BUILDING TIMBRE FROM NOTHING
SPECTRAL SCULPTOR EXERCISE — FUZZBOX PHYSICS
A spectrogram is a complete picture of a sound. Every sound you have ever heard could
be drawn as a pattern on this grid — frequency on the vertical axis, time on the horizontal,
brightness for loudness. In this exercise you will build sounds from silence by painting
directly into the frequency domain, then take apart a real sound to see what it's made of.
Open the Spectral Sculptor in a new tab
TOOLS REFERENCE
The Brush and the Rake
The Sculptor has two painting modes. Learn both before you start.
BRUSH — The default. Paints energy at the frequency under your cursor.
Adjust SIZE for width and INTENSITY for loudness.
Good for precise single-frequency work.
RAKE — Slide the RAKE control above 1.
Now when you paint at any frequency, the tool automatically paints harmonics above it
at 2×, 3×, 4×, etc. — all from a single drag. Higher harmonics are
progressively quieter, just like in nature. Set it to 8 for a full
harmonic series.
ODD ONLY — Click this button to make the rake skip even harmonics
(2nd, 4th, 6th). It only paints 1st, 3rd, 5th, 7th — the recipe for a square wave
or clarinet tone.
ERASE — Click ERASE (−) or hold
SHIFT while dragging to remove energy.
The rake works in erase mode too — it removes all harmonics at once.
The frequency readout at the bottom of the spectrogram shows the exact Hz under your
cursor. Use it to place your lines precisely.
PART 1
The Harmonic Series by Hand
Every pitched sound is built from a stack of sine waves at whole-number multiples
of a fundamental frequency. You are going to prove this by building one yourself.
1
In the Sculptor, click SILENCE to start with an empty spectrogram.
Set brush size to 3 and intensity to about 30%.
Make sure RAKE is set to 1 (off) — you'll
do this one by hand first.
2
Paint a thin horizontal line at 220 Hz (that's A3).
Drag slowly from left to right to fill the whole 2-second loop.
Press play.
You should hear a pure sine tone — smooth, featureless, almost electronic.
3
Now paint a second horizontal line at exactly 440 Hz (2 × 220).
This is the second harmonic. Listen again.
Same pitch, but it sounds different — brighter, more present. That's timbre changing.
4
Keep adding harmonics one at a time. Each one is at a whole-number multiple of 220 Hz.
Listen after each new harmonic and notice how the timbre evolves:
| HARMONIC | FREQUENCY | INTERVAL |
| 1st | 220 Hz | fundamental |
| 2nd | 440 Hz | octave |
| 3rd | 660 Hz | octave + fifth |
| 4th | 880 Hz | two octaves |
| 5th | 1100 Hz | two octaves + major third |
| 6th | 1320 Hz | two octaves + fifth |
| 7th | 1540 Hz | two octaves + minor seventh |
| 8th | 1760 Hz | three octaves |
5
After adding all eight, listen. You've built something close to a sawtooth wave
— the same waveform an analog synth generates. Every harmonic is present, each a little quieter than the last.
6
Now try it the fast way. Click SILENCE again. Set the
RAKE slider to 8.
Paint a single line at 220 Hz — all eight harmonics
appear at once, with natural rolloff.
Compare the sound to what you built by hand.
The rake applies 1/n rolloff — the 2nd harmonic is half as loud as the fundamental,
the 3rd is a third as loud, and so on. This is the natural spectrum of a sawtooth wave.
When you painted by hand, did you make the higher harmonics progressively quieter?
If not, your hand-built version probably sounded buzzier and brighter.
SCREENSHOT 1
Your hand-painted harmonic series spectrogram with all eight harmonics visible as horizontal lines.
PART 2
What Makes a Clarinet?
1
Click SILENCE to start fresh.
Set RAKE to 8 and click the
ODD ONLY button (it should highlight).
2
Paint a single horizontal line at 220 Hz.
The rake paints only the odd harmonics for you — 1st, 3rd, 5th, 7th:
| HARMONIC | FREQUENCY | EVEN/ODD |
| 1st | 220 Hz | odd — painted by rake |
| 2nd | 440 Hz | even — skipped |
| 3rd | 660 Hz | odd — painted by rake |
| 4th | 880 Hz | even — skipped |
| 5th | 1100 Hz | odd — painted by rake |
| 6th | 1320 Hz | even — skipped |
| 7th | 1540 Hz | odd — painted by rake |
3
Press play and compare to Part 1.
It sounds hollow, woody, nasal — like a clarinet or a square wave.
The missing even harmonics create gaps in the spectrum that give it that distinctive color.
4
Turn off ODD ONLY and set RAKE back
to 1. Now paint in one even harmonic by hand — a line at
440 Hz. Listen to how the character shifts.
The hollowness fills in.
A clarinet's cylindrical bore naturally cancels even harmonics. An oboe's conical bore
keeps them. This is why they sound so different playing the same note.
The physics of the tube determines which harmonics survive.
SCREENSHOT 2
Your odd-harmonics-only spectrogram. Write one sentence: how does removing the even
harmonics change the sound compared to Part 1?
PART 3
Destroy and Discover
1
Click the VOWEL preset. It generates a synthetic "ahh" sound.
Study the spectrogram before you touch anything.
You should see several bright horizontal bands at specific heights.
These are formants — resonant frequencies
created by the shape of the mouth and throat. Every vowel has a different
pattern of formants. This is how your brain tells "ah" from "ee" from "oo."
2
Switch to the eraser (hold Shift while dragging, or use a large brush
at high intensity to paint over areas). Erase the lowest bright band
(around 200–400 Hz). Listen.
The vowel should sound thinner, less full — you removed the first formant, which
carries the "body" of the voice.
3
Reload VOWEL. Now erase everything except the fundamental
(the very lowest line). Leave only that single horizontal stripe at the bottom.
Same pitch — but it doesn't sound like a voice anymore. It's just a sine tone.
Timbre is everything above the fundamental.
The fundamental tells you the note. Everything else tells you what's playing it.
4
Reload VOWEL one more time. Now erase the fundamental
(lowest line) and keep everything else.
Surprise — you still hear roughly the same pitch. Your brain reconstructs the fundamental
from the harmonic series above it. This is called the missing fundamental
effect. It's why tiny phone speakers can reproduce bass notes whose wavelengths
are longer than the speaker itself.
SCREENSHOT 3
Your vowel with the fundamental erased. Write one sentence: why can you still hear the pitch?
PART 4
Spectral Self-Portrait
1
Click RECORD and say "ahhh" steadily for 2 seconds. Let it loop
and study your spectrogram.
2
Find your fundamental frequency — the lowest bright horizontal line.
It's probably between 85 Hz (low male voice)
and 250 Hz (high female voice). Note the frequency.
3
Count the harmonics above it. Are they evenly spaced? (They should be — each one
is at a whole-number multiple of your fundamental.) How many can you see before
they fade into the noise floor?
4
Find the formants — the frequency regions where the harmonics are
brightest. For "ahh," you should see strong energy around
700–800 Hz (first formant) and
1000–1200 Hz (second formant).
5
Now click SILENCE. From memory and from what you observed,
try to repaint your own voice. Use the RAKE
set to about 10–12 harmonics and paint at your fundamental
frequency — you'll get the whole harmonic stack in one stroke. Then switch to the regular
brush (RAKE = 1) and boost the formant regions where the
harmonics should be brighter.
Don't expect a perfect recreation — the point is to see how close you can get and to
understand what your voice is actually made of. If it sounds vaguely vowel-like,
you've understood the physics.
Voices are rich and complex. A typical speaking voice has 20+ audible harmonics,
vibrato (slight pitch wavering), noise from breath, and formant transitions.
Getting even a rough approximation by painting teaches you more about voice
acoustics than any textbook diagram.
SCREENSHOT 4A
Your recorded voice spectrogram, with your fundamental frequency labeled.
SCREENSHOT 4B
Your painted recreation. Write 2–3 sentences: what was hardest to recreate? What does that
tell you about the complexity of human voice?
PART 5 — BONUS
Spectral Score
Composers like Iannis Xenakis and Aphex Twin have hidden images inside spectrograms.
The spectrogram is a canvas. Whatever you draw, you will hear.
1
Start from SILENCE. Using the brush, draw something — a shape,
a letter, an abstract pattern, a face. Don't think about what it will sound like.
Just draw.
2
Press play and listen. Now adjust your drawing to make it sound more interesting.
Notice: horizontal lines sustain, vertical lines click, diagonals sweep,
filled regions roar, thin lines sing.
3
Try to make a 2-second sound that tells a story or describes a place.
A rainstorm. A machine starting up. A creature calling. An explosion in reverse.
Think cinematically.
Horizontal = tone. Vertical = percussion. Diagonal = glissando.
Wide band = noise/texture. Narrow band = pitch. Low = rumble. High = shimmer.
You are composing with the raw physics of sound.
SCREENSHOT 5
Your spectral score. Title it and write one sentence describing what you were trying to evoke.
What to Turn In
Five screenshots with short written responses as described above. Submit as a single PDF or a series of images.
There are no wrong sounds — the goal is to understand the relationship between what you see in the spectrogram
and what you hear.