Next: 3D FFT, Previous: Multi-1D FFT, Up: Complex FFT Routines
The routines documented here compute the two-dimensional discrete
Fourier transforms (DFT) of a two-dimensional array of complex numbers
in either single or double precision arithmetic. The 2D DFT is
computed using a highly-efficient FFT algorithm.
There are two sets
of interfaces available: simple drivers and expert drivers. The simple
drivers perform in-place transforms on data held contiguously in
memory using a fixed scaling factor; these are simpler to use and are
sufficient for many problems. The expert drivers offer greater
flexibility by including a number of additional arguments. These allow
you to control: the scaling factor applied; whether the result should
be output to a separate array; the increments used in storing
successive elements in each dimension (for both input and output); and
the facility to not perform a final transposition. This final facility
is useful for those cases where a forward and backward transform are
to be applied with some data manipulations in between; here two whole
transpositions can be saved.