Additive synthesis builds complex timbres by summing many sine partials with individual amplitude envelopes

Additive synthesis constructs sounds from the bottom up by summing many simple waveforms — typically sine waves (partials) — each with its own frequency, amplitude, and phase, and usually its own time-varying amplitude envelope. It is the direct inverse of Fourier analysis: analysis decomposes a periodic sound into sinusoidal components, additive synthesis reassembles a sound from them. Because any periodic waveform can in principle be built this way, the method is highly general and can approximate almost any tone given enough partials — but cost scales with partial count, and a rich acoustic instrument may need 50–100+ partials each individually enveloped. A common misconception is that ‘more partials = better’; partials must be precisely tuned and enveloped over time. Too few partials, or a bank of flat-amplitude sines, yields a thin, static, organ-like sound rather than a recognisable evolving timbre — a real piano’s high partials decay faster than its low ones, and capturing that spectral evolution is what makes the sound convincing.

Examples

A Hammond organ does additive synthesis mechanically: each drawbar adds a harmonic partial at a fixed amplitude. A sawtooth = sine partials at integer multiples of the fundamental, amplitude 1/n. In SuperCollider: Mix.fill(15, {|i| SinOsc.ar(440 * (i+1), mul: 0.5/(i+1)) }) sums many partials; truncating to a few partials sounds noticeably different from the full sawtooth.

Assessment

Given a sawtooth, list the first four partials (frequencies and relative amplitudes) and explain why truncating to four sounds different from the full wave. State the relationship between additive synthesis and Fourier analysis, and explain why too few or flat-amplitude partials produce an ‘organ-like’ rather than a convincing instrument tone.