A package for computing the FFT with arbitrary floating point numbers
Author JuliaApproximation
11 Stars
Updated Last
9 Months Ago
Started In
July 2022


Build Status codecov

GenericFFT.jl implements a Fast Fourier Transform for generic floating point number types.

The transforms provided have the right computational complexity. However, for the time being, the implementations are crude and far from optimal. Please consider contributing improvements to the package and filing issues for missing functionality.


Installation is straightforward:

pkg> add GenericFFT

julia> using GenericFFT

Usage for high-precision FFTs

The main reason for using GenericFFT is high-precision calculations. For example:

julia> using GenericFFT

julia> fft(rand(Complex{BigFloat}, 2))
2-element Vector{Complex{BigFloat}}:
 0.8071607526060331187983248443648586158893950448440777116281652091029932491471374 + 1.058204007570364569492040922226041865648762106924785198005758849420721686004251im
 0.3195699335469630499276014344115859560577992018210584550701583748039853943955188 + 0.196737316420669631800810230623687615407691727320510522950346182385847637522683im

Other packages provide high-precision floating point numbers, including QuadMath.jl and DoubleFloats.jl.

julia> using GenericFFT, DoubleFloats

julia> fft(rand(Double64, 2))
2-element Vector{Complex{Double64}}:
 0.4026739024263829 + 0.0im
 0.3969515892883767 + 0.0im


The code in this package was developed in the FastTransforms.jl package by Mikael Slevinsky. The code was moved to a separate package in July 2022 from this file.