ComplexFFT

class ComplexFFT

This class represents a tool for Fast Fourier Transformation

Inheritance:

ComplexFFT

Public Classes
enum WindowFunc: windowing functions which can be used.

Public Fields
typedef Complex <Float> ComplexSample: complex type used by the FFT.
typedef vector <ComplexSample> ComplexWave: vector used by the FFT to store a single "window".
typedef double Float: floating point type used by FFT

Public Methods
size_t bits(): returns the window size in bits
static size_t bitsFor(size_t samples): calculates the number of bits which is needed to represent the given number of samples
ComplexFFT(const size_t samples, const WindowFunc winFunc = NONE): creates the tool for the FFT
virtual void fft(ComplexWave& wave): Here is the real work-horse
virtual void invfft(ComplexWave& wave): performs the inverse FFT on the given vector
virtual void printOn(ostream& os): prints the FFT parameter on the given output stream
size_t samples(): returns the window size in samples
virtual void set(const size_t samples, const WindowFunc winFunc): specifies FFT parameters
WindowFunc windowFunc(): returns the used windowing function
virtual ~ComplexFFT(): destroy the object.

Public
enum WindowFunc: windowing functions which can be used.

Protected Fields
size_t vBits: Number of bits in one "frame".
size_t vSamples: Number of samples in one "frame".
WindowFunc vWindowFunc: Windowing function used by the FFT.

Protected Methods
virtual void init(const size_t samples, const WindowFunc winFunc): initializer for FFT routines

Private Fields
Float* CosineTable: Table of values used for faster computation.
Float* SineTable: Table of values used for faster computation.
Float* WinTable: Table of values used for faster computation.

Private Methods
void reorder(ComplexWave& wave): put the samples back in order after the FFT scrambles them
const unsigned int reverse(unsigned int val): bit reverser for unsigned ints
void scale(ComplexWave& wave): scale sampled values
void unwindow(ComplexWave& wave): undo the window function to restore full magnitude
void window(ComplexWave& wave): apply some windowing function to the samples

Documentation

This class represents a tool for Fast Fourier Transformation.

enum WindowFunc

windowing functions which can be used.

typedef double Float

floating point type used by FFT

typedef Complex <Float> ComplexSample

complex type used by the FFT.

typedef vector <ComplexSample> ComplexWave

vector used by the FFT to store a single "window".

ComplexFFT(const size_t samples, const WindowFunc winFunc = NONE)

creates the tool for the FFT.

Parameters:: samples - specifies the used window size
winFunc - specifies the windowing function

virtual ~ComplexFFT()

destroy the object.

virtual void set(const size_t samples, const WindowFunc winFunc)

specifies FFT parameters. This method can be used to change the window size or the windowing function after the tool was created.

Parameters:: samples - specifies the used window size
winFunc - specifies the windowing function

virtual void fft(ComplexWave& wave)

Here is the real work-horse. It's a generic FFT, so nothing is lost or approximated. The samples in wave should be in order, and they will be decimated when fft() returns.

Parameters:: wave - vector containing samples (time domain of a signal)

virtual void invfft(ComplexWave& wave)

performs the inverse FFT on the given vector. The values in wave should be in order, and they will be decimated when invfft() returns.

Parameters:: vector - containg the frequency domain of a signal

size_t bits()

returns the window size in bits. samples = pow(2, bits).

Returns:: window size in bits

size_t samples()

returns the window size in samples. samples = pow(2, bits).

Returns:: window size in samples.

WindowFunc windowFunc()

returns the used windowing function.

Returns:: kind of windowing function

static size_t bitsFor(size_t samples)

calculates the number of bits which is needed to represent the given number of samples. bits = Max { x | pow(2,x) <= samples }

Returns:: number of bits needed
Parameters:: samples - number of samples

virtual void printOn(ostream& os)

prints the FFT parameter on the given output stream.

Parameters:: os - output stream

virtual void init(const size_t samples, const WindowFunc winFunc)

initializer for FFT routines. Currently only sets up tables. Generates scaled lookup tables for sin() and cos(). Fills a table for a window function; possible windows: Hamming, Hanning, None

Throws:: Exception , if memory allocation fails
Parameters:: samples - specifies the used window size
winFunc - specifies the windowing function

size_t vBits