Package fftManager

Class FastFFT

java.lang.Object

fftManager.FastFFT

public class FastFFT
extends Object

FFT Wrapper which uses the edu.emory.mathcs.jtransforms.fft transforms library for the actual FFT calculations.
These are simple wrappers to use the transforms library with standard PAMGUARD transform classes.

Author:

Doug Gillespie

Constructor Summary

Constructors

Constructor	Description
FastFFT()

Method Summary

Modifier and Type	Method	Description
void	fft(Complex[] x)	In place fft of complex data.
void	fft(Complex[][] x)	In lace FFT of a 2D complex array.
void	fft(ComplexArray x)	In place complex fft of complex data.
void	ifft(Complex[] x, int m)	Inverse FFT for Complex data.
void	ifft(ComplexArray x, int n)	Inverse FFT for Complex data.
void	ifft(ComplexArray x, int n, boolean scale)	Inverse FFT for Complex data.
static int	log2(int num)
static int	nextBinaryExp(int sourceNumber)	Finds the next highest binary exponential of the input integer.
static int	nextBinaryExp(int startPower, int sourceNumber)	Finds the next highest binary exponential of the input integer.
double[]	realInverse(ComplexArray x)	Inverse transform of what was real data.
ComplexArray	rfft(double[] x, int n)	FFT of real data.
Complex[]	rfft(double[] x, Complex[] y, int m)	Fast FFT function for real data.
ComplexArray	rfftFull(double[] x, int n)

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Details

FastFFT

public FastFFT()

Method Details

rfft

public Complex[] rfft(double[] x,
 Complex[] y,
 int m)

Fast FFT function for real data.

Parameters:

x - real data array

y - preallocated Complex array for output data (can be null)

m - log2 of the FFT length (sorry !)

Returns:

Complex FFT data.

rfft

public ComplexArray rfft(double[] x,
 int n)

FFT of real data. Will return only the first half of the FFT, since second is simply complex conjugate of the first.

Parameters:

x - waveform (will be padded or truncated to length n)

n - length of FFT

Returns:

Complex spectrum

rfftFull

public ComplexArray rfftFull(double[] x,
 int n)

fft

public void fft(Complex[] x)

In place fft of complex data.

Parameters:

x - complex array

fft

public void fft(ComplexArray x)

In place complex fft of complex data.

Parameters:

x - Complex data array.

fft

public void fft(Complex[][] x)

In lace FFT of a 2D complex array. Will use the multithreading abilities of the JTransofrms library.

Parameters:

x -

ifft

public void ifft(Complex[] x,
 int m)

Inverse FFT for Complex data.
I FFT is performed 'in place' so data are overwritten

Parameters:

x - Complex Data

m - log2 of the FFT length (sorry !)

ifft

public void ifft(ComplexArray x,
 int n)

Inverse FFT for Complex data.
I FFT is performed 'in place' so data are overwritten

Parameters:

x - ComplexArray - the input data.

n - the FFT length

realInverse

public double[] realInverse(ComplexArray x)

Inverse transform of what was real data. Will automatically assume a second half complex conj of the first half. This can be used to transform back data used in cross correlations rather than trying to fill in the second half of the conjugate, which is not really possible with the missing middle point of the FFT.

Parameters:

x - Complex input data, which originally came from real.

Returns:

real data array 2x the length of the Complex input (i.e. same length as double array within complex input).

ifft

public void ifft(ComplexArray x,
 int n,
 boolean scale)

Inverse FFT for Complex data.
I FFT is performed 'in place' so data are overwritten

Parameters:

x - ComplexArray - the input data

n - the FFT length

scale - - true for scaling to be performed

nextBinaryExp

public static int nextBinaryExp(int sourceNumber)

Finds the next highest binary exponential of the input integer. If the input is itself a binary exponential, then the result is itself. E.g. given 7 returns 8, 8 returns 8, 9 returns 16. Notes has limit of 2^100. Matlab calls this function nextpow2; it's also akin to frexp in C.

Parameters:

sourceNumber -

Returns:

The next highest 2^ of the input, unless input is itself a binary exponential.

nextBinaryExp

public static int nextBinaryExp(int startPower,
 int sourceNumber)

Finds the next highest binary exponential of the input integer. If the input is itself a binary exponential, then the result is itself. E.g. given 7 returns 8, 8 returns 8, 9 returns 16. Notes has limit of 2^100. Matlab calls this function nextpow2; it's also akin to frexp in C.

Parameters:

startPower - power of 2 to start at (e,g. 1 will return a minimum of 2)

sourceNumber -

Returns:

The next highest 2^ of the input, unless input is itself a binary exponential.

log2

public static int log2(int num)

Parameters:

FFTlength -

Returns:

log2 of FFTlength, -1 if not natural power of 2