FIR filter coefficients

[Racia]

Hi all,

I'm trying to include a few FIR filters into my program. I used the Tfilter website to design the filters however, it output the code but not the filter coefficients like what the FIR examples in arduino. Did i missed out any steps?

what i want:

/* .fir file parameters which generated these coefficients
100 1 2 16 32768 44100
0 1000 2000 
 1.0000  0.0000 
 1.0000 70.0000 
*/
	(short)0x0015,
	(short)0x0011,
	(short)0x0017,
	(short)0x001E,
	(short)0x0025,
	(short)0x002C,
	(short)0x0032,
	(short)0x0037,
	(short)0x003B,
	(short)0x003C,
	(short)0x003B,
	(short)0x0036,
	(short)0x002E,
	(short)0x0021,
	(short)0x000F,
	(short)0xFFF9,
	(short)0xFFDE,
	(short)0xFFBF,
	(short)0xFF9C,
	(short)0xFF76,
	(short)0xFF4F,
	(short)0xFF27,
	(short)0xFF02,
	(short)0xFEDF,
	(short)0xFEC2,
	(short)0xFEAC,
	(short)0xFEA0,
	(short)0xFE9F,
	(short)0xFEAC,
	(short)0xFEC8,
	(short)0xFEF4,
	(short)0xFF31,
	(short)0xFF7F,
	(short)0xFFDF,
	(short)0x0050,
	(short)0x00D0,
	(short)0x015E,
	(short)0x01F8,
	(short)0x029A,
	(short)0x0341,
	(short)0x03EA,
	(short)0x0491,
	(short)0x0532,
	(short)0x05CA,
	(short)0x0653,
	(short)0x06CC,
	(short)0x0731,
	(short)0x077F,
	(short)0x07B4,
	(short)0x07CF,
	(short)0x07CF,
	(short)0x07B4,
	(short)0x077F,
	(short)0x0731,
	(short)0x06CC,
	(short)0x0653,
	(short)0x05CA,
	(short)0x0532,
	(short)0x0491,
	(short)0x03EA,
	(short)0x0341,
	(short)0x029A,
	(short)0x01F8,
	(short)0x015E,
	(short)0x00D0,
	(short)0x0050,
	(short)0xFFDF,
	(short)0xFF7F,
	(short)0xFF31,
	(short)0xFEF4,
	(short)0xFEC8,
	(short)0xFEAC,
	(short)0xFE9F,
	(short)0xFEA0,
	(short)0xFEAC,
	(short)0xFEC2,
	(short)0xFEDF,
	(short)0xFF02,
	(short)0xFF27,
	(short)0xFF4F,
	(short)0xFF76,
	(short)0xFF9C,
	(short)0xFFBF,
	(short)0xFFDE,
	(short)0xFFF9,
	(short)0x000F,
	(short)0x0021,
	(short)0x002E,
	(short)0x0036,
	(short)0x003B,
	(short)0x003C,
	(short)0x003B,
	(short)0x0037,
	(short)0x0032,
	(short)0x002C,
	(short)0x0025,
	(short)0x001E,
	(short)0x0017,
	(short)0x0011,
	(short)0x0015,

and this is what i get from the website:

[CODE]/*
	This is a very simple C++ implementation of a floating point FIR filter.
 */

template <typename FloatType, int num_coeffs, const FloatType* coeffs>
class FirFilter {
public:
	
  FirFilter():
      current_index_(0) {
    for(int i = 0; i < num_coeffs; ++i)
      history_[i] = 0.0;
  }
	
  void put(FloatType value) {
    history_[current_index_++] = value;
    if(current_index_ == num_coeffs)
      current_index_ = 0;
  }

  FloatType get() {
    FloatType output = 0.0;
    int index = current_index_;
    for(int i = 0; i < num_coeffs; ++i) {
      if(index != 0) {
        --index;
      } else {
        index = num_coeffs - 1;
      }
      output += history_[index] * coeffs[i];
    }
    return output;
  }	

private:
	FloatType history_[num_coeffs];
	int current_index_;
};

/* This is a test program that shows how to use the FirFilter class. */

#include <iostream>
#include <assert.h>
#include <math.h>

#define FILTER1_LENGTH 4

float filter1_coeffs[FILTER1_LENGTH] = {
	0.1,
	0.2,
	0.3,
	0.4
};

bool approx_equal(float a, float b) {
	return fabs(a-b) < 1e-6;
}

int main() {
	FirFilter<float, FILTER1_LENGTH, filter1_coeffs> filter1;
	
	filter1.put(10.0);
	assert(approx_equal(filter1.get(), 1));
	filter1.put(10.0);
	assert(approx_equal(filter1.get(), 3));
	filter1.put(10.0);
	assert(approx_equal(filter1.get(), 6));
	filter1.put(10.0);
	assert(approx_equal(filter1.get(), 10));
	
	filter1.put(0.0);
	assert(approx_equal(filter1.get(), 9));
	filter1.put(0.0);
	assert(approx_equal(filter1.get(), 7));
	filter1.put(0.0);
	assert(approx_equal(filter1.get(), 4));
	filter1.put(0.0);
	assert(approx_equal(filter1.get(), 0));
	
	std::cout << "Tests OK." << std::endl;
	
	return 0;
}

[/CODE]

 

[WMXZ]

Hi all,

I'm trying to include a few FIR filters into my program. I used the Tfilter website to design the filters however, it output the code but not the filter coefficients like what the FIR examples in arduino. Did i missed out any steps?

Please tell us first what you wanted to do, Low-pass/bandpass , bass-band/ stop-band frequencies, ripple etc, i.e. all the parameters you put into TFilter website. With the information you provided I could not reproduce your data. (you could have presented the complete *.h file generated by TFilter)

Note also the sample C++ code is NOT a generated C++ code of your filter design, it is ONLY a sample /template of how you could convert the provided C-code into C++. Note also that there are notational differences between C and C++ version (e.g. filter_taps vs filter_coeffs)