Subtractive synthesis explained: oscillator, filter and VCA

Subtractive synthesis is how most analog synths work, from the 1970 Minimoog to today’s polysynths. The idea fits in one picture: start with a sound rich in harmonics, produced by an oscillator, remove part of it with a filter, the way a sculptor removes material, then shape the volume over time with an amplifier driven by an envelope. Three blocks and a handful of controls: understanding them is enough to program your own basses, pads and leads instead of relying on presets.

The principle: start with a rich sound

Every musical sound breaks down into a fundamental, the frequency that sets the pitch of the note, and harmonics, multiples of the fundamental that shape the timbre. An A at 110 Hz therefore contains components at 220, 330, 440 Hz and beyond. The more high harmonics a sound holds, the brighter it seems; the fewer, the rounder or duller it gets.

Subtractive synthesis uses that idea in one direction only: the oscillator supplies plenty of harmonics, and the filter decides which ones stay. It is the opposite of FM synthesis, which creates harmonics by modulating simple waves, and of additive synthesis, which stacks them one by one.

The oscillator and its waveforms

The oscillator (VCO, voltage controlled oscillator: an oscillator whose pitch is set by a voltage) repeats a waveform at the frequency of the note being played. Each waveform has its own harmonic signature:

The sawtooth is the raw material of choice: it contains everything, so the filter has plenty to work with. Several controls make the source richer still:

• detune: two oscillators slightly out of tune beat against each other and thicken the sound;
• sub-oscillator: a wave one or two octaves below the note, for weight in the low end;
• pulse width modulation (PWM): the width of the square wave moves continuously, for a natural chorus-like effect;
• sync: one oscillator forces the other to restart its cycle, the source of the aggressive leads typical of the 1980s.

A mixer then balances the oscillators and the noise before the filter. Pushing its level sometimes drives the filter input into gentle saturation, a sound many players look for on certain instruments.

The filter: the heart of the sound

The filter (VCF) attenuates part of the spectrum. The most common is the low-pass: it lets the lows through and removes the highs above a chosen frequency. Two controls do most of the work:

• cutoff: the point where the filter starts acting. Low, the sound is muffled; high, it is open and bright;
• resonance (also called emphasis or Q): it boosts the frequencies right around the cutoff. Turned all the way up, many analog filters self-oscillate and produce a sine wave you can play as an extra oscillator.

The slope: 12 or 24 dB per octave

The slope tells you how fast the filter attenuates past the cutoff. A 2-pole filter removes 12 dB per octave: the sound stays more open and nasal, as with the multimode filter of the Oberheim SEM. A 4-pole filter removes 24 dB per octave: the low end stays full while the highs fall away sharply, as with the famous ladder filter of the Minimoog. Neither slope is better: they serve different sounds, and many recent synths let you choose.

Other filter types

• high-pass: removes the lows, to thin out a pad or a cymbal sound;
• band-pass: keeps only a zone around the cutoff, for telephone-like or vocal timbres;
• notch (band-reject): carves out a narrow zone, useful in motion for a phasing effect.

One last control worth knowing: key tracking. It opens the filter as you play higher, so the top notes do not sound smothered.

The amplifier and the envelopes

The amplifier (VCA, voltage controlled amplifier) sets the volume of the voice. On its own it does nothing interesting: the envelope gives it a shape over time, with its four stages attack, decay, sustain and release, covered in our article on the ADSR envelope. A zero attack gives a percussive sound, a slow one a swelling pad.

Most subtractive synths have a second envelope dedicated to the filter. With an amount control, it opens the cutoff on each note and then closes it again: that is the “wah” of a funk bass or the pluck of a harp-like sound. Some filters accept a negative amount, which closes the filter on the attack instead of opening it; the Behringer Polyeight, for instance, offers positive or negative envelope control.

The LFO: movement

The LFO (low frequency oscillator) is an oscillator too slow to hear, from a fraction of a hertz to a few dozen. You do not listen to it: you use it to move another parameter. Depending on where it goes, it produces:

• on oscillator pitch: vibrato;
• on filter cutoff: a cyclic wah;
• on the VCA: tremolo;
• on pulse width: the PWM mentioned above.

Assigning the mod wheel or aftertouch to the LFO amount lets you dial the effect in while playing; our guide to aftertouch and MPE explains what each keyboard sends.

The full chain

Put end to end, these blocks make up a voice. A monosynth has one; a polysynth lines up several, one per note played at the same time, as explained in our article on polyphony and paraphony.

On a semi-modular synth or a Eurorack system, these connections run through patch cables and control voltages, so you can reorder them freely; our article on CV/Gate explains how that works.

Three recipes to get started

Start from an initialised patch (sawtooth oscillator, open filter, neutral envelopes), then apply these settings:

These values are starting points: every filter responds in its own way, and that is exactly what gives each instrument its character.

Where to practise

Nearly every analog synth works this way, from the entry-level models in our first synth guide to big polysynths such as the Roland JUPITER-08. To go deeper into the theory, Gordon Reid’s “Synth Secrets” series, published by Sound On Sound and free to read online, is still the most complete reference on the subject. Upcoming articles in our Subtractive Synthesis section will cover each block in detail, starting with the filter.