FourierCoefficient Class Reference

This class contains a set of fourier coefficients as well as methods transform them from and to a period of function values. More...

#include <FourierCoefficient.h>

List of all members.

Public Types

enum { numOfCoeffs = 80 }

Public Member Functions

FourierCoefficient ()

Constructor.

~FourierCoefficient ()

Destructor.

long fourierSynth (JointData::JointID joint, unsigned long time, int cMax, double scalingFactor=1.0)

Synthesizes the value of the original function for a given joint using the fourier coefficients.

long fourierSynth (JointData *jointData, unsigned long time, int cMax)

Like the other fourierSynth, a change to the joint parameter and no return value.

bool calcLUT (JointData::JointID joint, int cMax)

Calculate LUT for the function values so they don't have to be computed every time the synth.

bool calcAllLUTs (int cMax)

Calculate all LUTs.

bool calculate (long *functionValues, JointData::JointID joint)

Calculate coefficients from the function values over one period.

bool calcPhiAndR (int numOfC=numOfCoeffs)

Calculates phi and r for the first numOfC coefficients only Calculates the coefficients in the exponential form, i.e.

bool calcReAndIm (int numOfC=numOfCoeffs)

Calculates the coefficients in the vector form, i.e.

bool merge (FourierCoefficient *fc0, FourierCoefficient *fc1, double w1, FourierCoefficient *fc2, double w2, FourierCoefficient *fc3, double w3, int numOfC)

Takes 1+3=4 sets of Fourier coefficients and averages them according to the weights.

bool merge (FourierCoefficient *fc1, double w1, FourierCoefficient *fc2, double w2, int numOfC)

Similar to merge(many parameters) but only merges two sets!

double phaseAlign (JointData::JointID joint)

Function shifts all phases so that the 1 Hz frequency of the joint is a sine function starting at t=0 (i.e.

Public Attributes

int lengthOfPeriod

length of period these coefficients are designed for

double real [JointData::numOfJoint][numOfCoeffs]

real and imaginary parts of the coefficients

double imaginary [JointData::numOfJoint][numOfCoeffs]

double phi [JointData::numOfJoint][numOfCoeffs]

alternative representation: length r and phase phi of the coefficients

double r [JointData::numOfJoint][numOfCoeffs]

double functionLUT [JointData::numOfJoint][numOfCoeffs]

LUT that stores the function calculated from the coeffs.

bool useLookUp [numOfCoeffs]

Detailed Description

This class contains a set of fourier coefficients as well as methods transform them from and to a period of function values.

Definition at line 20 of file FourierCoefficient.h.

Member Enumeration Documentation

anonymous enum

Enumeration values:

numOfCoeffs

Definition at line 23 of file FourierCoefficient.h.

Constructor & Destructor Documentation

FourierCoefficient::FourierCoefficient ( )

Constructor.

init all coefficients to zero
Definition at line 12 of file FourierCoefficient.cpp.
References functionLUT, imaginary, numOfCoeffs, phi, r, real, and useLookUp.

FourierCoefficient::~FourierCoefficient ( )

Destructor.

Definition at line 30 of file FourierCoefficient.cpp.

Member Function Documentation

long FourierCoefficient::fourierSynth ( JointData::JointID joint,

unsigned long time,

int cMax,

double scalingFactor = 1.0

)

Synthesizes the value of the original function for a given joint using the fourier coefficients.
A scaling factor is added. This allows the amplitude of the curve to be changed while keeping the origin/offset constant. This factor is also used for inverting the direction of the function (in essence playing the function backward if the factor is smaller than 0)
Parameters:

joint The joint for which the synthesis is calculated.

time Time (in the format returned by "getSystemTime"). This time is used to generate the curves as a function of that time.

cMax Number of coefficients used. Small numbers result in a low pass behavior, numbers larger than numOfCoeffs/2 result in oversampling effects. (This is due to the fact that a continous Fourier (back) transformation is calculated from descrete FT coefficients.)

scalingFactor is used to vary the amplitude of the wave without touching the offset. 1.0 = 100%, values greater than 1 possible. Negative numbers will be interpreted as running the function backwards (negative time line)

Returns:
The synthesized function value.

Definition at line 36 of file FourierCoefficient.cpp.
References imaginary, numOfCoeffs, pi2, and real.
Referenced by calcLUT().

long FourierCoefficient::fourierSynth ( JointData * jointData,

unsigned long time,

int cMax

)

Like the other fourierSynth, a change to the joint parameter and no return value.
In contrast to the above, a reference to the entire joint data array is passed in order to minimize function calls. All joint values are computed in one go.
Also scalingfactor is no longer used. don't need it, really.

Parameters:

jointData Array of joint datas to be calculated.

time This time is used to generate the curves as a function of that time.

cMax Number of coefficients used.

Returns:
nothing

Definition at line 74 of file FourierCoefficient.cpp.
References JointData::data, imaginary, numOfCoeffs, pi2, and real.

bool FourierCoefficient::calcLUT ( JointData::JointID joint,

int cMax

)

Calculate LUT for the function values so they don't have to be computed every time the synth.
is used
Parameters:

joint The joint that is described by the function

cMax Number of coefficients used.

Returns:
true if successful, false if failed

Definition at line 115 of file FourierCoefficient.cpp.
References fourierSynth(), functionLUT, lengthOfPeriod, numOfCoeffs, real, and useLookUp.
Referenced by calcAllLUTs().