How to correctly apply the discrete inverse Fourier transform?

21 Views Asked by At

I'm just getting to know the Fourier transform. I computed $64$ values of the periodic sequence $[1, 2, 4, 8, 16, 32, 29, 23, 11, 22, 9, 18, 1, 2, ...]$ with period length of $12$. Using a machine, I computed

x = [1, 2, 4, 8, 16, 32, 29, 23, 11, 22, 9, 18, 1, 2, ...]
y = fft(x)

Looking at $y$, I see the following complex numbers.

[ 890.          +0.j        ,  -44.38255743  -5.25243373j,
  -48.01118625 -11.76210086j,  -56.39223493 -22.07948054j,
  -78.85632437 -45.70630911j, -237.45489019-204.72085508j, ...]

Computing the inverse, I find

array([ 1.-9.99200722e-16j,  2.+2.77555756e-16j,  4.-1.06143818e-15j,
        8.+1.52655666e-15j, 16.-6.66133815e-16j, 32.+1.27675648e-15j,
       29.-6.03896362e-16j, 23.+1.13797860e-15j, 11.+6.58557145e-16j,
       22.+2.62402416e-16j,  9.+1.44243956e-15j, 18.+4.08756964e-16j,
        1.-1.44328993e-15j,  2.-2.72004641e-15j, ...]

So I can see my list is back in the real parts, but that's not quite an inverse. What am I missing here?