Wind-turbine sound is amplitude-modulated: its level rises and falls at the blade-passing rhythm, the swish. An annual average records only the mean, and so cannot see the modulation at all. This synthesiser lets you hear what the mean conceals — the modulated and steady versions below are held at an identical mean level, so the only thing that changes when you switch between them is the fluctuation itself.
Model and assumptions. The carrier is broadband noise shaped to the mid-frequency band where the swish is heard (a gentle low-pass around 1 kHz over a 120 Hz high-pass); it is amplitude-modulated by a sinusoid at the blade-passing rate. Depth is defined as the peak-to-trough level difference; a value of 5 dB is a typical audible swish (Hansen & Hansen, 2020). The modulated signal is normalised so that the time-mean of its squared envelope equals that of the steady signal — the two therefore share an identical Leq, which is the point of the comparison and was verified in Node.js before publication. This is a synthetic illustration, generated in the browser, not a field recording; real turbine sound also carries tonal and lower-frequency content this demonstration omits.