Rfftrad4 - Analog Devices VISUALDSP++ 3.5 Manual

DSP Run-Time Library Reference

rfftrad4

N point real input FFT
Synopsis
#include <filter.h>
void rfftrad4_fr16(in[], t[], out[], w[], wst, n, block_exponent,
const fract16 in[];
complex_fract16 t[];
complex_fract16 out[];
const complex_fract16 w[];
int wst;
int n;
int *block_exponent;
int scale method;
Description
This function transforms the time domain real input signal sequence to
the frequency domain by using the radix-4 Fast Fourier Transform. The
rfftrad4_fr16
part of the input equals zero, which in turn eliminates half of the multipli-
cations in the butterfly.
The size of the input array
working buffer
The function does not impose any special memory alignment require-
ments on the arrays. However, benefits in run-time performance will be
realized if the output array is allocated on an address boundary that is
multiple of twice the FFT size. If the input data can be overwritten, then
optimum memory usage can be achieved by specifying the input array as
either the output array or as the temporary array provided that the mem-
ory size of the input array is at least
temporary array will also result in increased run-time performance.
4-96
scale_method)
/* input/output sequence
/* temporary working buffer
/* working buffer
/* twiddle sequence
/* twiddle factor stride
/* number of FFT points
/* block exponent of output data */
/* scaling method desired 0-none,
function takes advantage of the fact that the imaginary
, the output array
in
is , where represents the number of points in the FFT.
t n n
VisualDSP++ 3.5 C/C++ Compiler and Library Manual
1-static, 2-dynamic
, and the temporary
out
. Specifying the input array as the
2*n
for ADSP-219x DSPs
*/
*/
*/
*/
*/
*/
*/