MATLAB Functions

Inverse multidimensional fast Fourier transform

Y = ifftn(X)
Y = ifftn(X,siz)

Description

Y = ifftn(X)

performs the N-dimensional inverse fast Fourier transform. The result Y is the same size as X.

Y = ifftn(X,siz)

pads X with zeros, or truncates X, to create a multidimensional array of size siz before performing the inverse transform. The size of the result Y is siz.

Remarks

For any X, ifftn(fftn(X)) equals X within roundoff error. If X is real, ifftn(fftn(X)) may have small imaginary parts.

Algorithm

ifftn(X) is equivalent to

Y = X;
for p = 1:length(size(X))
    Y = ifft(Y,[],p);
end

This computes in-place the one-dimensional inverse fast Fourier transform along each dimension of X. The time required to compute ifftn(X) depends strongly on the number of prime factors of the dimensions of X. It is fastest when all of the dimensions are powers of 2.