Additive Synthesis

The Fourier Transform in Plain Language

In 1822, Joseph Fourier demonstrated that any periodic waveform — no matter how complex — can be broken down into a series of simple sine waves. Each sine wave has a specific frequency, amplitude, and phase. Add them all together and you reconstruct the original waveform.

This is not an approximation. It is mathematically exact. A sawtooth wave is the sum of all integer harmonics (1, 2, 3, 4, 5…) with amplitudes that decrease as 1/n (the second harmonic is half the amplitude of the first, the third is one-third, and so on). A square wave is the sum of only the odd harmonics (1, 3, 5, 7…) with the same 1/n amplitude relationship. A triangle wave is also odd harmonics only, but with amplitudes decreasing as 1/n squared — the upper harmonics are much quieter.

This matters because it means the waveforms you have been using as raw material in subtractive synthesis — saw, square, triangle — are themselves composed of individual sine waves. The filter in subtractive synthesis reduces the amplitude of some of those component sines. Additive synthesis skips the filter entirely and lets you set each sine wave’s amplitude independently.

Building a Sound Harmonic by Harmonic

In VCV Rack, the Bogaudio Additator module is a dedicated additive oscillator. It gives you independent control over multiple harmonics — you can set the amplitude of each one individually.

Start by creating a patch with only the fundamental (harmonic 1) audible. You hear a pure sine wave. Now bring in harmonic 2 at about half the amplitude of the fundamental. The tone gains an octave overtone — it sounds fuller, slightly hollow. Add harmonic 3 at one-third amplitude. The sound develops more body. Continue adding harmonics 4, 5, 6, and beyond.

If you add all harmonics with amplitudes decreasing as 1/n, you end up with a sawtooth wave. You have just built it from scratch. But the power of additive is that you do not have to follow that recipe. You can set harmonic 5 louder than harmonic 3. You can leave out harmonic 4 entirely. You can boost harmonic 7 and suppress everything above harmonic 10. Each combination produces a unique timbre that no standard waveform and no filter curve could create.

Try building some recognizable timbres:

Clarinet-like: Emphasize odd harmonics (1, 3, 5, 7) and suppress even harmonics (2, 4, 6, 8). The clarinet’s cylindrical bore naturally reinforces odd harmonics, and this harmonic recipe captures that nasal, woody quality.

Bright string-like: All harmonics present, with a slight emphasis on harmonics 2-5 relative to the fundamental. The upper harmonics should be present but gradually decreasing. This produces a tone with the brightness of a bowed string.

Hollow organ-like: Fundamental strong, harmonic 2 absent, harmonic 3 at moderate level, harmonic 4 absent, harmonic 5 at low level. Gaps in the harmonic series produce the distinctive hollow quality of certain organ registrations.

Tuning Partials: Harmonic and Inharmonic Spectra

When partials sit at exact integer multiples of the fundamental, the result is a pitched, musical tone. But additive synthesis does not require you to stay harmonic.

Shift a partial slightly off its integer position — make the third partial 3.02 times the fundamental instead of exactly 3 — and the sound develops a subtle shimmer. The two nearly-coincident frequencies beat against each other. Push a partial further off — make it 3.5 or 4.7 times the fundamental — and the sound becomes inharmonic. The ear stops hearing a clear pitch and starts hearing something metallic, bell-like, or percussive.

This is the same territory FM synthesis covers with non-integer ratios, but additive gives you finer control. In FM, changing one ratio reshuffles the entire sideband structure. In additive, you can move a single partial without affecting any of the others.

Independent Envelopes: The Real Power

The additive approach gets genuinely powerful when each partial has its own amplitude envelope. In subtractive synthesis, the filter envelope shapes the entire spectrum at once — all harmonics follow the same brightness contour. In additive, each harmonic can evolve independently.

Consider how a piano note behaves. The attack contains a burst of high harmonics from the hammer strike. Those upper partials decay within the first 100 milliseconds. The middle harmonics sustain longer. The fundamental sustains longest of all. The timbral evolution of a piano note is not a simple bright-to-dark sweep — it is a complex, harmonic-by-harmonic decay where each partial has its own timeline.

Additive synthesis can model this. Assign fast-decaying envelopes to the upper partials, slower decay to the middle partials, and the slowest decay to the fundamental. The result captures the natural behavior of vibrating objects in a way that a single filter envelope cannot.

This is also where additive synthesis becomes labor-intensive. Setting individual envelopes for 20 or 30 partials is slow, detailed work. The payoff is precision. The cost is time.

Drawbar Organs: Additive Synthesis in Disguise

The Hammond organ, introduced in 1935, is an additive synthesizer. Each key triggers a set of tonewheels — rotating metal discs that generate sine waves at specific frequencies. The drawbars on the console control the volume of each tonewheel relative to the fundamental. Pull out the 8’ drawbar and you hear the fundamental. Pull out the 4’ drawbar and you add the octave harmonic. Pull out the 2-2/3’ drawbar and you add the fifth above the octave. Each drawbar corresponds to a harmonic partial, and the organist builds timbres by mixing them in real time.

The standard Hammond drawbar labeling (16’, 5-1/3’, 8’, 4’, 2-2/3’, 2’, 1-3/5’, 1-1/3’, 1’) maps to sub-octave, fifth, fundamental, octave, octave+fifth, two octaves, two octaves+major third, two octaves+fifth, and three octaves. These are harmonics 1 through 8 of the fundamental (with the sub-octave as a bonus), each independently controllable from 0 to 8 in amplitude.

Organists develop registrations — specific combinations of drawbar settings — the way synthesists develop patches. The classic “full organ” registration (all drawbars at 8) is every harmonic at full volume, producing a bright, buzzy tone. A more restrained jazz organ registration might use only the fundamental, octave, and a touch of the fifth — warm and round.

In VCV Rack, the Squinky Labs Organ Three module gives you a drawbar-style additive engine. Experiment with different drawbar combinations and listen to how each harmonic changes the overall character. Pull out one drawbar at a time and learn what each partial contributes to the whole. This is additive synthesis at its most tactile.

Resynthesis: From Recording to Partials and Back

The most powerful application of additive synthesis is resynthesis: analyze a recording’s harmonic content frame by frame, then reconstruct it as a set of individual partials you can manipulate. Stretch harmonics apart for an otherworldly effect. Freeze a single frame and sustain it indefinitely. Crossfade the partials of two different sounds to morph one into the other.

This is how additive synthesis connects to the Musician Basics concept of the harmonic series — every sound contains partials, and resynthesis lets you see and reshape them individually. In your DAW, Logic’s Alchemy and Ableton’s Wavetable both use spectral analysis under the hood. VCV Rack modules that perform FFT analysis can do the same thing explicitly.

Additive vs. Subtractive: Trade-Offs

Additive synthesis offers total control over the frequency spectrum. You can create any timbre that exists and many that do not. You can model the behavior of acoustic instruments with more accuracy than subtractive synthesis allows, because you can give each partial its own amplitude contour.

The trade-off is complexity. A subtractive patch has a handful of parameters: oscillator waveform, filter cutoff, resonance, two or three envelope shapes, maybe some LFO modulation. An additive patch with 32 partials, each with its own amplitude, frequency offset, and envelope, has hundreds of parameters. Designing sounds from scratch in additive requires patience and a clear mental model of what you are trying to build.

In practice, most musicians encounter additive synthesis in specific contexts rather than as a primary sound design tool. Organ emulations use it. Spectral processing plugins use it under the hood. Resynthesis tools use it. FM synthesis is, in a mathematical sense, a shortcut for generating complex additive spectra with far fewer controls. Wavetable synthesis (next chapter) offers another solution to the complexity problem by storing pre-computed spectra and morphing between them.

Understanding additive gives you the conceptual foundation for all of these. When you know that every sound is a collection of sine waves at specific frequencies and amplitudes, you understand what every other synthesis method is actually doing — just by different means.

What to Practice

• Using the Bogaudio Additator (or multiple sine oscillators tuned to harmonic intervals), build a sawtooth wave from scratch by adding harmonics one at a time with amplitudes at 1/n. Verify against a real sawtooth oscillator using the Bogaudio Analyzer module — the spectra should match.
• Build a square wave by adding only odd harmonics. Compare against a real square oscillator.
• Create a timbre that does not correspond to any standard waveform: emphasize harmonics 3, 5, and 7 while suppressing harmonics 1, 2, 4, and 6. Listen to the result. Then rearrange — make different harmonics prominent. Develop an intuition for what each harmonic contributes to the overall sound.
• Experiment with inharmonic tuning. Take a harmonic additive patch and shift one partial slightly off its integer frequency. Listen for the beating. Now shift it further. At what point does the sound stop registering as pitched and start sounding like a bell or metallic object?
• If you have access to the Squinky Labs Organ Three module, learn three drawbar registrations by ear: a full bright organ, a warm jazz organ, and a pipe-organ-like principal stop. Write down the drawbar settings for each. You are programming an additive synthesizer with nine parameters — considerably more manageable than 32 independent partials.
• Take any short recorded sound (a spoken vowel, a plucked string, a horn note) and look at its spectrum in an analyzer. Try to approximate that spectrum using additive partials. You will not match it exactly, but the attempt will teach you more about the relationship between spectra and timbre than any amount of reading.