Modulating partial frequencies and amplitudes with noise spreads each sine into a band, letting additive synthesis approximate noisy timbres

Pure additive synthesis struggles with inherently noisy instruments (drums, flutes, cymbals) because a noise-like spectrum needs a huge number of overtones. A practical workaround is to modulate the frequency and amplitude of the partials with noise signals: a sine wave is a single impulse in the frequency domain, and noise-modulating it spreads that impulse into a broader band of frequencies. Applying this per-partial replaces an impractically large partial count with a smaller set of noise-broadened partials, giving the breathy or metallic energy those instruments need. Additive synthesis is otherwise at its best on quasi-periodic (pitched, stable) sounds.

Examples

Nord Modular: patch a noise source into the FM and AM inputs of slave sine oscillators so each partial jitters in pitch and level, broadening the flute’s blow-noise band. Contrast with an un-modulated additive patch, which stays glassy and pure.

Assessment

Explain why a flute or cymbal is hard to render with pure sine partials, and describe how noise-modulating the partials’ frequency and amplitude helps — what does it do to a single sine in the frequency domain?