Signal processessing

Classes for signal processessing. More...

Classes
struct	fir_t
	FIR filter structure. More...
struct	sos_t
	second order section structure More...
struct	iir_t
	IIR filter structure. More...
Functions
void	free_iir_filter (iir_t *)
iir_t *	init_iir_filter (double c[][6], double, int)
double	iir_filter (iir_t *, double)
	filter data with an iir filter
void	set_iir_initial_values (iir_t *, double)
double	fir_filter (fir_t *, double)
fir_t *	init_fir_filter (double *, int)
void	free_fir_filter (fir_t *)

---

Detailed Description

Classes for signal processessing.

---

Function Documentation

double fir_filter	(	fir_t *	pf,
		double	val
	)

filter a data value with an fir filter

Parameters:

	pf	a pointer to an fir filter
	val	the data value to be filtered

Returns:

the filtered data

void free_fir_filter

(

fir_t *

f

)

free an fir filter

Parameters:

f

a pointer to an fir filter

void free_iir_filter

(

iir_t *

filt

)

free memory associated with an iir filter

Parameters:

filt

a pointer to an iir filter

double iir_filter	(	iir_t *	f,
		double	xin
	)

filter data with an iir filter

Parameters:

	f	a pointer to an iir filter
	xin	the data value to be filtered

Returns:

the filtered value

fir_t* init_fir_filter	(	double *	b,
		int	ord
	)

create an fir filter

Parameters:

	b	an array of filter coefficients
	ord	the filter order

iir_t* init_iir_filter	(	double	coef[][6],
		double	gain,
		int	nsects
	)

create an iir filter

Parameters:

	coef	a two dimensional array of second order section coefficients
	gain	the filter gain associated with the second order section representation
	nsects	the number of second order sections

usage

call init_iir_filter with an nsects x 6 array of doubles as filter coefficients. Assuming that you have created a filter in MATLAB and obtained transfer function coefficients, the following m-code will generate the appropriate declaration:

---------------------------------------------------------------
    function writesos(b,a,filtername,filename)

    if (nargin < 3)
        disp('usage: writesos(b,a,filtername,filename)');
        return
    end
    if (nargin < 4) 
        fid=1;
    else
        fid=fopen(filename,'w');
    end

    [sos g] = tf2sos(b,a);
    [nsects ncol]=size(sos);

    fprintf(fid,['iir_t *init_' filtername '()\n']);
    fprintf(fid,'{ double coeff[][6]= {\n');

    for i=1:nsects
        fprintf(fid,'                   ');
        fprintf(fid,'    {%.15g, %.15g, %.15g, %.15g, %.15g, %.15g}',sos(i,:));
        if (i ~= nsects)
            fprintf(fid,',');
        end
        fprintf(fid,'\n');
    end

    fprintf(fid,'                     };\n');

    fprintf(fid,'  double gain=%.15g;\n',g);
    fprintf(fid,'  iir_t *f;\n\n');
    fprintf(fid,'  f=init_iir_filter(coeff, gain, %d);\n\n',nsects);
    fprintf(fid,'  return f;\n');
    fprintf(fid,'}\n');
    if (fid ~= 1)
        fclose(fid);
    end
---------------------------------------------------------------

Example

Example: 5th order low-pass butterworth filter with break frequency at 10% of the Nyquist frequency

Save the m-code above into a file called writesos.m. Then execute these commands in matlab:

[b a] = butter(5,0.1);
writesos(b,a,'myLowPassFilter')

Matlab outputs the following text in the command window:

iir_t *init_myLowPassFilter()
{ double coeff[][6]= {
{1, 0.998743271944812, 0, 1, -0.726542528005698, 0},
{1, 2.00203542892064, 1.00203701235769, 1, -1.52169042607194, 0.599999999999708},
{1, 1.99922129913454, 0.999222880053167, 1, -1.7363101655347, 0.825664548620679}
};
double gain=5.97957803700113e-05;
iir_t *f;

f=init_iir_filter(coeff, gain, 3);

return f;
}

Call this function to initialize your filter.

Adding a 4th parameter to the writesos call allows output to a text file instead of the matlab console. writesos(b,a,'myLowPassFilter','butter.c') for example, write the init call to the file butter.c.

void set_iir_initial_values	(	iir_t *	f,
		double	x
	)

set the output of an iir filter to an initial value

Parameters:

	f	a pointer to an iir filter
	x	the initial value