be.tarsos.dsp.util.fft

Class FFT

java.lang.Object

be.tarsos.dsp.util.fft.FFT

public final class FFT
extends java.lang.Object

Wrapper for calling a hopefully Fast Fourier transform. Makes it easy to switch FFT algorithm with minimal overhead. Support for window functions is also present.

Author:

Joren Six

Constructor Summary

Constructors

Constructor and Description
FFT(int size)
FFT(int size, WindowFunction windowFunction) Create a new fft of the specified size.

Method Summary

Modifier and Type	Method and Description
void	backwardsTransform(float[] data) Computes inverse DFT.
double	binToHz(int binIndex, float sampleRate)
void	complexForwardTransform(float[] data)
void	forwardTransform(float[] data) Computes forward DFT.
void	modulus(float[] data, float[] amplitudes) Calculates the the modulus for each element in data and stores the result in amplitudes.
float	modulus(float[] data, int index) Returns the modulus of the element at index bufferCount.
void	multiply(float[] data, float[] other) Multiplies to arrays containing imaginary numbers.
void	powerAndPhaseFromFFT(float[] data, float[] power, float[] phase) Returns magnitude (or power) and phase for the FFT transformed data.
void	powerPhaseFFT(float[] data, float[] power, float[] phase) Computes an FFT and converts the results to polar coordinates (power and phase).
void	powerPhaseFFTBeatRootOnset(float[] data, float[] power, float[] phase)
int	size()

Methods inherited from class java.lang.Object

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

Constructor Detail

FFT

public FFT(int size)

FFT

public FFT(int size,
           WindowFunction windowFunction)

Create a new fft of the specified size. Apply the specified window on the samples before a forward transform. arning: the window is not applied in reverse when a backwards transform is requested.

Parameters:

size - The size of the fft.

windowFunction - Apply the specified window on the samples before a forward transform. arning: the window is not applied in reverse when a backwards transform is requested.

Method Detail

forwardTransform

public void forwardTransform(float[] data)

Computes forward DFT.

Parameters:

data - data to transform.

complexForwardTransform

public void complexForwardTransform(float[] data)

backwardsTransform

public void backwardsTransform(float[] data)

Computes inverse DFT. Warning, does not reverse the window function.

Parameters:

data - data to transform

binToHz

public double binToHz(int binIndex,
                      float sampleRate)

size

public int size()

modulus

public float modulus(float[] data,
                     int index)

Returns the modulus of the element at index bufferCount. The modulus, magnitude or absolute value is (a²+b²) ^ 0.5 with a being the real part and b the imaginary part of a complex number.

Parameters:

data - The FFT transformed data.

index - The index of the element.

Returns:

The modulus, magnitude or absolute value of the element at index bufferCount

modulus

public void modulus(float[] data,
                    float[] amplitudes)

Calculates the the modulus for each element in data and stores the result in amplitudes.

Parameters:

data - The input data.

amplitudes - The output modulus info or amplitude.

powerPhaseFFT

public void powerPhaseFFT(float[] data,
                          float[] power,
                          float[] phase)

Computes an FFT and converts the results to polar coordinates (power and phase). Both the power and phase arrays must be the same length, data should be double the length.

Parameters:

data - The input audio signal.

power - The power (modulus) of the data.

phase - The phase of the data

powerAndPhaseFromFFT

public void powerAndPhaseFromFFT(float[] data,
                                 float[] power,
                                 float[] phase)

Returns magnitude (or power) and phase for the FFT transformed data.

Parameters:

data - The FFT transformed data.

power - The array where the magnitudes or powers are going to be stored. It is half the length of data (FFT size).

phase - The array where the phases are going to be stored. It is half the length of data (FFT size).

powerPhaseFFTBeatRootOnset

public void powerPhaseFFTBeatRootOnset(float[] data,
                                       float[] power,
                                       float[] phase)

multiply

public void multiply(float[] data,
                     float[] other)

Multiplies to arrays containing imaginary numbers. The data in the first argument is modified! The real part is stored at 2*i, the imaginary part 2*i+i

Parameters:

data - The array with imaginary numbers that is modified.

other - The array with imaginary numbers that is not modified. Data and other need to be the same length.