Subtractive Synthesis

The Subtractive Architecture

The signal flow is straightforward. An oscillator generates a raw waveform — sawtooth, square, pulse, or sometimes a combination. That signal passes through a filter, which removes some of the harmonics. The filtered signal then passes through an amplifier, which controls its volume over time via an envelope. Modulation sources (LFOs, additional envelopes, performance controllers) connect to any of these stages to create movement.

Oscillator into filter into amplifier. Three stages, in that order. Every subtractive synthesizer ever made follows this skeleton, from a Minimoog to the soft synths in your DAW.

The reason the architecture works so well is that it mirrors how many acoustic instruments behave. A plucked guitar string starts bright (rich in harmonics) and mellows as it decays (the upper harmonics die off first). A subtractive patch does the same thing: a sawtooth wave through a low-pass filter with a decaying envelope on the filter cutoff. The parallel is not accidental — subtractive synthesis was designed to approximate what acoustic instruments do naturally, and its vocabulary of controls maps onto our intuition about how sounds behave.

Audio Path vs. Modulation Path

In VCV Rack, you have been working with two kinds of cables: audio signals and control voltages. In a subtractive voice, these two paths serve distinct purposes.

The audio path carries the sound itself. It runs from the oscillator output, through the filter input/output, and into the amplifier. The signal in this path is the waveform you will eventually hear through your speakers.

The modulation path carries control signals — envelopes, LFOs, keyboard voltages — that shape the audio path’s behavior without being heard directly. An envelope connected to the filter cutoff does not add its waveform to the sound. It tells the filter where to open and close.

Separating these two paths in your mind — and eventually in your patching habits — is the single most important conceptual tool in modular synthesis. A subtractive voice is a clear place to internalize it because the roles are unambiguous: the oscillator, filter, and amplifier are the audio path. Everything else is modulation.

Building a Subtractive Voice in VCV Rack

Open VCV Rack and place the following modules:

1. A VCO (voltage-controlled oscillator) — use the one you are comfortable with from Chapter 2. Select the sawtooth output.
2. A VCF (voltage-controlled filter) — low-pass mode. Set the cutoff somewhere around the midpoint.
3. A VCA (voltage-controlled amplifier).
4. Two envelope generators — one for the filter, one for the amplifier.
5. A MIDI-to-CV module to turn your keyboard into pitch and gate signals.

Connect the MIDI-to-CV pitch output to the VCO’s V/Oct input. Connect the VCO’s sawtooth output to the VCF’s audio input. Connect the VCF’s output to the VCA’s audio input. Connect the VCA’s output to your audio interface.

Now connect the gate output from MIDI-to-CV to both envelope generators’ gate inputs. Patch one envelope’s output to the VCF’s cutoff CV input. Patch the other envelope’s output to the VCA’s CV input.

Play a note. You should hear a sawtooth wave shaped by both envelopes — one controlling brightness, the other controlling volume.

This is a complete subtractive voice. Everything else is refinement.

Shaping the Filter Envelope

The amplifier envelope controls when and how long the sound plays. The filter envelope controls how the timbre changes during that time. These two envelopes working together are what give a subtractive patch its character.

Set your amplifier envelope to something simple: fast attack, no decay, full sustain, moderate release. This gives you a straightforward organ-like volume shape — the sound comes on when you press the key and fades when you let go.

Now experiment with the filter envelope. This is where subtractive synthesis gets expressive.

A short, punchy filter envelope (fast attack, moderate decay, low sustain) produces a plucky sound. The filter opens briefly on the attack, exposing the bright harmonics, then closes down. Bass patches, plucked sounds, and percussive leads use this shape. The initial burst of brightness gives the ear something to grab onto, then the sound settles into a warmer, more sustained tone.

A slow filter envelope (slow attack, no decay, full sustain) produces a gradual brightening — the sound starts dark and opens up over time. Pad sounds and evolving textures use this approach.

A fast attack with moderate decay and medium sustain gives you a brass-like quality. The sound pops open bright, then settles into a sustained tone that retains some harmonic content. This is the classic Minimoog lead sound.

The amount of envelope modulation matters as much as the shape. Turn the filter envelope’s modulation depth up and the effect is dramatic — the cutoff sweeps over a wide range. Turn it down and the effect is subtle, just a hint of timbral shift on the attack. Start with moderate depth and adjust by ear.

Resonance in Context

You explored resonance in the filters chapter. In a complete subtractive voice, resonance takes on a musical role beyond just boosting frequencies near the cutoff.

Adding moderate resonance to a bass patch emphasizes the point where the filter is cutting, which gives the bass a nasal, focused quality. Acid bass lines — the 303 sound — depend on high resonance combined with a filter envelope that sweeps the cutoff up and down. The resonant peak follows the cutoff as the envelope moves it, creating that distinctive squelchy character.

On pad sounds, a touch of resonance adds definition to the filter movement. Without resonance, a slow filter sweep sounds smooth but vague. With resonance, the sweep has a point — a specific frequency that gets emphasized as the cutoff moves through it.

Be careful with high resonance values. At extreme settings, many filters self-oscillate — they start producing their own pitched tone that can overpower the oscillator’s output entirely. This is sometimes useful (you can play the filter as a sine wave oscillator by feeding pitch CV to the cutoff), but it will blow up your patch if you are not expecting it.

Adding a Second Oscillator

Most classic subtractive synthesizers have two or three oscillators, and for good reason. A single oscillator produces a clean, focused tone. Two oscillators detuned slightly against each other produce a thicker, warmer sound because of the beating between their nearly-identical frequencies.

Add a second VCO to your patch. Connect the same V/Oct signal to both oscillators. Route both outputs into a mixer, then send the mixed signal to the filter.

Now detune the second oscillator by a few cents — just enough that you can hear the pitch relationship wobble. The beating between the two oscillators creates movement that no amount of modulation on a single oscillator can replicate. This is the foundation of the fat analog sound that people associate with vintage synthesizers.

Try setting the two oscillators to different waveforms: sawtooth on one, square on the other. The combination creates a spectrum that neither waveform has on its own. You can also set the second oscillator an octave lower for a beefier bass foundation, or a fifth above for a more harmonically complex tone.

Classic Subtractive Instruments

The instruments that defined subtractive synthesis are worth knowing, because their architectures show up as templates in nearly every soft synth you will encounter.

The Minimoog (1970) established the three-oscillator, one-filter architecture that became the standard. Its ladder filter — a four-pole low-pass with a distinctive warm rolloff — is one of the most imitated circuits in synthesizer history. The Minimoog is monophonic: one note at a time. That constraint forced players to develop expressive single-line techniques, and the Minimoog’s pitch bend and modulation wheels became standard performance controls on every synthesizer that followed.

The ARP 2600 (1971) was semi-modular — it had a default signal path (oscillators into filter into amp, like the Minimoog) but also exposed patch points for every module. You could override the default routing by plugging in cables. This design philosophy is exactly what VCV Rack gives you: a default nothing, requiring you to build every connection yourself.

The Prophet-5 (1978) was the first polyphonic synthesizer with patch memory. Five voices, each with two oscillators and a low-pass filter, and the ability to store and recall 40 patches. Before the Prophet-5, every patch was gone the moment you turned the machine off.

The Roland SH-101 (1982) stripped subtractive synthesis down to one oscillator, one filter, one envelope, and a built-in sequencer. It was inexpensive, portable, and its simplicity made it fast to program. The 101 became a staple of early electronic music precisely because it did not try to do everything — it did one thing with clarity and character.

These instruments share the same signal path you just built in VCV Rack. The differences are in the specific character of their oscillators and filters, their modulation options, and their interface design. When you see a soft synth modeled after any of these instruments, you already know the architecture.

Common Subtractive Patches

With the two-oscillator voice you built, try these starting points:

Fat bass: Both oscillators on sawtooth, second oscillator one octave below the first. Filter cutoff low, moderate resonance. Filter envelope with fast attack, moderate decay, zero sustain. Amplifier envelope with fast attack, no decay, full sustain, short release. The filter envelope amount controls how much the attack pops — more modulation for a punchier bass, less for a rounder one.

Brass lead: Both oscillators on sawtooth, slightly detuned. Filter cutoff at about 40%, moderate resonance. Filter envelope with medium attack (maybe 100-200ms), short decay, high sustain. This gives the attack a swell that mimics the way a brass player pushes into a note. Add some vibrato with an LFO on the oscillator pitch — not too fast, and delayed so it comes in after the note has been held for a moment.

Warm pad: Both oscillators on sawtooth or a sawtooth/pulse combination, detuned a few cents apart. Filter cutoff relatively low, minimal resonance. Filter envelope with slow attack (1-2 seconds), no decay, full sustain. Amplifier envelope with slow attack and long release. The sound fades in gradually and lingers when you release the key.

Pluck: Single oscillator, square wave. Filter cutoff low, no resonance. Filter envelope with instant attack, fast decay (100-200ms), zero sustain. Amplifier envelope with instant attack, moderate decay, zero sustain. The result is a short, bright pop that decays quickly — the filter and amplifier both close at roughly the same rate, giving the sound a natural plucked quality.

Limitations of Subtractive Synthesis

Subtractive synthesis is powerful, but it cannot do everything. Its palette is constrained by what a filter can do to a harmonically rich source.

You cannot easily create metallic, bell-like, or inharmonic tones. Filters remove harmonics — they do not create new ones or rearrange existing ones into non-integer relationships. A bell sound requires inharmonic partials (frequencies that are not integer multiples of the fundamental), and a low-pass filter on a sawtooth wave will never produce that.

You cannot independently control individual harmonics. The filter acts on the entire spectrum at once — everything above the cutoff gets attenuated. If you want harmonic 3 louder and harmonic 7 quieter while leaving harmonics 1, 2, 4, 5, and 6 unchanged, subtractive synthesis has no mechanism for that.

Evolving, morphing timbres are possible but limited. You can sweep a filter and modulate pulse width, which creates movement, but the timbral vocabulary is bounded by the waveforms your oscillators produce and the shapes your filter can carve from them.

These limitations are not flaws — they are boundaries. The chapters that follow explore synthesis methods that address each gap: FM synthesis for inharmonic and metallic tones, additive synthesis for per-harmonic control, wavetable synthesis for complex timbral morphing, and sampling for capturing any sound at all.

Subtractive synthesis remains the backbone. The other methods extend what it can do. Understanding this architecture thoroughly means you will recognize its signal path inside every other synthesis method you encounter — because most of them bolt their unique features onto a subtractive framework at some point in the signal chain.

What to Practice

• Build the complete subtractive voice described in this chapter from scratch in VCV Rack. Do not use a template — patch every connection yourself. Once it works, save it. This is your reference patch for the rest of the guide.
• With the amplifier envelope set to a simple sustaining shape, experiment with at least five different filter envelope shapes. For each one, write down (or just remember) what kind of sound it produces — pluck, brass, pad, etc. The goal is to internalize the relationship between filter envelope shape and timbral character.
• Add a second oscillator and experiment with detuning. Start with both oscillators at the same pitch and slowly increase the detune on the second one. Listen for the point where the beating becomes a chorus effect, and the point where it becomes two distinct pitches. The musical sweet spot is usually well before that second threshold.
• Try different waveform combinations on your two oscillators: saw + saw, saw + square, square + triangle, saw + sub-octave saw. Each combination has a different harmonic profile, and the filter will respond differently to each one.
• Recreate the four common patches described above (fat bass, brass lead, warm pad, pluck). Once you have each one working, start modifying — change the oscillator waveform, adjust the envelope times, move the filter cutoff. The point is not to memorize recipes but to develop intuition for how each parameter affects the result.
• If you have a soft synth (or a hardware synth) that uses subtractive architecture, open it alongside your VCV patch and find the same controls. Identify the oscillator section, the filter, the amplifier envelope, and the filter envelope. Every subtractive synth organizes these differently on its panel, but the signal path underneath is what you just built.