The Resonance Labfor Vibrant Souls
"Every voice has a diamond in the rough. The work here is to uncover it."
Welcome, Vibrant Soul Vocalists, to the Resonance Laboratory — the place where we study the ingredients of resonance and learn the alchemy of mixing them, guiding you to uncover the diamond in your voice, and set it in gold.
Let's begin. Stations 1–6 cover the science — the ingredients. Stations 7–9 are where you mix those ingredients and the artistry comes alive. Station 10 is where additional magic arrives, the listener's ear. And the last station hands you the mission board — where everything you've learned goes to work in the real world. Go in order the first time; wander freely ever after. (The Chords in Your Voice is the printable companion, carrying the whole journey.)
Wait, what? Did the piano just sing with me on its own?
Here's a hands-on experiment to show that your voice carries more than one note: sing next to a piano. Listen closely, and you will hear more notes coming from the piano than just the one you thought you were singing. To create this, push your foot down on the pedal farthest to the right — there are three, and the right-hand one is the sustain pedal. Sing into the cavity of the piano, and many of the piano's strings will vibrate with your voice, creating sound. This is sympathetic vibration.
Below is a simulation of this experiment, without the real piano. The sounds are computer tones playing what you would expect to hear — and know that an actual piano's sympathetic vibration is a lot more generous. The strings that vibrate with your voice are not random. Only notes that are relatives — like family members — to the pitch you sang will vibrate. Are you ready to start experimenting, and see if you can begin to notice who will be a family member, and why? Let's play.
The strings that answer will be listed here, with how in-tune each one was.
Try your own vocal alchemy — the control test: switch to "Only keys I hold", press Hold wrong keys, then sing. Almost nothing rings. That silence is the gift: it's what makes the first result mean something.
I can take a voice apart, note by note?
You can — and you're about to. Every slider is one of the hidden notes in a middle C — a note most people can sing. The bottom note never changes. Only how much of each upper note is in the mix changes — and that mix is the biggest part of what "tone of voice" is.
This is what the sound wave looks like as it leaves your vocal cords.
Why does "ee" sound higher than "oo" on the very same note?
Because each vowel shape magnifies its own set of your hidden notes — no matter what pitch you sing. Your voice is made in two parts. The cords make a buzz full of hidden notes — that's the stack. Then the buzz has to travel up a tube: your throat and your mouth. Here's the rule: as long as its walls stay quiet, the tube can't add any new notes. It can only turn some of them up and some of them down, like volume sliders. The tube has two favorite spots (scientists call them resonances), and notes that land near a favorite spot get loud. Change your mouth shape and the favorite spots move. That's all a vowel is.
Words like eye, ay, oy, ow, oh are diphthongs — two vowels with a glide between them. In real speech the whole journey takes a tenth or two of a second; here it's slowed to almost three, so you can hear the tube traveling. The circle tours the three corners of vowel space without ever stopping — there are no walls between vowels, only places the tube passes through.
Top: the smooth buzz the cords send into the tube. Bottom: the same wave coming out of your mouth — the tube turned some hidden notes up, some down, and nudged some in time, so the wave gets bent into a whole new shape. Same note, same cords — only the tube changed. And every vowel bends the wave its own way.
Try your own vocal alchemy — proof the tube has pitches of its own: whisper "ee — ah — oo". When you whisper, your cords don't buzz at all, so there's no stack. But you still hear the vowels change! So where is the sound coming from? Rushing air makes a soft hiss in your throat — a noisy source instead of a buzzing one — and the tube filters the hiss exactly the way it filters the buzz. Even a whisper is still source plus filter; what you're hearing is the tube's favorite spots, traced in hiss.
That's what my voice looks like?
It is — every line of it is you. Sing a steady "ah" and hold it. That first spike on the left? The note you think you're singing — though it isn't always the tallest; a bright voice can make ×2 or ×3 the loudest of all. Every spike after it is a note you didn't know you were singing, spaced at exact whole-number multiples. The moving picture below is called a spectrum.
The experiment: keep the same pitch and loudness, and slide from a soft breathy "ah" to a bright buzzy one. The first spike stays put; the count grows. That's the second test from the project, live.
What is that wobbling sound?
That wobble is a fight between two notes that are close but not the same. Your voice's seventh harmonic lands 31 cents below the piano's nearest string. And this isn't special to one note — every seventh harmonic of every piano-tuned note lands about 31 cents flat, because the mismatch is baked into how pianos divide the octave. Play the two together and listen to them fight: the loudness pulses as the waves drift in and out of step. Then drag the slider and tune the fight away.
The piano's string isn't wrong and neither is your voice. Pianos are tuned so every key works in every key signature, which nudges nearly every note off its natural spot. Number 7 is where the nudge gets big enough to hear.
The same note just changed color — how?
Here's the secret at the top of all of this. A trained singer holding one note is still making choices every instant: how firmly the cords meet, how long and wide the tube is, whether to gather the high harmonics into a ring — a bell-like shine floating on top of the note. None of it changes the pitch — all of it changes the color. Singers call the classic blend chiaroscuro — Italian for "bright-dark" — because the goal is both at once: a warm, deep tube and a gleam of ring on top. And here is the alchemy, spinning straw into gold: begin with a low-grade buzz, finish with a diamond ring.
Part 1 · The palette — three knobs, one note
Part 2 · The payoff — why ring carries
Why ring means being heard: the listener's ear is a tube too, and it amplifies exactly this band — roughly +20 dB near 3,000 Hz, before the eardrum even moves (station 10 shows why). Meanwhile an orchestra's energy piles up near 500 Hz and leaves this zone quiet. So the boosts chain together:
Boosts in a chain simply add their decibels — slide the ring dial and watch. That's how one singer sails over eighty instruments without singing louder: in measured opera voices the ring band grows about 16 dB when the whole voice grows only 10 (Sundberg 2001). (The ear's +20 dB is measured; the dial now spans the evidence — modeling puts the cluster's gain at 7–19 dB, and recorded opera voices run about 15–19 dB above untrained voices near 3 kHz.)
Why even +15 is more impressive than it sounds: harmonics arrive in the ring band already roughly 25–30 dB quieter than the bottom note — that's the natural downhill slope from station 3. So think of ring as rescue, not gilding: it lifts faint harmonics into competition, right where the listener's ear is waiting to double the favor.
What this chain counts — and what it doesn't: the two links are the trained gathering (your dial) and the listener's ear. They sit on top of boosts every voice gets free: the tube's upper favorite spots (3, 4 and 5) already live around the band, and leaving the lips lifts high harmonics by +6 dB per octave. And real throats fine-tune the gleam with sub-chambers this app doesn't model yet — the piriform sinuses, the epiglottic curl, the ventricle. Each is its own knob; they're getting their own app.
What the dial is shorthand for: speech keeps favorite spots 3, 4 and 5 spread out (gray). Classical training draws the same three together (green) — no new energy, no extra breath: one tall peak where their skirts overlap. The gathering is done with tube shape — a small larynx-tube outlet opening into a wide pharynx. (Heights are sketched to show the idea, not to scale; measured gains land at 7–19 dB.)
Try your own vocal alchemy — the studio experiment: turn the ring all the way up while the cords stay hazy, and… almost nothing. The meter barely moves. The tube is only an equalizer — it can't boost harmonics the cords never supplied — and that's exactly why a teacher asks for a clear, firm (not squeezed) sound before ring work ever starts.
Why did my sound collapse when the note climbed?
Because the note moved and your favorite spots didn't. So far you've moved the tube under a stack that stayed put. Now the note moves. As pitch climbs, the whole comb slides to the right, and if your favorite spots stay parked where regular speech left them, the harmonics slide right past them and the sound collapses — the harmonic windows were left empty. A harmonic window is a favorite spot seen from the comb's side: a region of brightness, and harmonics light up when they come into the window's view. And windows don't stay in one spot — they slide with your tube's shape, just as the comb slides with pitch. The fix is the deepest trick in singing: move the window with the note — sliding it upward as the pitch climbs, the way a surfer stays with a wave. Singers call this tracking the note. Slide the pitch up on each strategy and watch the meter.
How do you actually move a window? With the same handles that make vowels. Spot 1 lives in your jaw: drop it and let the mouth open tall, and spot 1 climbs — letting the tongue's body lie low helps, and some singers add spread lips (measured sopranos split between the two routes). Spot 2 lives in your tongue: arch it forward with its sides resting on your upper side teeth — the "ee" position — and spot 2 rises; draw it back toward "oo" and spot 2 falls. (An honesty note in passing: the jaw and the tongue are each spot's lead player, not its only one. These knobs are really directions — up-and-down, forward-and-back, side-to-side — and the lips, the soft palate, even the tongue's flex all share the controls. When you're ready for the full cast, the mission board's resonance tuners hand it to you.) Try your own vocal alchemy — the separation drill: keep the jaw loose and still, and glide the tongue between "ee" and "oo" — you just moved spot 2 all by itself. Then park the tongue and swing only the jaw — that's spot 1 alone. Two windows, two independent handles. And to hear a window move with no note in the way, whisper while you do it — the hiss itself sweeps as the spot slides. One more handle moves the whole rack at once: round and push out your lips, or let the larynx settle low, and every spot slides down together — because the tube just got longer. That's the length slider from station 7, played on your own face.
Why sopranos have no choice: by around C5–E5 the note itself is higher than where "ah" keeps favorite spot 1. Keep the speech shape and the strongest harmonics have nothing under them. Real sopranos open the jaw and let spot 1 track the note itself — measured all the way up to about 1,000 Hz (Joliveau, Smith & Wolfe 2004, Nature), worth up to 20–30 dB. Experts start tuning around B4–D5 — and here's the encouraging part: the untrained sopranos in the study did it too, from about E5, without ever being taught. Your body may already know this trick; training just makes it arrive earlier and hold steadier.
The price of the power: when the window leaves its speech position, the vowel goes with it. In listening tests, tuned notes were preferred — but listeners could no longer tell which vowel was being sung (Vos et al. 2018). That's not a flaw; it's the trade every high soprano line makes, and why the words get hard to catch. Belting makes the opposite bargain: it keeps a maintained "yell" shape with spot 1 parked on the second harmonic — huge, speech-like, and bright, but it only works as high as the yell can reach.
Good notes by accident: even the frozen speech shape gets alignment boosts — whenever the climbing comb drags a harmonic across a fixed spot, the sound blooms for free, then fades as it drifts past. That's why untrained voices have lucky notes and dead notes scattered through their range, and can't say why. The meter shows the speech shape as 0 dB because it is the baseline — its accidents are already counted in. Training doesn't invent the boost; it makes the accident happen on purpose, on every note.
Why does this one spot in my voice fight me?
I hear you — and that fighting spot is not a flaw in you. It's a challenge that shows up in nearly every voice until a certain set of factors gets handled. The answer is different for males than for females, even though the two share something deep — so let's talk through the men first, then the women, and see if you can spot the difference and what they have in common.
Part 1 · The men's turn — slip down and hand off
Around E4–G4, the male voice meets the wall singers call the passaggio: the second harmonic climbs up to favorite spot 1 and keeps going. Something has to give. Do nothing and the sound cracks open. Chase it and you're yelling. Or make the classical move — the turn: let the vowel round (“ah” softening toward “aw”), let spot 1 slip down out of H2's way, and hand the job to spot 2, which catches the third harmonic instead. Same wall, three answers. Drag the pitch through the zone on each one.
This is measured, not folklore: when professional tenors sang an ascending line, only 2% of tones at E♭4 had H3 in charge — by G4 it was 99%, and their favorite spot 1 had dropped from about 725 to 573 Hz (Castle 2014). The turn is that handoff. Singers don't feel it as math; they feel the vowel rounding itself — "ah" drifting toward "aw/uh" — which is why teachers call it passive vowel migration (Bozeman): you allow it rather than force it.
Neither answer is wrong. "Stay open" is a real style — it's the call/belt family from station 8, bright and speech-like, and it costs more the higher it goes. The turn buys evenness and ease at the top of the staff, at the price of a rounder, less word-like vowel. What ruins voices is not choosing: hanging on to the open shape past its ceiling, or covering so early the middle voice goes woolly.
Part 2 · The women's turn — open and soar
Now the women. Same wall, different collider: a woman's speaking spot 1 sits higher, so her second harmonic clears it — but as she climbs toward the top of the staff, H1, the note itself, comes up underneath spot 1. Do nothing, and the very note you're singing slides past its own window. The women's escape runs the opposite direction from the men's: instead of slipping spot 1 down and handing off, she opens the jaw tall and takes spot 1 up to travel with the note itself. Singers call this open, soaring setup whoop — the same move as station 8's “drop the jaw,” arriving here as the women's turn. (Women feel an earlier, gentler shift lower in the range too; this part models the big turn near the top of the staff.)
What this feels like in a real voice — because computer tones understate it: when nothing is done, the timbre thins. The synergy is gone, and the body feels it before the mind does. You start pushing with more effort, stiffening, trying to rebuild something you can't quite locate — and without clarity about where the effort belongs, the result is a harsh, constricted yell, or, with enough squeeze, no sound at all. The whoop replaces all of that effort with placement. (A note on speakers: this simulation changes only the tube, never the pushing, and ordinary computer audio flattens the difference further — so let the picture, the meter, and this description carry what small speakers can't.)
Did you spot it? The similarity: both voices hit the same wall — a powerful low harmonic colliding with spot 1 — and both turns are window moves, not muscle moves. The difference: which harmonic collides (H2 for men; H1, the note itself, for women), and which direction the window escapes (men slide spot 1 down and hand the shine to H3; women take spot 1 up to rise with H1). Opposite directions. Same artistry: keep a window where your power is.
Wait — my ear has favorite sounds too?
It does — and it plays favorites hard. Every sound in this lab ends its journey inside one more tube: the listener's ear canal, about as long as the last joint of your little finger and closed at one end by the eardrum. Like every tube you've met here, it has favorite frequencies, and it amplifies them before the eardrum even moves. Find your own ear's sweet zone at this station, and you'll see why singers, babies, and alarm designers all aim for the same place.
Start with your volume low. Every tone here has the same wave size — any loudness difference you hear between frequencies is your own ear's doing. The A/B test is the cleanest proof: two waves of identical size, and almost everyone hears the second as far louder.
Why the peak is where it is: a tube open at one end and closed at the other loves the note whose quarter-wavelength fits inside it. Your ear canal is roughly 2.5 cm, and 343 m/s ÷ (4 × 2.5 cm) ≈ 3,400 Hz — and sure enough, the measured boost of the outer ear peaks near 3,000 Hz at about +20 dB. That's the second link of the amplification chain drawn live in station 7: an opera-level ring (+15 dB, the measured tenor figure) plus your ear (+20 dB) totals +35 dB — ring-band waves arriving about 56× taller at the eardrum than with neither boost. Your head grew an amplifier tuned to the exact band singers spend years learning to fill. Or rather — singers learned to fill the band your head amplifies.
The flip side — handle with care: because the canal delivers the most energy near 3–4 kHz, years of loud sound wear out the hearing sensors for that zone first. Audiologists literally look for a "4 kHz notch" when they test for noise damage. It's why musicians wear earplugs at rehearsal: the same amplifier that makes ring free also makes loudness costly, right where a musician can least afford it.
Okay — but what do I actually DO with all of this?
You take it into the world — and this board is how. A board of missions, each one circling back to a station you've already walked and pointing it at real rooms, real songs, real ears. Check missions off as you accomplish them — this page will remember.
And the board will grow: as we build new exercises together in class, they'll earn their pins here.
Prove it — take the science into your own hands
Train it — the studio exercises
Mix it — the artistry, on purpose
Carry it — into rooms, songs, and other people
About the science
About the sounds: everything here is computer-made from pure sine waves so each frequency can be heard on its own — a real voice and a real piano are messier and richer. In the Build-a-Voice mixer, the two lowest harmonics get a small fixed audibility lift in the audio (identical in every preset, so preset comparisons stay fair) because small speakers under-deliver low frequencies. The math is the same as the project poster: harmonics at f × 1, 2, 3 …, compared to a piano tuned to A = 440. The piano and tube stations use the poster's low C (130.81 Hz), whose dense harmonic comb samples the tube's resonances finely; the Build-a-Voice mixer uses middle C (261.63 Hz) so its bottom note stays audible on small speakers. About the hidden-melody station: the field trip links to Wolfgang Saus's own recorded hearing test (oberton.org/en/hearing-test-saus — his recordings, hosted on his own site). The listener figures (≈5% hear the melody through the full speech disguise; 20–30% once only the second resonance moves) are Saus's reported results from his sung test, not a controlled census. The two-camp finding — fundamental listeners vs. overtone listeners, with opposite left/right asymmetries in the auditory cortex, a balance that tracks musical experience — is Schneider et al. 2005 (Nature Neuroscience); Saus, Seither-Preisler & Schneider 2025 (Frontiers in Neuroscience) extends that work to overtone perception, with Saus himself among the authors. About the tube station: the buzz falls off at −12 dB per octave (Sundberg), the two movable favorite spots use classic measured vowel values for adult male voices (Peterson & Barney 1952), a third spot sits fixed near 2,500 Hz, and the sound leaving the lips gets a +6 dB-per-octave lift (the radiation effect) — a simplified but carefully built model; real tracts have more resonances and every voice differs. About the painting station: the cord slider spans measured source slopes (breathy ≈ −18 to firm ≈ −9.5 dB/oct); the length slider uses Fₙ ∝ 1/L (each 1% of length moves every resonance ≈1%); and the "ring" control is a deliberately simplified stand-in for the singer's-formant cluster — in a real voice that gleam is made by tract shape (a small larynx-tube outlet opening into a wide pharynx draws F3–F5 together; Sundberg 1974, refined by MRI studies since), not by a knob; its 0–19 dB range follows computational modeling of the cluster (7–19 dB gains; Yoshinaga & Zhang 2025) and matches long-term-average-spectrum comparisons in which trained opera voices run about 15–19 dB above untrained voices near 3 kHz. The painting model also carries the neutral tract's upper favorite spots near 3,500 and 4,500 Hz (scaled by the length slider) so the ring band isn't artificially empty; even so, simple textbook resonances roll off harder than living tracts, so the ring meter's dB readings run deeper than measured voices (real ring peaks sit roughly 6–20 dB below the main spectral peak, varying by vowel) — read that meter to compare presets against each other, not as a clinical value. Sopranos above roughly C5 use a different strategy entirely (tuning the first resonance to the note itself, worth +20–30 dB — station 8). The 16 dB / 10 dB figure is Sundberg 2001. About the tuning station: the R1:f₀ tracking follows Joliveau, Smith & Wolfe 2004 (Nature) and Garnier et al. 2010; the belt model (F1 near H2, "maintained yell") follows Bourne & Garnier and Schutte & Miller, with the ~980 Hz ceiling as a teaching simplification. The meter holds the cord behavior identical across strategies so it shows the filter's contribution alone — a real belt also uses firmer cord closure and higher breath pressure on top of the tuning shown here. About the turn station: the open/close-timbre framework is Bozeman's; the measured numbers (spot 1 dropping ≈725 → 573 Hz, H3 dominance 2% at E♭4 → 99% at G4 in professional tenors on an ascending line) are Castle 2014. The straight-line drop through the zone and the exact zone edges are teaching simplifications — real singers turn at different notes on different vowels, and register (cord) changes travel with the turn in ways this filter-only model leaves out. The women's part models a female "ah" with favorite spots about 15% above the male values and treats whoop as spot 1 tracking H1 — the R1:f₀ tuning measured by Joliveau, Smith & Wolfe and Garnier et al.; the same simplifications apply. About the listener station: the sensitivity curve is a smoothed teaching curve shaped on the ISO 226:2023 equal-loudness data, drawn relative to 1 kHz; the ~+20 dB outer-ear boost near 3 kHz and the quarter-wave estimate (343 m/s ÷ 4 × 2.5 cm ≈ 3,400 Hz) come from the measured literature; individual ears differ; the curve is drawn for comfortable listening levels (at very loud levels the ear's low-frequency disadvantage shrinks); and small speakers under-deliver the lowest tones, which exaggerates the low-end effect. The microphone station analyzes sound locally in your browser; nothing is recorded or uploaded. About the mission board: every mission repeats an experiment or listening habit already taught at its home station — nothing new is prescribed; check-offs are saved only in your own browser and never leave your device.
