[webaudio] AnalyserNode output arrays

Hi,

Could we please explicitly describe the format of the data that goes into
getFloatFrequencyData() and friends?

I'd like to figure out the pitch of input sound. The current recommendation
document (18 Sept. 2018) walks us through a 4 step technical process
(1.8.6. FFT Windowing and Smoothing over Time), ending with "This array,
Y[k], is copied to the output array for getFloatFrequencyData()". It would
be great if we could describe what's in this final array so that we could
use it without understanding the implementation details. I'm hoping to
learn how Fourier transforms work some day and understand these details,
but we should be able to leverage AnalyserNodes without this knowledge.

Cheers,
Eric