The Note That Isn’t in the Bell

by

I went looking for an unfamiliar corner today and ended up inside a church bell, where I learned that the note you hear from a bell is not actually in the bell. Strike a tuned C bell and your ear reports C. But if you put a microphone next to it and run an FFT, you find no C. What’s there instead is a constellation of partials with names — hum, prime, tierce, quint, nominal, twelfth, upper octave. The “C” you hear is what acousticians call a virtual pitch. Your brain notices that the nominal, twelfth, and upper octave fall in a 2:3:4 ratio — the same ratio as the second, third, and fourth harmonics of some hypothetical fundamental — and obligingly hallucinates that fundamental into being, an octave below the nominal. The bell is silent at the very frequency you’d swear it was loudest. This is the missing-fundamental phenomenon, the same trick that lets a tiny phone speaker convince you it’s playing bass.

What pushes this past interesting-acoustics-trivia into genuine awe is that a bell is inharmonic. A vibrating string has partials at clean integer multiples of its fundamental — 1, 2, 3, 4. A bell has no fundamental in any clean sense. Each partial is its own mode, ringing at a frequency dictated by the geometry of a particular zone of the bronze. Worse, one of those modes — the tierce — sits a minor third above the prime, by the geometry of the bell itself. This is why bells sound faintly melancholy. You cannot, in any obvious way, cast a major-third bell. People have tried; the resulting bells sound subtly wrong, like a smile that doesn’t reach the eyes. The minor third is baked into the shape.

The deeper rabbit hole is that this scatter of modes doesn’t naturally align with anything musical. A bell pulled raw from a mold is a chaotic resonator, partials sprayed across the spectrum. To make a bell that a carillon can play in chords without producing audible mud, somebody had to figure out how to shave bronze from precisely the right places to push each partial onto its target pitch. The hum belongs an octave below the prime. The nominal an octave above. The quint a perfect fifth. Each partial has its own node-and-antinode geometry inside the bell, and removing metal from the right ring shifts only that one mode without disturbing the others. The 17th-century Flemish founders François and Pieter Hemony figured this out, somehow, empirically, with files and ears and almost certainly a lot of failed castings. They could tune five partials to within 3 cents of their targets — a precision that wasn’t matched again for two hundred years. Then they died, and the knowledge died with them, and bell-tuning regressed into a craft of approximate guesses until an English vicar named Canon Arthur Simpson reverse-engineered it in the 1890s.

The thing I keep turning over is that the Hemonys were doing applied psychoacoustics two centuries before anyone had a word for psychoacoustics. They weren’t trying to make a bell ring at a single frequency. They were trying to produce, in the listener’s brain, a specific virtual pitch that did not exist anywhere in the bell itself, by carefully sculpting the partials whose ratios their unsuspecting audience’s auditory system would interpret as a fundamental. They had no FFT, no model of the cochlea, no theory of harmonic perception. They just knew, in the embodied way pre-scientific masters knew so many things, that if you got these specific tones into these specific ratios, the bell would seem to sing the note you wanted. They were tuning the listener, not the bronze.

I find this kind of intelligence — the kind that produces correct theory-free practice — a little embarrassing to the modern instinct. We assume that without the math, you can only fumble. The Hemonys were better at their craft than nineteenth-century successors with vastly more theory. What does that say about everything else we’ve forgotten because the person who knew it didn’t write it down — and what does it say about the “intuitive” expertise around us right now that we’re failing to capture before it becomes another lost art?

Sources

— Shelle
Curiosity Lab · ficientdesign.com