Sound Design & SynthesisLesson 1 of 5

Oscillators and Waveforms

The four classic waveforms and the harmonics in each, plus how detune and unison turn one thin oscillator into a wide, rich sound.

8 min read

Every sound in a subtractive synth starts at the oscillator. It produces a repeating waveform at the pitch you play, and the shape of that waveform decides which harmonics are available for the rest of the synth to shape. You cannot filter in harmonics that are not there, so choosing the waveform is the first real sound design decision.

Harmonics in one paragraph

A pitched sound is a stack of sine waves: the fundamental at the note's frequency, plus harmonics at whole-number multiples of it. Play A2 at 110 Hz and the harmonics sit at 220, 330, 440, 550 Hz and so on. Which harmonics are present, and how loud each one is, is what makes a saw sound different from a flute at the same pitch.

The four classic waveforms

WaveformHarmonics presentLevel of harmonic nSound
SineFundamental only—Pure, round, sub
TriangleOdd only (1, 3, 5…)1/n² (falls fast)Soft, flute-like
SquareOdd only (1, 3, 5…)1/nHollow, woody, clarinet-like
SawtoothAll (1, 2, 3, 4…)1/nBright, buzzy, brassy

Sine

A sine has no harmonics at all. It is the right choice for sub bass, where you want weight with nothing to clash with other instruments, and for building other tones additively. On small speakers a pure low sine can almost disappear, because there are no upper harmonics to imply the pitch.

Triangle

A triangle has the same odd harmonics as a square, but they fall away much faster. It sounds like a slightly brighter sine. It works for soft leads, gentle sub layers and flute-like tones.

Square and pulse

A square has odd harmonics only, at fairly strong levels, which gives the hollow "woody" character of chiptune and clarinet sounds. A pulse wave is a square with an uneven duty cycle. As the width moves away from 50%, even harmonics appear and the sound gets thinner and nasal. Slowly modulating the width, called pulse width modulation or PWM, produces a classic chorused string sound.

Sawtooth

The saw contains every harmonic, which makes it the most useful starting point for subtractive synthesis: there is plenty to filter away. Strings, brass, supersaws, reese basses and most leads start from a saw.

Detune: two oscillators are wider than one

Two oscillators at exactly the same pitch just make a louder single tone. Detune one by a few cents (hundredths of a semitone) and they drift in and out of phase, producing a slow beating that sounds thicker and more alive.

  • 3–8 cents: gentle thickening, good for basses and leads that need to stay focused.
  • 10–20 cents: obvious chorus, good for pads and big leads.
  • More than ~25 cents: starts to sound out of tune.

Oscillators can also be offset by whole intervals. A second saw one octave up adds brightness; one a fifth up (+7 semitones) gives the hollow "power chord" lead often heard in trance and old rave music.

Unison: many voices per note

Unison stacks several copies of an oscillator per note, each detuned slightly and often panned across the stereo field. Five to nine detuned saws make the classic supersaw.

SettingWhat it doesTypical value
VoicesCopies per note3–9
DetuneSpread of pitch between copies10–30% (synth-specific scale)
Spread / widthStereo panning of copies50–100%
BlendLevel of the side voices vs the centre30–70%

Unison costs clarity and low-end focus. Wide, detuned voices partially cancel when summed to mono, which makes them a bad idea for bass. Keep bass oscillators mono and in tune, and save unison for pads and leads above roughly 150 Hz.

Noise

Most synths also have a noise generator. It contains all frequencies at once and has no pitch. A little noise mixed under a pad adds breath; filtered noise with a short envelope makes hats, snares and risers.

Choosing quickly

  • Need weight and nothing else? Sine.
  • Soft and mellow? Triangle.
  • Hollow, retro or reedy? Square or pulse.
  • Bright, rich, or about to be filtered? Saw.

Next, in filters, you will cut those harmonics back into shape.

Sources

  1. Gordon Reid (1999). Synth Secrets, Part 1: What's In A Sound?. Sound On Sound.
  2. Joe Wolfe. What is a Sound Spectrum?. UNSW Sydney.
  3. Adam Szabo (2010). How to Emulate the Super Saw. KTH Royal Institute of Technology.