Fast Fourier Transform Object Data Structure. More...

#include <dspl.h>

Data Fields
complex_t *	w

complex_t *	t0

complex_t *	t1

complex_t	w32 [32]

complex_t	w64 [64]

complex_t	w128 [128]

complex_t	w256 [256]

complex_t	w512 [512]

complex_t *	w1024

complex_t *	w2048

complex_t *	w4096

int	n

Detailed Description

Fast Fourier Transform Object Data Structure.

The structure stores pointers to twiddle factors and arrays of intermediate data of the fast Fourier transform algorithm.

The libdspl-2.0 library uses an FFT algorithm for composite size.

Parameters

n	The size of the FFT vector for which memory is allocated in the structure arrays. The parameter `n` must be equal to an integer power of two (radix 2).
w	Pointer to the vector of twiddle factors. The size of the vector is `[n x 1]`. The memory must be allocated and an array of twiddle factors must be filled with the fft_create function.
t0	Pointer to the vector of intermediate results of the FFT algorithm. The size of the vector is `[n x 1]`. Memory must be allocated by fft_create function.
t1	Pointer to the vector of intermediate results. The size of the vector is `[n x 1]`. The memory must be allocated with the fft_create function.
w32	Static twiddle factors vector for 32-points FFT.
w64	Static twiddle factors vector for 32-points FFT.
w128	Static twiddle factors vector for 32-points FFT.
w256	Static twiddle factors vector for 32-points FFT.
w512	Static twiddle factors vector for 32-points FFT.
w1024	Dynamic twiddle factors vector for 32-points FFT.
w2048	Dynamic twiddle factors vector for 32-points FFT.
w4096	Dynamic twiddle factors vector for 32-points FFT. The structure is calculated with the fft_create function once before using the FFT algorithm. A pointer to an object of this structure may be reused when calling FFT functions. Before exiting the program, dedicated memory for twiddle factors and arrays of intermediate data must be cleared by the fft_free function.

For example:

fft_t pfft = {0};     // Structure fft_t and clear all fields
int n = 64;           // FFT size
 
int err;
 
// Create and fill FFT structure for 64-points FFT
err = fft_create(&pfft, n);
 
// FFT calculation here
// FFT calculation here one more
// ...
 
// Clear fft structure
fft_free(&pfft);

Note: It is important to note that if the object fft_t was created for the FFT size equal ton, it can only be used for FFT of size n.

It’s also worth noting that the FFT functions independently control the size, and independently allocate the memory of the FFT object, if necessary. So if you call any function using the fft_t structure with filled data for the FFT length k for calculating the FFT of lengthn, then the structure arrays will be automatically recreated for the length n.