A classic subtractive synthesizer builds sound along a fixed chain: oscillators produce a harmonically rich raw waveform, a mixer blends them, a filter removes highs, an amplifier controls loudness, and envelopes and LFOs move all of it over time. Once you understand this chain you can build sounds on purpose instead of scrolling through presets. This article walks through every stage, explains ADSR envelopes in detail, compares the main synthesis types and gives starting points for bass, pad and lead sounds.
The signal path at a glance
From key press to sound, the signal in a subtractive synth usually passes these stages:
- Oscillators generate the waveform and set the pitch.
- Mixer sets the level balance between the oscillators (often including a sub-oscillator or noise source).
- Filter removes or emphasizes frequencies and shapes the tone colour.
- Amplifier controls the volume contour.
- Effects (chorus, delay, reverb) sit at the end of the chain on many instruments.
Envelopes and LFOs are not stages in the audio path but control sources: they move parameters such as filter cutoff, volume or pitch automatically. In analog instruments voltages control these blocks, hence the names VCO (oscillator), VCF (filter) and VCA (amplifier). A digital instrument does the same jobs in software; the principle is identical.
Oscillators and waveforms
The oscillator is the sound source and sets the basic character. The four standard waveforms differ mainly in how many harmonics they contain:
- Sawtooth: the richest in harmonics, bright and buzzy, string-like. The default for basses, leads and pads.
- Square/pulse: hollow and reed-like. Changing the duty cycle (pulse width) makes the tone thinner and more nasal; slowly modulating the pulse width gives the typical chorus-like movement.
- Triangle: soft, flute-like, few harmonics.
- Sine: the purest tone with no harmonics at all, ideal for sub bass.
Detune, unison and sub-oscillator
Two oscillators tuned slightly apart create beating: the sound gets wider and more alive. A few cents are enough; one lead example uses about 25 cents (rule of thumb). Unison stacks several voices per note with slight detuning and sounds correspondingly thick. A sub-oscillator runs an octave lower and gives bass sounds a foundation. With extreme detuning the pitch quickly turns unclear, which is a deliberate effect rather than a fault.
Mixer
The mixer decides which oscillator dominates. A 50:50 balance gives an even blend, while a quiet second oscillator only adds subtle colour. When the sum gets very loud, analog instruments can distort slightly in the filter, which some players like.
The filter in detail
For most players the filter is the most expressive control. The low-pass filter, the most common type, lets low frequencies through and cuts everything above the cutoff frequency. The further you turn it down, the duller a bright saw becomes. Other types are high-pass (removes bass), band-pass (passes only one band) and notch (blocks a narrow band).
Slope: 12 dB or 24 dB
The slope tells you how quickly the filter cuts above the cutoff, measured in decibels per octave. Every 6 dB corresponds to one pole: 6 dB (one pole) is gentle, 12 dB (two poles) is considered soft and warm, and 24 dB (four poles) cuts much harder and sounds more dramatic. As a rule of thumb, 12 dB suits natural-sounding tones, while 24 dB suits classic fat basses and sweeps. The filter design also shapes the character: a ladder design is often described as warm and squelchy, other circuits as smooth or as aggressive and distorted.
Resonance
Resonance (also called emphasis or Q) boosts the frequencies around the cutoff, essentially feedback within the filter. A little resonance makes the tone more pronounced, more makes it biting, nasal and vocal. At full resonance many filters begin to oscillate, producing a whistling sine tone at the cutoff frequency. Because resonance emphasizes the mids, high-resonance patches often sound sharper; check the volume after setting it.
Keytracking and the filter envelope
With keytracking the cutoff follows the note you play: high notes sound brighter, low notes darker. Without it, high notes sound dull at a low cutoff while low notes sound disproportionately bright. The filter envelope moves the cutoff over the duration of every note and creates the typical flash at the start of a tone, see below.
The amplifier and loudness
The amplifier (VCA) controls the loudness of each voice. In analog instruments it often works more as a controllable attenuator. What matters is what drives it: usually its own envelope, which decides whether the note starts instantly and dies away or swells slowly. If an LFO also modulates the amplifier, you get tremolo.
ADSR envelopes in detail
An envelope describes how a parameter develops over the length of a note. The classic ADSR has four stages, three of which are times and one is a level:
- Attack: time from key press to the peak. Short gives a percussive start, long gives a swell.
- Decay: time from the peak down to the sustain level.
- Sustain: the level held as long as the key stays down. It is the only stage that is not a time.
- Release: time to fade out after the key is released.
An everyday detail: with sustain at zero, the note dies away after the decay even if you keep holding the key, so the envelope behaves like a plucked or struck sound. With sustain at maximum, the decay has no audible effect.
Amp envelope and filter envelope
Almost every synthesizer has at least two envelopes. The amp envelope controls volume and decides whether a sound feels plucked, held or pad-like. The filter envelope controls the cutoff: a short decay at a medium cutoff with a high envelope amount gives the classic "wow" start, where the sound flashes briefly and then turns duller. The interplay of the two is the most important trick in sound programming: same waveform, same filter setting, different envelopes, and a bass turns into a pad. Many instruments offer a further envelope for pitch, for example for a percussive pitch drop.
Starting points for typical sounds
The table gives starting points (rules of thumb) that you adapt to your instrument and taste. Scales differ between synths, so treat the times as orders of magnitude.
| Sound | Oscillators | Filter | Amp envelope | Extras |
|---|---|---|---|---|
| Bass | Saw or square plus sub-oscillator | Low-pass, medium to low cutoff, low to medium resonance, filter envelope with short decay | Attack 0, short to medium decay, high sustain, short release | play monophonic, maybe portamento |
| Pad | Two detuned saw or triangle waves | Low-pass, fairly low cutoff (roughly 600 to 800 Hz), little or no resonance | Attack from about 1.5 s, decay about 1 s, sustain about 50 %, release about 1 s or longer | chorus, reverb, slow LFO |
| Lead | Two slightly detuned saws, or unison | Low-pass, fairly open cutoff, little resonance | Attack 0, high sustain | LFO on pitch for vibrato, portamento, delay |
| Pluck | Saw or square | Low-pass with a short filter envelope | Attack 0, decay about 200 to 700 ms, sustain 0, short release | percussive, good for arpeggios |
A well-known special case is the acid bass: medium cutoff, high resonance and a strongly modulated filter frequency. A Reese bass combines saw or square waves with unison and detuning, with the filter set low, roughly in the 200 to 800 Hz range.
LFO and modulation
An LFO (low-frequency oscillator) oscillates so slowly that you do not hear it as a pitch but as a repeating movement. Most LFOs run below about 20 Hz, and some instruments extend into the audible range, which yields new tone colours. The main destinations:
- Pitch: vibrato, with a slow rate and small depth.
- Filter cutoff: wah effects and breathing pads.
- Volume: tremolo.
- Pulse width: lively, chorus-like square tones.
The LFO waveform defines the character: sine and triangle wobble smoothly, square jumps between two values (trills, rhythmic effects), sawtooth rises slowly and drops abruptly, and sample & hold produces random steps. Rate sets the speed and depth (amount) the strength. Many instruments offer tempo sync to a MIDI clock, so the LFO runs in note values, plus key sync, which restarts it with every new note. For connecting gear, see also MIDI vs. CV/Gate explained.
Modulation in general means a source (envelope, LFO, velocity, mod wheel, aftertouch) controls a destination (cutoff, pitch, volume and more). Small, well-chosen modulations make the difference between static and alive.
Synthesis types in brief
- Subtractive synthesis: harmonically rich waveforms are filtered. The classic and easiest principle to grasp; the sound is often described as warm and rich.
- Additive synthesis: many sine waves are summed into one sound. Very precise, but laborious to operate.
- FM synthesis: oscillators modulate each other's frequency following fixed algorithms. It yields glassy, harmonically rich, percussive sounds and is less intuitive to program. The first widely popular digital synthesizer of this kind appeared in 1983 and sold over 100,000 units.
- Wavetable synthesis: the oscillator reads a table of many short waveforms and blends between them, so the timbre changes smoothly.
- Granular and sample-based methods: recordings are cut into very short grains (one hundredth to one tenth of a second) or used as an oscillator source. Modern hybrid instruments combine samples with many filter types, several envelopes and LFOs.
How analog, digital and virtual-analog designs differ in daily use is covered in analog vs. digital vs. virtual analog.
Polyphony and voices
Polyphony is the number of notes that can sound at once. An eight-voice synth plays eight notes in parallel; on the ninth note the oldest is usually cut off. Every voice has its own chain of oscillators, filter and envelopes. A monophonic instrument is often enough for bass and lead, while chords and pads need several voices. If you use long release times, plan for more voices, since fading notes keep occupying voices. More in monophonic vs. polyphonic.
Typical beginner mistakes
- Changing everything at once: change one control per pass and listen.
- Too much resonance: the sound turns shrill and can get uncomfortably loud. Start low.
- Mixing up sustain and release: sustain is a level, release a time. If a note never stops or cuts off at once, check there first.
- Filter envelope with no amount: if the envelope amount at the filter is zero, nothing happens however you set the values.
- Forgetting keytracking: high notes then sound dull, low notes harsh.
- Buying by demos alone: sound examples are often heavily produced. What counts is how quickly you reach good sounds yourself.
- Too many effects: reverb and delay can hide a weak basic tone but not improve it.
What matters when buying hardware
When buying a synthesizer, handling matters more than the feature count: more features do not automatically mean better sound, and a unit where every change is several menu pages deep slows creativity down. Direct access to filter, envelopes, oscillators and modulation, ideally with dedicated knobs, is what you want. Also check:
- How many voices and oscillators per voice the instrument has, and whether that fits your music.
- Whether it offers 12 dB, 24 dB or switchable filters.
- How many envelopes and LFOs there are and how flexible the routing is.
- Key feel, velocity response and aftertouch, ideally tried in person.
- Connections such as MIDI, USB and headphone output, plus accessory costs for case, cables, pedal and audio interface.
When buying used, especially older models, look at known weak spots such as unstable filter chips. You will find a checklist in buying a used synthesizer. Slight drift in older oscillators is even valued by some players as lively character, as long as tuning stays playable.
Frequently asked questions
Which waveform is best for beginners?
The sawtooth, because it offers the most harmonics and responds most clearly to the filter.
Do I need two oscillators?
Not strictly, but two detuned oscillators give a full sound with little effort. Unison or a sub-oscillator can do something similar.
Why does my sound change after the attack?
Most likely a filter envelope or an LFO is at work. Check the modulation amount at the filter.
Does FM or wavetable sound "better" than subtractive?
No, they sound different. Subtractive is the easiest entry point; FM and wavetable open other sound worlds but need more learning time.
Practice checklist
- Choose a sawtooth wave, filter wide open, envelopes neutral.
- Slowly turn the cutoff down and listen to the highs disappear.
- Raise resonance carefully and sweep the cutoff again.
- Add a filter envelope with short decay and medium amount: bass.
- Lengthen attack and release and detune a second oscillator slightly: pad.
- Route an LFO to pitch with a small depth: vibrato.
