Fir1 Class Reference

#include <Fir1.h>

Collaboration diagram for Fir1:

Public Member Functions
template<unsigned nTaps>
	Fir1 (const double(&_coefficients)[nTaps])
	Fir1 (std::vector< double > _coefficients)
	Fir1 (double *coefficients, unsigned number_of_taps)
	Fir1 (const char *coeffFile, unsigned number_of_taps=0)
	Fir1 (unsigned number_of_taps)
	~Fir1 ()
double	filter (double input)
void	lms_update (double error)
void	setLearningRate (double _mu)
double	getLearningRate ()
void	reset ()
void	zeroCoeff ()
void	getCoeff (double *coeff_data, unsigned number_of_taps) const
std::vector< double >	getCoeffVector () const
unsigned	getTaps ()
double	getTapInputPower ()

Detailed Description

Finite impulse response filter. The precision is double. It takes as an input a file with coefficients or an double array.

Constructor & Destructor Documentation

Fir1() [1/5]

template<unsigned nTaps>

Fir1::Fir1 ( const double(&)  _coefficients[nTaps] )

inline

Coefficients as a const double array. Because the array is const the number of taps is identical to the length of the array.

Parameters

_coefficients

A const double array with the impulse response.

Fir1() [2/5]

Fir1::Fir1 ( std::vector< double >  _coefficients )

inline

Coefficients as a C++ vector

Parameters

_coefficients

is a Vector of doubles.

Fir1() [3/5]

Fir1::Fir1	(	double *	_coefficients,
		unsigned	number_of_taps
	)

Coefficients as a (non-constant-) double array where the length needs to be specified.

Parameters

coefficients	Coefficients as double array.
number_of_taps	Number of taps (needs to match the number of coefficients

Initialise the filter from coefficients, assigned as array

Parameters

_coefficients	coeffs to be assigned
number_of_taps	nr of taps (array length)

Fir1() [4/5]

Fir1::Fir1	(	const char *	coeffFile,
		unsigned	number_of_taps = 0
	)

Coefficients as a text file (for example from Python) The number of taps is automatically detected when the taps are kept zero.

Parameters

coeffFile	Patht to textfile where every line contains one coefficient
number_of_taps	Number of taps (0 = autodetect)

Initialise the filter coefficients with a text file where each line is has one coefficient

Parameters

coeffFile	text file for initialisation
number_of_taps	nr of fiter taps (order)

Fir1() [5/5]

Fir1::Fir1

(

unsigned

number_of_taps

)

Inits all coefficients and the buffer to zero

This is useful for adaptive filters where we start with zero valued coefficients.

Initalise the filter with empty coefficients

Parameters

number_of_taps

nr of taps (filter order)

~Fir1()

Fir1::~Fir1

(

)

Releases the coefficients and buffer.

Filter destructor

Member Function Documentation

filter()

double Fir1::filter ( double  input )

inline

The actual filter function operation: it receives one sample and returns one sample.

Parameters

input

The input sample.

getCoeff()

void Fir1::getCoeff	(	double *	coeff_data,
		unsigned	number_of_taps
	)		const

Copies the current filter coefficients into a provided array. Useful after an adaptive filter has been trained to query the result of its training.

Parameters

coeff_data	target where coefficients are copied
number_of_taps	number of doubles to be copied

Exceptions

std::out_of_range

number_of_taps is less the actual number of taps.

Copies a coeffs to coeff_data array

Parameters

coeff_data
number_of_taps

getCoeffVector()

std::vector<double> Fir1::getCoeffVector ( ) const

inline

Returns the coefficients as a vector

getLearningRate()

double Fir1::getLearningRate ( )

inline

Getting the learning rate for the adaptive filter.

getTapInputPower()

double Fir1::getTapInputPower ( )

inline

Returns the power of the of the buffer content:

sum_k buffer[k]^2 which is needed to implement a normalised LMS algorithm.

getTaps()

unsigned Fir1::getTaps ( )

inline

Returns the number of taps.

lms_update()

void Fir1::lms_update ( double  error )

inline

LMS adaptive filter weight update:

Every filter coefficient is updated with: w_k(n+1) = w_k(n) + learning_rate * buffer_k(n) * error(n)

Parameters

error

Is the term error(n), the error which adjusts the FIR conefficients.

reset()

void Fir1::reset

(

)

Resets the buffer (but not the coefficients)

sets all tap outputs to zero

setLearningRate()

void Fir1::setLearningRate ( double  _mu )

inline

Setting the learning rate for the adaptive filter.

Parameters

_mu	The learning rate (i.e. rate of the change by the error signal)

zeroCoeff()

void Fir1::zeroCoeff

(

)

Sets all coefficients to zero

Sets all coeffs to zero

---

The documentation for this class was generated from the following files:

• src/Fir1.h
• src/Fir1.cpp