I do not believe that "amplitude modulation" is the real problem with wind turbine noise. I suspect that the wind industry do not believe it either, but are simply using it as a smokescreen to cover the real problem which is low frequency noise (in the 1 Hz to 20 Hz band, which covers the blade passing frequency and harmonics).
It is difficult to find reliable independent measured sound power data at frequencies below 20 Hz, but I have found some (attached), appended to a Canadian planning application, for a REpower MM92 - 2 MW wind turbine at Chemin d'Ablis, France.
The plot on page 14 (above) shows how the sound power level increases as the frequency decreases towards 2 Hz; this is the lower frequency limit of the Norsonic 110 sound level meter used for the measurements. The blade passing frequency of the turbine is 0.75 Hz maximum.
Note that the usual plot on page 25 of the noise power level from 20 Hz upwards is, following standard practice, A-weighted. The sound power level at 20 Hz is the same on both plots, from which it follows that they are both A-weighted. Removal of the A-weighting would reveal very high sound power levels at the blade passing frequency and its harmonics.”MM92 acoustic report