nobody warns you about.

Take a note with a full harmonic series and delete the fundamental — remove the lowest frequency entirely and leave everything above it. You still hear the original pitch. The brain works backwards from the spacing of the harmonics and reconstructs the missing note, and it does it so convincingly that you'd have to be told to notice.

This is why a phone speaker, which can't physically move enough air to produce a low bass note, still gives you a bass line you can follow. It isn't producing the note. It's producing the harmonics, and your auditory system fills in the rest without asking permission. Organ builders exploited the same effect for centuries by sounding two pipes together to imply a pipe far too enormous to build.

Why simple ratios settle and complicated ones don't

The octave is two to one, and it's the most consonant interval there is because the ratio is as simple as a ratio can get. Every second peak of the lower wave arrives at the same moment as a peak of the higher one. Nothing argues.

The fifth is three to two, and it's next in line for the same reason. As the numbers get bigger and less tidy — six to five, then worse — the intervals sound progressively less settled. There's a rough but real correlation between arithmetical simplicity and how stable a combination sounds.

The harmonic series gives a deeper reason for it. When two notes are an octave apart, the harmonics of the upper note are a subset of the lower one's — the two share most of their content, and the ear receives one coherent stack of frequencies rather than two competing ones. A fifth shares a good deal too. Awkward ratios share almost nothing, and their harmonics land near each other without matching.

Dissonance is roughness, and you can count it

Two frequencies close together but not identical do something specific. They drift in and out of phase, reinforcing and cancelling, and the loudness pulses at a rate equal to the difference between them.

Play four hundred and forty hertz against four hundred and forty-three and you'll hear three pulses a second. That's beating, and it's not a metaphor — piano tuners literally count those pulses with a stopwatch to set intervals, which is an oddly mechanical thing at the heart of a supposedly artistic trade.

Hermann von Helmholtz worked out in the eighteen-sixties that this is where dissonance largely comes from. When two notes sound together, it isn't only the fundamentals that interact — all their harmonics do. Intervals with awkward ratios have harmonics that land a few hertz apart, beating fast enough to be heard as a grinding roughness rather than a pulse. Dissonance, to a substantial degree, is the sound of harmonics colliding.

Later work refined this. The roughness depends on how close the clashing frequencies are relative to the ear's resolving power, which is finer at high frequencies than low. That's why a chord voiced tightly in the bass sounds muddy and the identical chord voiced high sounds fine, and why arrangers space low chords out without necessarily knowing the reason.

The sum that refuses to close

Now the awkward part, and it's the reason every tuning system since antiquity has been a compromise.

Stack twelve perfect fifths, each a clean three to two. Go up C, G, D, A and so on and you should come back round to C, seven octaves higher. You don't. You land sharp.

The arithmetic is brutally simple. Three halves multiplied by itself twelve times gives about a hundred and twenty-nine point seven. Seven octaves is two to the power of seven, which is a hundred and twenty-eight. Those aren't the same number, and they never will be, because powers of three cannot equal powers of two. The gap is the Pythagorean comma, and it's roughly a quarter of a semitone — not subtle, not a rounding error, and audible to anybody.

There's a second, similar problem hiding elsewhere. Four pure fifths stacked up don't quite produce a pure major third either; they overshoot by a slightly different amount called the syntonic comma. So the system doesn't just fail to close, it fails to close in more than one direction at once.

This has been known for well over two thousand years. Every tuning system in history is a different answer to it, and none of them is right, because there's no right answer available.

Cents, and the man who made this arguable

Comparing tuning systems used to be miserable, because you were comparing fractions with different denominators and no intuition for how they'd sound.

Alexander Ellis fixed that in the eighteen-eighties, while translating Helmholtz into English. He proposed dividing the octave into twelve hundred equal units and called them cents. A semitone in modern tuning is a hundred cents. Anything under about five cents is roughly the limit of what most people notice as a pitch change in isolation.

Suddenly you could say how wrong a thing was. The Pythagorean comma is about twenty-three cents. The syntonic comma about twenty-two. Those numbers are the reason the rest of this argument is possible to have at all.

What people tried before giving up

Tune everything from pure fifths and you get Pythagorean tuning. The fifths are beautiful, the major thirds are noticeably wide and harsh, and the entire accumulated error gets dumped into one interval somewhere in the circle, which howls. Tuners call it the wolf.

Renaissance musicians cared far more about thirds, so they went the other way. Quarter-comma meantone narrows every fifth by a small amount in order to make the major thirds pure at five to four. The result is genuinely gorgeous in the keys it was set up for. It's also unusable in the remote ones, and it still has a wolf.

Then came the well temperaments, which spread the error unevenly on purpose. Every key becomes playable, but no two keys sound quite alike — some are smooth, some are spiky. Composers wrote to that. When eighteenth-century writers describe a key as having a particular character, they may well be describing something that was physically there in the tuning rather than being fanciful.

Which is where a very common claim goes wrong. Bach's Das Wohltemperirte Clavier, from seventeen twenty-two, is not evidence for equal temperament. Well-tempered means what it says: a temperament in which all keys work. It doesn't mean all keys are identical, and that distinction is the whole point of the collection.

Equal temperament, and exactly what it costs

The solution Western music eventually settled on is to give up. Divide the octave into twelve mathematically identical steps, each one the twelfth root of two, and accept that almost nothing is in tune.

The maths had been worked out long before anyone adopted it. Zhu Zaiyu calculated the twelfth root of two in Ming-dynasty China in fifteen eighty-four, and Simon Stevin arrived at much the same thing in the Netherlands around the same time. It took European keyboards another couple of centuries to come round.

Now, the damage, in numbers.

Only the octave survives untouched. The fifths are each narrowed by about two cents, which is genuinely almost nothing — you'd struggle to pick it out even side by side. The major thirds are the casualties: about fourteen cents wide of a pure third, which is around a seventh of a semitone and quite obvious once you've been told to listen for it. Every major chord on a piano has a third in it that's beating away, and you've heard it your whole life without registering it as wrong.

What you buy for that is enormous. Every key becomes exactly as in-tune as every other, so music can wander anywhere it likes and modulate freely, and no instrument needs retuning between pieces. A great deal of later Western music simply couldn't exist otherwise. Whether that trade was worth it is a real argument, and people still have it.

The cost is easiest to hear in ensembles that don't have fixed pitches. A good choir, a string quartet, a barbershop group — none of them has frets or keys, and they drift instinctively toward the pure ratios, tuning chords more sweetly than a piano is capable of. If unaccompanied singing has ever struck you as richer than the same harmony on a keyboard, that's not romanticism. It's a measurable difference of about fourteen cents.

Real strings don't obey the theory either

All of the above assumes an ideal string, perfectly flexible, whose harmonics are exact whole-number multiples of its fundamental. Actual strings are made of steel and have stiffness, and stiffness pushes the upper harmonics slightly sharp of where the theory puts them.

That's inharmonicity, and on a piano it's significant, especially on the short thick strings at the bottom. It means that if you tune a piano's octaves to be mathematically exact, they'll sound flat, because the ear is judging them against the harmonics it can hear rather than against the fundamentals.

So tuners stretch. The top of a piano is tuned progressively sharp and the bottom progressively flat, by an amount that can reach tens of cents at the extremes. O. L. Railsback measured what tuners were actually doing in the nineteen-thirties and found they were consistently doing this without having been taught to, purely by ear.

A piano tuned by a machine to mathematically correct frequencies sounds wrong. A well-tuned piano is not in tune with itself in any strict sense, and it's shorter and stiffer strings that make the problem worse — which is why a small upright needs more stretch than a concert grand.

Even the starting note is a committee decision

All of this is about relationships, and none of it says where to start. That was up for grabs for most of musical history.

The A above middle C has been set anywhere from around four hundred hertz to close to four hundred and eighty, depending on the century, the country and whether the instrument belonged to a church or an opera house. Pitch drifted upwards over time, largely because brighter sounded better and orchestras were competing.

Four hundred and forty hertz became the standard by international agreement in nineteen fifty-five. It isn't a fact about the universe. It's an administrative decision that happens to be nearly universal now, and Baroque ensembles routinely ignore it and tune lower, which is why the same piece can sound noticeably darker on a period recording.

Consonance isn't fixed, and the tritone proves it

The physics explains why certain combinations are rough. It has nothing whatsoever to say about whether roughness is bad.

The interval spanning three whole tones was avoided in medieval counterpoint, and it eventually picked up the nickname diabolus in musica — though the phrase itself turns up in teaching manuals considerably later than the legend implies, so treat the devil story as folklore about theory rather than history. That same interval is now the opening move of a great deal of jazz and sits at the heart of ordinary blues harmony. Nobody flinches.

The acoustics didn't change. The convention did.

Across the last few centuries the direction of travel has been toward tolerating more and more dissonance, with each generation's difficult harmony turning into the next one's background music. If you find a piece harsh, that's a real perceptual response — it just happens to be shaped by what you've heard before, which is why exposure shifts the verdict more reliably than argument ever does.

Rhythm is doing at least as much work

Harmony gets most of the attention and rhythm does at least half the job.

Metre is the underlying grid of strong and weak beats. Rhythm is what's actually played against that grid. The interest lives in the gap between the two, which is why a drum part can be gripping without a single pitch in it.

Syncopation is the clearest case: accent a beat the grid says is weak, and the listener's expectation and the music pull in different directions. It's fundamental to jazz, funk, most Latin American traditions and a great deal of West African music, and it's why notating those styles produces a page that looks far more complicated than the music feels to play.

There's a reason rhythm is easier to feel than to explain. The perception of a steady pulse shows up very early in life, engages motor systems as well as auditory ones — people move to it involuntarily — and is one of the few musical responses found reliably across cultures. Someone who insists they're unmusical can almost always clap in time. That's a larger part of musicianship than they're giving themselves credit for.

How much of this is actually universal

It would be tidy to finish by saying consonance is physics and therefore the same everywhere. The honest answer is messier.

The octave does appear near-universally across musical cultures, and the fifth very widely, which is roughly what the harmonic-series argument predicts. Beyond that, practice diverges enormously. Plenty of traditions divide the octave differently, use intervals that Western ears hear as badly out of tune, or treat as unstable the very intervals Western theory calls restful.

Gamelan is the standard example, and it's a good one. Its tuning systems don't map onto twelve equal steps at all, the tuning varies from one ensemble to the next, and the metallophones themselves have inharmonic overtones — so the acoustic starting conditions are different, not just the taste.

The physics constrains without dictating. There are real acoustic reasons why some combinations sound stable, and a huge amount of room in what a culture chooses to build on top of them. Which is about the most interesting answer available: not arbitrary, not determined, but a very long negotiation between what vibrating objects do and what people decided to do about it.