pub fn fft_convolve(a: &[f64], b: &[f64]) -> Vec<f64>
Convolve two real signals using FFT.
Returns the linear convolution a * b of length a.len() + b.len() - 1.
a * b
a.len() + b.len() - 1