In audio-rate FM the carrier and modulator are both audible, so the side frequencies are the spectrum itself

Chowning’s reframing of frequency modulation is what makes it a synthesis technique rather than a radio one. In radio FM the modulating signal is sub-audio and the side frequencies are an unwanted by-product to be minimised. In audio FM both the carrier frequency c and the modulating frequency m lie in the audio band, so the side frequencies that appear above and below the carrier — spaced at integer multiples of m (c±m, c±2m, …) — are not noise but the audible partials that constitute the timbre. This single shift is why two sine oscillators, one modulating the frequency of the other, can generate a rich harmonic (or inharmonic) spectrum directly, with no filtering or partial-by-partial summing. The common misconception, imported from radio, is that FM ‘distorts’ or smears a tone; in audio FM the sidebands ARE the intended sound.

Examples

Two Strudel/SC sine oscillators: one at c=200 Hz whose frequency is modulated by a 200 Hz sinusoid produces audible partials at 200, 400, 600 Hz … — you hear a harmonic tone, not a warbling 200 Hz sine, because m is itself audio-rate.

Assessment

Explain why a 5 Hz vibrato LFO on an oscillator does NOT create an FM spectrum of discrete audible sidebands, whereas a 200 Hz modulator does. What property of m is decisive?