Gwyddion – Free SPM (AFM, SNOM/NSOM, STM, MFM, …) data analysis software

simpleFFT (HEAD)

simpleFFT — Simple FFT algorithm

Functions

void gwy_fft_simple ()
void gwy_fft_window ()
void gwy_fft_window_data_field ()

Includes

#include <libprocess/gwyprocess.h>

Description

The simple one-dimensional FFT algorithm gwy_fft_simple() used to be employed as a fallback by other functions when a better implementation (FFTW3) was not available. Since version 2.49 it just calls the corresponding FFTW routine.

Generally, you should either use high-level Gwyddion functions such as gwy_data_field_2dfft_raw() or, if they are insufficient, FFTW routines directly.

Up to version 2.7 simpleFFT required certain tranform sizes, mostly powers of 2. Since 2.8 it can handle arbitrary tranform sizes, although sizes with large prime factors can be quite slow (still O(n*log(n)) though).

Functions

gwy_fft_simple ()

void
gwy_fft_simple (GwyTransformDirection dir,
                gint n,
                gint istride,
                const gdouble *in_re,
                const gdouble *in_im,
                gint ostride,
                gdouble *out_re,
                gdouble *out_im);

Performs a DFT algorithm.

This is a low-level function that used to be employed by other FFT functions when no better backend was available. Since version 2.49 it just calls the corresponding FFTW routine.

Strides are distances between samples in input and output arrays. Use 1 for normal `dense' arrays. To use gwy_fft_simple() with interleaved arrays, that is with alternating real and imaginary data, call it with istride =2, in_re =complex_array , in_im =complex_array +1 (and similarly for output arrays).

The output is symmetrically normalized by square root of n for both transform directions. By performing forward and then backward transform, you will obtain the original array (up to rounding errors).

Parameters

dir

Transformation direction.

 

n

Number of data points. Note only certain transform sizes are implemented. If gwy_fft_simple() is the current FFT backend, then gwy_fft_find_nice_size() can provide accepted transform sizes. If gwy_fft_simple() is not the current FFT backend, you should not use it.

 

istride

Input data stride.

 

in_re

Real part of input data.

 

in_im

Imaginary part of input data.

 

ostride

Output data stride.

 

out_re

Real part of output data.

 

out_im

Imaginary part of output data.

 

gwy_fft_window ()

void
gwy_fft_window (gint n,
                gdouble *data,
                GwyWindowingType windowing);

Multiplies data by given window.

Parameters

n

Number of data values.

 

data

Data values.

 

windowing

Method used for windowing.

 

gwy_fft_window_data_field ()

void
gwy_fft_window_data_field (GwyDataField *dfield,
                           GwyOrientation orientation,
                           GwyWindowingType windowing);

Performs windowing of a data field in given direction.

Parameters

dfield

A data field.

 

orientation

Windowing orientation (the same as corresponding FFT orientation).

 

windowing

The windowing type to use.

 

See Also

inttrans

-- high-level integral transform functions

© David Nečas and Petr Klapetek

Home Download News Features Screenshots Documentation Communicate Participate Resources Publications Applications Site Map

Valid XHTML 1.0 Valid CSS