KaliVeda
Toolkit for HIC analysis
KVDigitalFilter.cpp
1 /*
2  $Header: /home/CVS/luigi/FClasses/KVDigitalFilter.cxx,v 1.6 2013-05-23 19:40:20 garfield Exp $
3  $Log: KVDigitalFilter.cxx,v $
4  Revision 1.6 2013-05-23 19:40:20 garfield
5  KVDigitalFilter: added methods to apply a FIR filter (no b coefficients)
6  added methods to read/write FIR coefficients for MatLab
7 
8  Revision 1.5 2013-05-23 11:29:09 garfield
9  Added method to draw the step response of the filter.
10 
11  Revision 1.4 2010-01-26 09:48:12 bini
12  Ottimizzazioni in KVSignal
13  Patch per interpolazione cubica in KVDigitalFilterWizard
14 
15  Revision 1.3 2009/10/12 12:57:31 bardelli
16  Aggiunto il metodo static KVDigitalFilter BuildInverse(KVDigitalFilter *filter)
17 
18  Revision 1.2 2006/05/05 13:10:26 bardelli
19  Aggiunto Header e Log in testa a tutti i files
20 
21 */
22 
23 
24 //=======================
25 // Author: G.Pasquali & L.Bardelli
26 // @(#) 17 11 2003
27 //=======================
29 // //
30 // KVDigitalFilter //
31 // //
32 // Class for digital filtering //
33 // //
34 // No limits on number of a[] and b[] coeffs. //
35 // //
36 // b[0] = 0. (this is different from the sources of KVSignal!) //
37 // //
38 // INTERNAL COMPUTATIONS DONE USING "DspGuide notation". //
39 // //
40 // Example: //
41 // //
42 // // Low pass RC with tau=2 us, coeff. for a 10ns sampling period //
43 //KVDigitalFilter rc=KVDigitalFilter::BuildRCLowPass( 2, 10 ); //
44 // //
45 // // Apply it to a signal: intermediate results are "double" //
46 // // type (i.e. 64 bit). //
47 // KVSignal s; //
48 // rc.ApplyTo( &s ); //
49 // //
50 // // Apply it to generic data: intermediate results are "double" //
51 // // type (i.e. 64 bit). //
52 // void ApplyTo(double *data, const int N, int reverse=0) const; //
53 // void ApplyTo(float *data, const int N, int reverse=0) const; //
54 // void ApplyTo(int *data, const int N, int reverse=0) const; //
55 // //
56 // // Combine two filters //
57 //KVDigitalFilter rc=KVDigitalFilter::BuildRCLowPass( 2, 10 ); //
58 //KVDigitalFilter cr=KVDigitalFilter::BuildRCHighPass( 2, 10 ); //
59 //KVDigitalFilter total=KVDigitalFilter::CombineStages( &rc, &cr ); //
60 // //
61 // // Combine Many series (parallel=0) //
62 //KVDigitalFilter total4=KVDigitalFilter::CombineStagesMany(&cr,&rc,&rc,&rc,&rc);//
63 // //
64 // // Write Coeffs //
65 // total.PrintCoeffs() //
66 // Coefficients valid with tau_clk=10.000 ns //
67 // a[ 0]= 4.975083e-03 - //
68 // a[ 1]= -4.975083e-03 b[ 1]= 1.990025e+00 //
69 // a[ 2]= 0.000000e+00 b[ 2]= -9.900498e-01 //
70 // TOTAL: 4 coefficients (a+b). //
71 // //
72 // - Various "Build*" methods... //
73 // - ComputeGainDC(); // gain at 0 frequency //
74 // - ComputeGainNy(); // gain at Nyquist frequency (0.5) //
75 // //
76 // //
78 #include "KVDigitalFilter.h"
79 #define DUEPI 6.28318530717958623
80 #define PI (DUEPI/2.)
81 
82 #define LINE() // printf("Now %s line %d...\n",__PRETTY_FUNCTION__, __LINE__);
83 
84 ClassImp(KVDigitalFilter);
85 
86 //============================================
87 
89 
90 KVDigitalFilter KVDigitalFilter::BuildRCLowPass(const double& tau_usec, const double& tau_clk)
91 {
92 
93  double f = 1. / (DUEPI * tau_usec * 1000.) * tau_clk;
94  double x = exp(-DUEPI * f);
95  KVDigitalFilter filter(tau_clk);
96  filter.Alloc(2);
97  filter.a[0] = 1 - x;
98  filter.b[1] = x;
99 
100  return filter;
101 }
102 
103 //============================================
104 
106 
107 KVDigitalFilter KVDigitalFilter::BuildRCLowPassDeconv(const double& tau_usec, const double& tau_clk)
108 {
109 
110  double f = 1. / (DUEPI * tau_usec * 1000.) * tau_clk;
111  double x = exp(-DUEPI * f);
112  KVDigitalFilter filter(tau_clk);
113  filter.Alloc(2);
114  filter.a[0] = 1 / (1 - x);
115  filter.a[1] = -x / (1 - x);
116 
117  return filter;
118 }
119 
120 //============================================
121 
123 
124 KVDigitalFilter KVDigitalFilter::BuildRCHighPassWithPZ(const double& tau_usec, const double& preamp_decay_usec, const double& tau_clk)
125 {
126  KVDigitalFilter highpass = KVDigitalFilter::BuildRCHighPass(tau_usec, tau_clk);
127  double ycoef[] = {0};
128  double xcoef[] = {tau_usec / (preamp_decay_usec + tau_usec)};
129  KVDigitalFilter polez(tau_clk, 1, xcoef, ycoef);
130  return KVDigitalFilter::CombineStages(&highpass, &polez, 1);
131 }
132 
133 //============================================
134 
136 
137 KVDigitalFilter KVDigitalFilter::BuildRCHighPass(const double& tau_usec, const double& tau_clk)
138 {
139 
140  double f = 1. / (DUEPI * tau_usec * 1000.) * tau_clk;
141  double x = exp(-DUEPI * f);
142  KVDigitalFilter filter(tau_clk);
143  filter.Alloc(2);
144  filter.a[0] = (1 + x) / 2.;
145  filter.a[1] = -(1 + x) / 2.;
146  filter.b[1] = x;
147 
148  return filter;
149 }
150 
151 //============================================
152 
154 
155 KVDigitalFilter KVDigitalFilter::BuildChebyshev(const double& freq_cutoff_mhz, int is_highpass,
156  const double& percent_ripple, int npoles,
157  const double& tau_clk)
158 {
159  KVDigitalFilter filter(tau_clk);
160  filter.Alloc(30);
161  ComputeChebyshevCoeffs(freq_cutoff_mhz * (tau_clk / 1e3),
162  is_highpass, percent_ripple, npoles, filter.a, filter.b);
163  filter.Compress();
164  return filter;
165 }
166 
167 //============================================
168 
169 
171 
172 void KVDigitalFilter::ComputeChebyshevCoeffs_serv(const double& fc, const double& lh,
173  const double& pr, const double& np,
174  int p,
175  double* a0, double* a1, double* a2,
176  double* b1, double* b2)
177 {
178  double rp, ip, es, vx, kx, t, w, m, d, k = 0;
179  double x0, x1, x2, y1, y2;
180 
181  // printf("%e %e\n",cos(PI/(np*2.)+(p-1)*PI/np),PI/(np*2.)+(p-1)*PI/np);
182 
183  /* calculate the pole location on the unit circle */
184  rp = -cos(PI / (np * 2.) + (p - 1) * PI / np);
185  ip = sin(PI / (np * 2.) + (p - 1) * PI / np);
186 
188 
189  /* warp from a circle to an ellipse */
190 
191  if (pr == 0) goto pr_zero;
192 
193  // printf("pr=%e\n",pr);
194 
195  es = sqrt((100. / (100. - pr)) * (100. / (100. - pr)) - 1);
196  vx = (1 / np) * log((1. / es) + sqrt(1. / es / es + 1.));
197  kx = (1 / np) * log((1. / es) + sqrt(1. / es / es - 1.));
198  kx = (exp(kx) + exp(-kx)) / 2.;
199  rp = rp * ((exp(vx) - exp(-vx)) / 2.) / kx;
200  ip = ip * ((exp(vx) + exp(-vx)) / 2.) / kx;
201 
202  /* s-domain to z-domain conversion */
204 
205 pr_zero:
206  t = 2 * tan(1 / 2.);
207  w = 2.*PI * fc;
208  m = rp * rp + ip * ip;
209  d = 4 - 4 * rp * t + m * t * t;
210 
211  x0 = t * t / d;
212  x1 = 2.*t * t / d;
213  x2 = t * t / d;
214 
215  y1 = (8 - 2.*m * t * t) / d;
216  y2 = (-4. - 4.*rp * t - m * t * t) / d;
217 
218  // printf("t=%e w=%e m=%e d=%e\n",t,w,m,d);
219  // printf("x0=%e x1=%e x2=%e y1=%e y2=%e\n",x0,x1,x2,y1,y2);
220  /* LP to LP or LP to HP transform */
221  if (lh == 1) k = -cos(w / 2. + 1 / 2.) / cos(w / 2. - 1 / 2.);
222  if (lh == 0) k = sin(-w / 2. + 1 / 2.) / sin(w / 2. + 1 / 2.);
223  d = 1 + y1 * k - y2 * k * k ;
224  *a0 = (x0 - x1 * k + x2 * k * k) / d;
225  *a1 = (-2.*x0 * k + x1 + x1 * k * k - 2.*x2 * k) / d;
226  *a2 = (x0 * k * k - x1 * k + x2) / d;
227  *b1 = (2.*k + y1 + y1 * k * k - 2.*y2 * k) / d;
228  *b2 = (-k * k - y1 * k + y2) / d;
229  if (lh == 1) *a1 = -(*a1);
230  if (lh == 1) *b1 = -(*b1);
231 
232 
233  return;
234 }
235 
236 
237 //============================================
238 
239 
241 
242 void KVDigitalFilter::ComputeChebyshevCoeffs(const double& freq_cutoff,
243  int is_highpass, const double& percent_ripple, int npoles,
244  double* a, double* b)
245 {
246  double fc = freq_cutoff;
247  double lh = is_highpass;
248  double pr = percent_ripple;
249  double np = npoles;
250 
251  double a0, a1, a2, b1, b2, gain;
252  double ta[22], tb[22];
253 
254 #define CHECK(A,MIN,MAX) if(A<MIN || A>MAX) printf("ERROR in %s: %s=%e! (ok is %e..%e)\n",__PRETTY_FUNCTION__,#A,(double)A,(double)MIN,(double)MAX);
255  CHECK(freq_cutoff, 0, 0.5);
256  CHECK(lh, 0, 1);
257  CHECK(pr, 0, 29);
258  CHECK(np, 2, 20);
259 #undef CHECK
260 
261  for (int i = 0; i < 22; i++) {
262  a[i] = 0.;
263  b[i] = 0.;
264  }
265  a[2] = 1.;
266  b[2] = 1.;
267 
268  for (int p = 1; p <= np / 2; p++) {
269  ComputeChebyshevCoeffs_serv(fc, lh, pr, np, p, &a0, &a1, &a2, &b1, &b2);
270  for (int i = 0; i < 22; i++) {
271  ta[i] = a[i];
272  tb[i] = b[i];
273  }
274  for (int i = 2; i < 22; i++) {
275  a[i] = a0 * ta[i] + a1 * ta[i - 1] + a2 * ta[i - 2];
276  b[i] = tb[i] - b1 * tb[i - 1] - b2 * tb[i - 2];
277  }
278  }
279  b[2] = 0.;
280  for (int i = 0; i < 20; i++) {
281  a[i] = a[i + 2];
282  b[i] = -b[i + 2];
283  }
284  double sa = 0;
285  double sb = 0;
286 
287  /***** ORIGINALE DI Gab.*********/
288 #define AAA
289 #ifdef AAA
290  for (int i = 0; i < 20; i++) {
291  if (lh == 0) sa = sa + a[i];
292  if (lh == 0) sb = sb + b[i];
293  if (lh == 1) sa = sa + a[i] * pow(-1., i);
294  if (lh == 1) sb = sb + b[i] * pow(-1., i);
295  }
296 #else
297  /**********************/
298  if (lh == 0) {
299  for (int i = 0; i < 20; i++) {
300  sa = sa + a[i];
301  sb = sb + b[i];
302  }
303  }//end of lh==0
304  else {
305  int sign = 1;
306  for (int i = 0; i < 20; i++) {
307  sa = sa + a[i] * sign;
308  sb = sb + b[i] * sign;
309  sign *= -1;
310  }
311  }
312 #endif
314  gain = sa / (1. - sb);
315  for (int i = 0; i < 20; i++) a[i] = a[i] / gain;
317  return;
318 }
319 
320 //============================================
321 
323 
324 KVDigitalFilter::KVDigitalFilter(const double& tau)
325 {
326 
327  a = b = NULL;
328  tau_clk = tau;
329  Ncoeff = 0;
330 
331 }
332 
333 //============================================
334 
336 
337 KVDigitalFilter::KVDigitalFilter(const double& tau, int N, const double* Xcoeffs, const double* Ycoeffs)
338 {
339 
340  a = b = NULL;
341  tau_clk = tau;
342  Ncoeff = 0;
343  Alloc(N);
344  memcpy(a, Xcoeffs, sizeof(double)*N);
345  memcpy(b, Ycoeffs, sizeof(double)*N);
346 
347 }
348 
349 //=============================================
350 
352 
353 KVDigitalFilter::~KVDigitalFilter()
354 {
355 
356  if (a != NULL) {
357  delete [] a;
358  delete [] b;
359  }
360 
361 }
362 
363 //=============================================
364 
367 
368 void KVDigitalFilter::Alloc(const int N)
369 {
370 
371  // printf("a=%p, N=%d, Ncoeff=%d\n", a, N, Ncoeff);
372  Ncoeff = N;
373  if (a != NULL) {
374  delete [] a;
375  delete [] b;
376  a = NULL;
377  b = NULL;
378  };
379  if (N > 0) {
380  a = new double [N];
381  b = new double [N];
382  Azzera();
383  }
384 
385 }
386 
387 //=============================================
388 
390 
391 void KVDigitalFilter::Azzera()
392 {
393 
394  memset(a, 0, sizeof(double)*Ncoeff);
395  memset(b, 0, sizeof(double)*Ncoeff);
396 
397 }
398 
399 //=============================================
400 
402 
403 KVDigitalFilter::KVDigitalFilter(const KVDigitalFilter& orig)
404  : KVBase()
405 {
406 
407  Ncoeff = 0;
408  tau_clk = orig.tau_clk;
409  a = b = NULL;
410  if (orig.Ncoeff > 0) {
411  Alloc(orig.Ncoeff);
412  memcpy(a, orig.a, sizeof(double)*Ncoeff);
413  memcpy(b, orig.b, sizeof(double)*Ncoeff);
414  }
415 
416 }
417 
418 //=============================================
419 
421 
422 KVDigitalFilter KVDigitalFilter::operator =(const KVDigitalFilter& orig)
423 {
424 
425  tau_clk = orig.tau_clk;
426  Alloc(orig.Ncoeff);
427  if (orig.Ncoeff > 0) {
428  memcpy(a, orig.a, sizeof(double)*Ncoeff);
429  memcpy(b, orig.b, sizeof(double)*Ncoeff);
430  }
431 
432  return *this;
433 }
434 
435 //=============================================
436 
440 
441 void KVDigitalFilter::ApplyTo(double* datax, const int NSamples, int reverse) const
442 {
443  // Copiato +- da KVSignal.cxx
444  // Diversa la convenzione per a0 b0!
445 
446  std::vector<long double> datay(NSamples);
447  int i = 0, k = 0;
448  switch (reverse) {
449  case 0:// direct
450  for (i = 0; i < Ncoeff; i++) { // primo loop su Npunti
451  datay[i] = a[0] * datax[i];
452  for (k = 0; k < i; k++)
453  datay[i] += a[k + 1] * datax[i - k - 1] + b[k + 1] * datay[i - k - 1];
454  }
455  for (i = Ncoeff; i < NSamples; i++) { //secondo loop. cambia l'indice interno.
456  datay[i] = a[0] * datax[i];
457  for (k = 0; k < Ncoeff - 1; k++)
458  datay[i] += a[k + 1] * datax[i - k - 1] + b[k + 1] * datay[i - k - 1];
459  }
460  break; // end of direct
461  case 1: //reverse: cut & paste from direct and NSamples-1-
462  for (i = 0; i < Ncoeff; i++) { // primo loop su Npunti
463  datay[NSamples - 1 - i] = a[0] * datax[NSamples - 1 - i];
464  for (k = 0; k < i; k++)
465  datay[NSamples - 1 - i] += a[k + 1] * datax[NSamples - 1 - (i - k - 1)]
466  + b[k + 1] * datay[NSamples - 1 - (i - k - 1)];
467  }
468  for (i = Ncoeff; i < NSamples; i++) { //secondo loop. cambia l'indice interno.
469  datay[NSamples - 1 - i] = a[0] * datax[NSamples - 1 - i];
470  for (k = 0; k < Ncoeff - 1; k++)
471  datay[NSamples - 1 - i] += a[k + 1] * datax[NSamples - 1 - (i - k - 1)]
472  + b[k + 1] * datay[NSamples - 1 - (i - k - 1)];
473  }
474  break;
475  case -1: // bidirectional
476  ApplyTo(datax, NSamples, 0);
477  ApplyTo(datax, NSamples, 1);
478  return;
479  default:
480  printf("ERROR in %s: reverse=%d not supported\n", __PRETTY_FUNCTION__, reverse);
481  }// end of reverse switch.
482  /*----------------------------------------------*/
483  // void *memcpy(void *dest, const void *src, size_t n);
484  //memcpy(datax, datay, NSamples * sizeof(double));
485  // can't use memcpy: datax & datay have different type sizes!
486  for (int i = 0; i < NSamples; i++)
487  datax[i] = (double)datay[i];
488 }
489 
490 //=============================================
491 
495 
496 void KVDigitalFilter::ApplyTo(float* datax, const int NSamples, int reverse) const
497 {
498  // Copiato +- da KVSignal.cxx
499  // Diversa la convenzione per a0 b0!
500 
501  std::vector<long double> datay(NSamples);
502  int i = 0, k = 0;
503  switch (reverse) {
504  case 0:// direct
505  for (i = 0; i < Ncoeff; i++) { // primo loop su Npunti
506  datay[i] = a[0] * datax[i];
507  for (k = 0; k < i; k++)
508  datay[i] += a[k + 1] * datax[i - k - 1] + b[k + 1] * datay[i - k - 1];
509  }
510  for (i = Ncoeff; i < NSamples; i++) { //secondo loop. cambia l'indice interno.
511  datay[i] = a[0] * datax[i];
512  for (k = 0; k < Ncoeff - 1; k++)
513  datay[i] += a[k + 1] * datax[i - k - 1] + b[k + 1] * datay[i - k - 1];
514  }
515  break; // end of direct
516  case 1: //reverse: cut & paste from direct and NSamples-1-
517  for (i = 0; i < Ncoeff; i++) { // primo loop su Npunti
518  datay[NSamples - 1 - i] = a[0] * datax[NSamples - 1 - i];
519  for (k = 0; k < i; k++)
520  datay[NSamples - 1 - i] += a[k + 1] * datax[NSamples - 1 - (i - k - 1)]
521  + b[k + 1] * datay[NSamples - 1 - (i - k - 1)];
522  }
523  for (i = Ncoeff; i < NSamples; i++) { //secondo loop. cambia l'indice interno.
524  datay[NSamples - 1 - i] = a[0] * datax[NSamples - 1 - i];
525  for (k = 0; k < Ncoeff - 1; k++)
526  datay[NSamples - 1 - i] += a[k + 1] * datax[NSamples - 1 - (i - k - 1)]
527  + b[k + 1] * datay[NSamples - 1 - (i - k - 1)];
528  }
529  break;
530  case -1: // bidirectional
531  ApplyTo(datax, NSamples, 0);
532  ApplyTo(datax, NSamples, 1);
533  return;
534  default:
535  printf("ERROR in %s: reverse=%d not supported\n", __PRETTY_FUNCTION__, reverse);
536  }// end of reverse switch.
537  /*----------------------------------------------*/
538  // void *memcpy(void *dest, const void *src, size_t n);
539  for (int i = 0; i < NSamples; i++)
540  datax[i] = (float)datay[i];
541 }
542 
543 //=============================================
544 
548 
549 void KVDigitalFilter::ApplyTo(int* datax, const int NSamples, int reverse) const
550 {
551  // Copiato +- da KVSignal.cxx
552  // Diversa la convenzione per a0 b0!
553 
554  std::vector<long double> datay(NSamples);
555  int i = 0, k = 0;
556  switch (reverse) {
557  case 0:// direct
558  for (i = 0; i < Ncoeff; i++) { // primo loop su Npunti
559  datay[i] = a[0] * datax[i];
560  for (k = 0; k < i; k++)
561  datay[i] += a[k + 1] * datax[i - k - 1] + b[k + 1] * datay[i - k - 1];
562  }
563  for (i = Ncoeff; i < NSamples; i++) { //secondo loop. cambia l'indice interno.
564  datay[i] = a[0] * datax[i];
565  for (k = 0; k < Ncoeff - 1; k++)
566  datay[i] += a[k + 1] * datax[i - k - 1] + b[k + 1] * datay[i - k - 1];
567  }
568  break; // end of direct
569  case 1: //reverse: cut & paste from direct and NSamples-1-
570  for (i = 0; i < Ncoeff; i++) { // primo loop su Npunti
571  datay[NSamples - 1 - i] = a[0] * datax[NSamples - 1 - i];
572  for (k = 0; k < i; k++)
573  datay[NSamples - 1 - i] += a[k + 1] * datax[NSamples - 1 - (i - k - 1)]
574  + b[k + 1] * datay[NSamples - 1 - (i - k - 1)];
575  }
576  for (i = Ncoeff; i < NSamples; i++) { //secondo loop. cambia l'indice interno.
577  datay[NSamples - 1 - i] = a[0] * datax[NSamples - 1 - i];
578  for (k = 0; k < Ncoeff - 1; k++)
579  datay[NSamples - 1 - i] += a[k + 1] * datax[NSamples - 1 - (i - k - 1)]
580  + b[k + 1] * datay[NSamples - 1 - (i - k - 1)];
581  }
582  break;
583  case -1: // bidirectional
584  ApplyTo(datax, NSamples, 0);
585  ApplyTo(datax, NSamples, 1);
586  return;
587  default:
588  printf("ERROR in %s: reverse=%d not supported\n", __PRETTY_FUNCTION__, reverse);
589  }// end of reverse switch.
590  /*----------------------------------------------*/
591  // void *memcpy(void *dest, const void *src, size_t n);
592  for (int i = 0; i < NSamples; i++)
593  datax[i] = (int)datay[i];
594 }
595 
596 /**********************************************/
597 
600 
601 void KVDigitalFilter::Compress()
602 {
603  // shorten filter. No deallocation of memory.
604 
605  int M = Ncoeff;
606  // printf("COmpress: %d\n", M);
607  const double zero_level = 1e-20;
608  for (; M >= 1; M--) {
609  if (fabs(a[M - 1]) > zero_level || fabs(b[M - 1]) > zero_level)
610  break;
611  }
613 
614  for (int i = 0; i < M; i++) {
615  if (fabs(a[i]) < zero_level)
616  a[i] = 0;
617  if (fabs(b[i]) < zero_level)
618  b[i] = 0;
619  }
620 
621  b[0] = 0;
622 
623  Ncoeff = M; // le lascio allocate, ma non le usero' nei conti...
624 
625 }
626 
627 
628 
629 /***************************************************/
630 /***************************************************/
631 
635 
636 KVDigitalFilter KVDigitalFilter::CombineStagesMany(const KVDigitalFilter* f1, const KVDigitalFilter* f2,
637  const KVDigitalFilter* f3, const KVDigitalFilter* f4,
638  const KVDigitalFilter* f5, const KVDigitalFilter* f6,
639  const KVDigitalFilter* f7, const KVDigitalFilter* f8,
640  const KVDigitalFilter* f9, const KVDigitalFilter* f10)
641 {
642  // se ne devi combinare + di 1 IN CASCATA!
643  // per fare parallelo usare l'altro...
644  const KVDigitalFilter* filters[10] = {f1, f2, f3, f4, f5, f6, f7, f8, f9, f10};
645  KVDigitalFilter total = BuildUnity(f1->tau_clk);
646  for (int i = 0; i < 10; i++)
647  if (filters[i] != NULL) {
648  total = CombineStages(&total, filters[i]);
649  }
650  total.Compress();
651  return total;
652 }
653 
654 /***************************************************/
655 /***************************************************/
656 
658 
659 KVDigitalFilter KVDigitalFilter::CombineStages(const KVDigitalFilter& f1,
660  const KVDigitalFilter& f2,
661  int parallel)
662 {
663  return CombineStages(&f1, &f2, parallel);
664 }
665 
666 /***************************************************/
667 
669 
670 KVDigitalFilter KVDigitalFilter::CombineStages(const KVDigitalFilter* f1,
671  const KVDigitalFilter* f2,
672  int parallel)
673 {
674  if (fabs(f1->tau_clk - f2->tau_clk) > 1e-4) {
675  printf("ERROR in %s: cannot combine with different taus! %e != %e\n",
676  __PRETTY_FUNCTION__, f1->tau_clk, f2->tau_clk);
677  return KVDigitalFilter(f1->tau_clk);
678  }
679  int Nmax = f1->Ncoeff;
680  int __N = f2->Ncoeff;
681  if (__N > Nmax)
682  Nmax = __N;
683 
684  Nmax *= 2;
685 
686  /******** now N is the max *********/
687  double a1[Nmax], a2[Nmax], a3[2 * Nmax];
688  double b1[Nmax], b2[Nmax], b3[2 * Nmax];
689  for (int i = 0; i < Nmax; i++) {
690  a1[i] = f1->GetXcoeff(i);
691  b1[i] = f1->GetYcoeff(i);
692  a2[i] = f2->GetXcoeff(i);
693  b2[i] = f2->GetYcoeff(i);
695  }
696 
697  for (int i = 0; i < Nmax; i++) {
698  b2[i] = - b2[i];
699  b1[i] = - b1[i];
700  }
701  b1[0] = 1;
702  b2[0] = 1;
703  for (int i = 0; i < 2 * Nmax; i++) {
704  a3[i] = 0;
705  b3[i] = 0;
706  for (int j = 0; j < Nmax; j++) {
707  if (j > i || (i - j) >= Nmax) continue;
708  if (!parallel)
709  a3[i] = a3[i] + a1[j] * a2[i - j];
710  else
711  a3[i] = a3[i] + a1[j] * b2[i - j] + a2[j] * b1[i - j];
712  b3[i] = b3[i] + b1[j] * b2[i - j];
713  }
714  }
715 
716  for (int i = 0; i < Nmax; i++) b3[i] = -b3[i];
717  b3[0] = 0;
718  /*
719  for(int i=0; i< Nmax*2; i++){
720  printf("Combine stages, output: a3[%2d]=%15.8e b3[%2d]=%15.8e\n",
721  i,a3[i],i,b3[i]);
722  }
723  */
724  /********** COPY INTO NEW FILTER ****************/
725  KVDigitalFilter filter(f1->tau_clk, 2 * Nmax, a3, b3);
726  filter.Compress();
727  return filter;
728 
729 }
730 
731 /***********************************************/
732 
736 
737 double KVDigitalFilter::ComputeGainDC()
738 {
739  // Gain of filter at f=0
740  // eq. 33.7 del solito libro
741  double num = 0, den = 0;
742  for (int i = 0; i < Ncoeff; i++) {
743  num += a[i];
744  den += b[i];
745  }
746  return num / (1 - den);
747 }
748 
749 /***********************************************/
750 
754 
755 double KVDigitalFilter::ComputeGainNy()
756 {
757  // Gain of filter at f=Nyquist (0.5)
758  // eq. 33.8 del solito libro
759  double num = 0, den = 0;
760  int sign = 1;
761  for (int i = 0; i < Ncoeff; i++) {
762  num += a[i] * sign;
763  den += b[i] * sign;
764  sign *= -1;
765  }
766  return num / (1 - den);
767 }
768 
769 
771 
772 KVDigitalFilter KVDigitalFilter ::BuildUnity(const double& tau_clk)
773 {
774  double a[2] = {1, 0};
775  double b[2] = {0, 0};
776  KVDigitalFilter filter(tau_clk, 2, a, b);
777  filter.Compress();
778  return filter;
779 }
780 
781 
783 
784 KVDigitalFilter KVDigitalFilter ::BuildIntegrator(const double& tau_clk)
785 {
786  double a[2] = {1, 0};
787  double b[2] = {0, 1};
788  KVDigitalFilter filter(tau_clk, 2, a, b);
789  return filter;
790 }
791 
792 /***********************************************/
793 
795 
796 void KVDigitalFilter::PrintCoeffs() const
797 {
798  printf("------------------------------------------------------\n");
799  printf("Coefficients valid with tau_clk=%.3f ns.\n", tau_clk);
800  int tot = 0;
801  for (int i = 0; i < Ncoeff; i++) {
802  if (i == 0)
803  printf("*** Xcoeff[%2d]= %20.20e -\n", i, a[i]);
804  else
805  printf("*** Xcoeff[%2d]= %20.20e Ycoeff[%2d]= %20.20e\n", i, a[i], i, b[i]);
806  if (fabs(a[i]) > 1e-20) tot++;
807  if (fabs(b[i]) > 1e-20) tot++;
808  }
809  printf("TOTAL: %d coefficients (a+b).\n", tot);
810  printf("(DspGuide : Xcoeff <-> a\n"
811  " Ycoeff <-> b\n"
812  " Oppenheim: Xcoeff <-> b\n"
813  " Ycoeff <-> a)\n"
814  "------------------------------------------------------\n"
815  );
816 }
817 
818 
820 
821 void KVDigitalFilter::PrintCoeffsDSP() const
822 {
823  printf("------------------------------------------------------\n");
824  printf("DSP Coefficients valid with tau_clk=%.3f ns.\n", tau_clk);
825  int tot = 0;
826  for (int i = 0; i < Ncoeff; i++) {
827  if (i == 0)
828  printf("*** Xcoeff[%2d]= 0x%6.6x -\n", i, Double2DSP(a[i]));
829  else
830  printf("*** Xcoeff[%2d]= 0x%6.6x Ycoeff[%2d]= 0x%6.6x\n",
831  i, Double2DSP(a[i]), i, Double2DSP(b[i]));
832  if (fabs(a[i]) > 1e-20) tot++;
833  if (fabs(b[i]) > 1e-20) tot++;
834  }
835  printf("TOTAL: %d coefficients (a+b).\n", tot);
836  printf("(DspGuide : Xcoeff <-> a\n"
837  " Ycoeff <-> b\n"
838  " Oppenheim: Xcoeff <-> b\n"
839  " Ycoeff <-> a)\n"
840  "------------------------------------------------------\n"
841  );
842 }
843 
844 /***********************************************/
845 
847 
848 void KVDigitalFilter::PrintCoeffs_AsC() const
849 {
850  printf("/****** filter valid for %.1f tau_clk *********/\n", tau_clk);
851  printf(" int Ncoeff=%d;\n", Ncoeff);
852  printf(" double Xcoeffs[%d]={\n", Ncoeff);
853  for (int i = 0; i < Ncoeff; i++)
854  printf("\t%20.20e%s\n", a[i], i == Ncoeff - 1 ? "};" : ",");
855  printf(" double Ycoeffs[%d]={\n", Ncoeff);
856  for (int i = 0; i < Ncoeff; i++)
857  printf("\t%20.20e%s\n", b[i], i == Ncoeff - 1 ? "};" : ",");
858  printf(" KVDigitalFilter filter(%d, Xcoeffs, Ycoeffs, %f);\n", Ncoeff, tau_clk);
859  printf("/**********************************************/\n");
860 }
861 
862 /***********************************************/
863 
865 
866 void KVDigitalFilter::Quantize(int nbits, int use_pow2,
867  double* xgain, int* x_out, int* y_out, int* x_scale, int* y_scale)
868 {
869 
870  if (nbits <= 0) return;
871  Compress();
872  double* x = GetXarray();
873  double* y = GetYarray();
874 
875  double Nlevel = pow(2.f, nbits - 1) - 1; // non posso fare con << xe' non funziona se nbits==32 !!
876  // normalizzazione delle X a 1:
877  double factor = fabs(x[0]);
878  // int imax=0;
879  for (int i = 1; i < GetNCoeff(); i++)
880  if (fabs(x[i]) > factor) {
881  factor = fabs(x[i]);
882  // imax=i;
883  }
884  if (factor <= 1e-16) {
885  printf("ERROR in %s: factor=%e?\n", __PRETTY_FUNCTION__, factor);
886  factor = 1;
887  }
888  if (xgain != NULL) {
889  *xgain = factor;
890  }
891  if (x_out != NULL) memset(x_out, 0, GetNCoeff()*sizeof(int));
892  if (y_out != NULL) memset(y_out, 0, GetNCoeff()*sizeof(int));
893  if (x_scale != NULL) memset(x_scale, 0, GetNCoeff()*sizeof(int));
894  if (y_scale != NULL) memset(y_scale, 0, GetNCoeff()*sizeof(int));
895 
896 
898  for (int i = 0; i < GetNCoeff(); i++) {
899  x[i] /= factor;
900 
901 
902  // quantizziamo il coefficiente N-esimo...
903  int N = 0;
904  if (use_pow2 == 1) {
905  N = (int)ceil(log(fabs(x[i])) / log(2.));
906  N = (1 << N);
907  }
908  else
909  N = (int)ceil(fabs(x[i])); // numero intero.
910  if (N == 0) {
911  x[i] *= factor;
912  continue;
913  }
914  x[i] /= N;
915  //-- adesso i bits: adesso y[i] e' un numero tra -1..1
916  int k = (int)rint(x[i] * Nlevel);
917  if (k >= Nlevel) k = (int)(Nlevel - 1.);
918  if (k < -Nlevel) k = (int)(-Nlevel);
919  // if(stampa && (k&0x00FFFFFF)!=0) printf("x[%d] 0x %6.6Xoo with factor %d %s\n",i,k&0x00FFFFFF,N,imax==i?" << MAX" : "");
920  if (x_out != NULL) {
922  x_out[i] = k;
923  }
924  if (x_scale != NULL) {
925  x_scale[i] = N;
926  }
927 
928  x[i] = (double)k / (double)Nlevel;
929  //--- ripristino N
930  x[i] *= N;
931  //--- ripristino factor
932  x[i] *= factor;
933 
934  }
935 
936  for (int i = 1; i < GetNCoeff(); i++) {
937 
938  // printf("Now coeff %d\n",i);
939  // quantizziamo il coefficiente N-esimo...
940 
941  int N = 0;
942  if (use_pow2 == 1) {
943  N = (int)ceil(log(fabs(y[i])) / log(2.));
944  N = (1 << N);
945  }
946  else
947  N = (int)ceil(fabs(y[i])); // numero intero.
948  if (N == 0) {
949  y[i] *= factor;
950  continue;
951  }
952  y[i] /= N;
953  //-- adesso i bits: adesso y[i] e' un numero tra -1..1
954  int k = (int)rint(y[i] * Nlevel);
955  if (k >= Nlevel) k = (int)(Nlevel - 1);
956  if (k < -Nlevel) k = (int)(-Nlevel);
958  if (y_out != NULL)
959  y_out[i] = k;
960  if (y_scale != NULL)
961  y_scale[i] = N;
962  y[i] = (double)k / (double)Nlevel;
963  //--- ripristino N
964  y[i] *= N;
965 
966  }
967 
968 }
969 
970 /*******************************************************/
971 
973 
974 void KVDigitalFilter::Draw(Option_t* option)
975 {
976  int Nbins = 10000;
977 
978  TH1F* h1 = new TH1F("hKVDigitalFilter", "Filter response", Nbins, 0, 1. / GetTauClk() / 2.*1000.);
979  h1->SetStats(kFALSE);
980  h1->GetXaxis()->SetTitle("Frequency (MHz)");
981  printf("NCoeffs: %d\n", GetNCoeff());
982  for (int k = 0; k < Nbins; k++) {
983  double angolo = 3.14159265358979312 * k / (Nbins - 1);
984 
985  double numRE = 0;
986  double numIM = 0;
987  double denRE = 0;
988  double denIM = 0;
989 
990  for (int i = 0; i < GetNCoeff(); i++) {
991  double a_re = cos(angolo * i);
992  double a_im = sin(angolo * i);
993 
994  numRE += GetXcoeff(i) * a_re;
995  numIM += GetXcoeff(i) * a_im;
996  if (i == 0) {
997  denRE += 1;
998  }
999  else {
1000  denRE -= GetYcoeff(i) * a_re;
1001  denIM -= GetYcoeff(i) * a_im;
1002  }
1003  }
1004  /*-------------------------*/
1005  double val = sqrt((numRE * numRE + numIM * numIM) / ((denRE * denRE + denIM * denIM)));
1006  h1->SetBinContent(k + 1, val);
1007  /*-------------------------*/
1008  }
1009  h1->Draw(option);
1010 }
1011 
1012 /**************************************/
1013 //void KVDigitalFilter::DrawResponse(Option_t* option)
1014 //{
1015 // KVSignal s(tau_clk);
1016 // s.SetMaxT(10000); // 10 us moooolto arbitrariamente...
1017 // s.SetData((int)(100/tau_clk), 1);
1018 // ApplyTo(&s);
1019 // s.Draw(option);
1020 //}
1022 //void KVDigitalFilter::DrawStepResponse(Option_t* option)
1023 //{
1024 // KVSignal s;
1025 // s.SetChannelWidth(tau_clk);
1026 // s.SetMaxT(10000); // 10 us moooolto arbitrariamente...
1027 // s.SetData((int)(100/tau_clk), 1);
1028 // s.Integra();
1029 // ApplyTo(&s);
1030 // s.Draw(option);
1031 //}
1032 /**************************************/
1033 
1035 
1036 KVDigitalFilter KVDigitalFilter::BuildInverse(KVDigitalFilter* filter)
1037 {
1038  KVDigitalFilter inv_filter(filter->GetTauClk());
1039  inv_filter.Alloc(filter->GetNCoeff());
1040  inv_filter.SetXcoeff(0, 1. / filter->GetXcoeff(0));
1041  for (int i = 1; i < filter->GetNCoeff(); i ++) {
1042  inv_filter.SetXcoeff(i, - filter->GetYcoeff(i)*inv_filter.GetXcoeff(0));
1043  inv_filter.SetYcoeff(i, - filter->GetXcoeff(i)*inv_filter.GetXcoeff(0));
1044  }
1045  return inv_filter;
1046 }
1047 
1048 
1049 
1050 /********************************************/
1051 
1052 // function to load filter coefficients
1053 
1056 
1057 int KVDigitalFilter::ReadMatlabFIR(char* filecoeff)
1058 {
1059 
1060  // FILE *fin = fopen("notch_coeffs.txt","r");
1061  FILE* fin = fopen(filecoeff, "r");
1062  int i = 0;
1063  int n;
1064  if (fscanf(fin, "%d", &n)) {
1065  ;
1066  }
1067  Alloc(n);
1068  // while(fscanf(fin, "%lg",&b[i])!=EOF){
1069  for (i = 0; i < n; i++) {
1070  if (fscanf(fin, "%lg", &a[i])) {
1071  ;
1072  }
1073  b[i] = 0.0;
1074  printf("%s: i=%d a[%d]=%20.20e\n", __PRETTY_FUNCTION__, i, i, a[i]);
1075  //i++;
1076  }
1077  printf("%s: Read %d coefficients\n", __PRETTY_FUNCTION__, i);
1078  Ncoeff = i ;
1079  fclose(fin);
1080  return Ncoeff;
1081 }
1082 
1083 
1084 /********************************************/
1085 
1086 // function to load filter coefficients
1087 
1089 
1090 int KVDigitalFilter::WriteMatlabFIR(char* filecoeff)
1091 {
1092 
1093  FILE* fout = fopen(filecoeff, "w");
1094  int i = 0;
1095  fprintf(fout, "%d\n", Ncoeff);
1096  for (i = 0; i < Ncoeff; i++) {
1097  fprintf(fout, "%20.20e\n", a[i]);
1098  }
1099  printf("%s: Written %d coefficients\n", __PRETTY_FUNCTION__, i);
1100  fclose(fout);
1101  return Ncoeff;
1102 }
1103 
1104 
1105 
1106 // =========================== ApplyTo per filtri FIR:
1107 // eliminata ricorsivita' per risparmiare tempo:
1108 //=============================================
1109 
1113 
1114 void KVDigitalFilter::FIRApplyTo(float* datax, const int NSamples, int reverse) const
1115 {
1116  // Copiato +- da KVSignal.cxx
1117  // Diversa la convenzione per a0 b0!
1118 
1119  std::vector<long double> datay(NSamples);
1120  int i = 0, k = 0;
1121  switch (reverse) {
1122  case 0:// direct
1123  for (i = 0; i < Ncoeff; i++) { // primo loop su Npunti
1124  datay[i] = a[0] * datax[i];
1125  for (k = 0; k < i; k++)
1126  datay[i] += a[k + 1] * datax[i - k - 1];
1127  }
1128  for (i = Ncoeff; i < NSamples; i++) { //secondo loop. cambia l'indice interno.
1129  datay[i] = a[0] * datax[i];
1130  for (k = 0; k < Ncoeff - 1; k++)
1131  datay[i] += a[k + 1] * datax[i - k - 1];
1132  }
1133  break; // end of direct
1134  case 1: //reverse: cut & paste from direct and NSamples-1-
1135  for (i = 0; i < Ncoeff; i++) { // primo loop su Npunti
1136  datay[NSamples - 1 - i] = a[0] * datax[NSamples - 1 - i];
1137  for (k = 0; k < i; k++)
1138  datay[NSamples - 1 - i] += a[k + 1] * datax[NSamples - 1 - (i - k - 1)];
1139  }
1140  for (i = Ncoeff; i < NSamples; i++) { //secondo loop. cambia l'indice interno.
1141  datay[NSamples - 1 - i] = a[0] * datax[NSamples - 1 - i];
1142  for (k = 0; k < Ncoeff - 1; k++)
1143  datay[NSamples - 1 - i] += a[k + 1] * datax[NSamples - 1 - (i - k - 1)];
1144  }
1145  break;
1146  case -1: // bidirectional
1147  FIRApplyTo(datax, NSamples, 0);
1148  FIRApplyTo(datax, NSamples, 1);
1149  return;
1150  default:
1151  printf("ERROR in %s: reverse=%d not supported\n", __PRETTY_FUNCTION__, reverse);
1152  }// end of reverse switch.
1153  /*----------------------------------------------*/
1154  // void *memcpy(void *dest, const void *src, size_t n);
1155  for (int i = 0; i < NSamples; i++)
1156  datax[i] = (float)datay[i];
1157 }
1158 
1159 //=============================================
1160 
1164 
1165 void KVDigitalFilter::FIRApplyTo(double* datax, const int NSamples, int reverse) const
1166 {
1167  // Copiato +- da KVSignal.cxx
1168  // Diversa la convenzione per a0 b0!
1169 
1170  std::vector<long double> datay(NSamples);
1171  int i = 0, k = 0;
1172  switch (reverse) {
1173  case 0:// direct
1174  for (i = 0; i < Ncoeff; i++) { // primo loop su Npunti
1175  datay[i] = a[0] * datax[i];
1176  for (k = 0; k < i; k++)
1177  datay[i] += a[k + 1] * datax[i - k - 1];
1178  }
1179  for (i = Ncoeff; i < NSamples; i++) { //secondo loop. cambia l'indice interno.
1180  datay[i] = a[0] * datax[i];
1181  for (k = 0; k < Ncoeff - 1; k++)
1182  datay[i] += a[k + 1] * datax[i - k - 1];
1183  }
1184  break; // end of direct
1185  case 1: //reverse: cut & paste from direct and NSamples-1-
1186  for (i = 0; i < Ncoeff; i++) { // primo loop su Npunti
1187  datay[NSamples - 1 - i] = a[0] * datax[NSamples - 1 - i];
1188  for (k = 0; k < i; k++)
1189  datay[NSamples - 1 - i] += a[k + 1] * datax[NSamples - 1 - (i - k - 1)];
1190  }
1191  for (i = Ncoeff; i < NSamples; i++) { //secondo loop. cambia l'indice interno.
1192  datay[NSamples - 1 - i] = a[0] * datax[NSamples - 1 - i];
1193  for (k = 0; k < Ncoeff - 1; k++)
1194  datay[NSamples - 1 - i] += a[k + 1] * datax[NSamples - 1 - (i - k - 1)];
1195  }
1196  break;
1197  case -1: // bidirectional
1198  FIRApplyTo(datax, NSamples, 0);
1199  FIRApplyTo(datax, NSamples, 1);
1200  return;
1201  default:
1202  printf("ERROR in %s: reverse=%d not supported\n", __PRETTY_FUNCTION__, reverse);
1203  }// end of reverse switch.
1204  /*----------------------------------------------*/
1205  // void *memcpy(void *dest, const void *src, size_t n);
1206  // can't use memcpy: datax & datay have different sized types
1207  // (double=64bits & long double=128bits)
1208  //memcpy(datax, datay, NSamples * sizeof(double));
1209  for (int i = 0; i < NSamples; i++)
1210  datax[i] = (double)datay[i];
1211 }
1212 
1213 
KVBase
Base class for KaliVeda framework.
Definition: KVBase.h:142
KVDigitalFilter::CombineStages
static KVDigitalFilter CombineStages(const KVDigitalFilter &f1, const KVDigitalFilter &f2, int parallel=0)
Definition: KVDigitalFilter.cpp:659
KVDigitalFilter::BuildRCHighPassWithPZ
static KVDigitalFilter BuildRCHighPassWithPZ(const double &tau_usec, const double &preamp_decay_usec, const double &tau_clk)
Definition: KVDigitalFilter.cpp:124
KVDigitalFilter::GetYarray
double * GetYarray()
Definition: KVDigitalFilter.h:117
KVDigitalFilter::Compress
void Compress()
shorten filter. No deallocation of memory.
Definition: KVDigitalFilter.cpp:601
KVDigitalFilter::Draw
virtual void Draw(Option_t *option="")
Definition: KVDigitalFilter.cpp:974
KVDigitalFilter::GetYcoeff
const double & GetYcoeff(int i) const
Definition: KVDigitalFilter.h:91
KVDigitalFilter::ComputeChebyshevCoeffs_serv
static void ComputeChebyshevCoeffs_serv(const double &fc, const double &lh, const double &pr, const double &np, int p, double *a0, double *a1, double *a2, double *b1, double *b2)
Definition: KVDigitalFilter.cpp:172
KVDigitalFilter::Double2DSP
static int Double2DSP(const double &val)
– conversion to/from DSP 1.15? notation
Definition: KVDigitalFilter.h:51
KVDigitalFilter::SetXcoeff
void SetXcoeff(int i, const double &val)
Definition: KVDigitalFilter.h:97
KVDigitalFilter::BuildRCLowPassDeconv
static KVDigitalFilter BuildRCLowPassDeconv(const double &tau_usec, const double &tau_clk)
Definition: KVDigitalFilter.cpp:107
KVDigitalFilter::Quantize
void Quantize(int nbits, int use_pow2=0, double *xgain=NULL, int *x_out=NULL, int *y_out=NULL, int *x_scale=NULL, int *y_scale=NULL)
Definition: KVDigitalFilter.cpp:866
KVDigitalFilter::BuildRCLowPass
static KVDigitalFilter BuildRCLowPass(const double &tau_usec, const double &tau_clk)
Definition: KVDigitalFilter.cpp:90
KVDigitalFilter::PrintCoeffsDSP
void PrintCoeffsDSP() const
Definition: KVDigitalFilter.cpp:821
KVDigitalFilter::ComputeChebyshevCoeffs
static void ComputeChebyshevCoeffs(const double &freq_cutoff, int is_highpass, const double &percent_ripple, int npoles, double *a, double *b)
Definition: KVDigitalFilter.cpp:242
KVDigitalFilter::BuildChebyshev
static KVDigitalFilter BuildChebyshev(const double &freq_cutoff_mhz, int is_highpass, const double &percent_ripple, int npoles, const double &tau_clk)
Definition: KVDigitalFilter.cpp:155
KVDigitalFilter::GetXarray
double * GetXarray()
Definition: KVDigitalFilter.h:113
KVDigitalFilter::ApplyTo
void ApplyTo(double *data, const int N, int reverse=0) const
Definition: KVDigitalFilter.cpp:441
KVDigitalFilter::a
double * a
coefficients.
Definition: KVDigitalFilter.h:174
KVDigitalFilter::operator=
KVDigitalFilter operator=(const KVDigitalFilter &)
Definition: KVDigitalFilter.cpp:422
KVDigitalFilter::ReadMatlabFIR
int ReadMatlabFIR(char *filecoeff)
FILE *fin = fopen("notch_coeffs.txt","r");.
Definition: KVDigitalFilter.cpp:1057
KVDigitalFilter::BuildUnity
static KVDigitalFilter BuildUnity(const double &tau_clk)
Definition: KVDigitalFilter.cpp:772
KVDigitalFilter::PrintCoeffs
void PrintCoeffs() const
Definition: KVDigitalFilter.cpp:796
KVDigitalFilter::GetXcoeff
const double & GetXcoeff(int i) const
Definition: KVDigitalFilter.h:85
KVDigitalFilter::BuildRCHighPass
static KVDigitalFilter BuildRCHighPass(const double &tau_usec, const double &tau_clk)
Definition: KVDigitalFilter.cpp:137
KVDigitalFilter::SetYcoeff
void SetYcoeff(int i, const double &val)
Definition: KVDigitalFilter.h:105
KVDigitalFilter::GetTauClk
const double & GetTauClk()
Definition: KVDigitalFilter.h:123
KVDigitalFilter::BuildIntegrator
static KVDigitalFilter BuildIntegrator(const double &tau_clk)
Definition: KVDigitalFilter.cpp:784
KVDigitalFilter::KVDigitalFilter
KVDigitalFilter(const double &tau=10)
Definition: KVDigitalFilter.cpp:324
KVDigitalFilter::WriteMatlabFIR
int WriteMatlabFIR(char *filecoeff)
Definition: KVDigitalFilter.cpp:1090
KVDigitalFilter::PrintCoeffs_AsC
void PrintCoeffs_AsC() const
Definition: KVDigitalFilter.cpp:848
KVDigitalFilter::Alloc
void Alloc(const int Ncoeff)
printf("a=%p, N=%d, Ncoeff=%d\n", a, N, Ncoeff);
Definition: KVDigitalFilter.cpp:368
KVDigitalFilter::FIRApplyTo
void FIRApplyTo(double *datax, const int NSamples, int reverse) const
Definition: KVDigitalFilter.cpp:1165
KVDigitalFilter::BuildInverse
static KVDigitalFilter BuildInverse(KVDigitalFilter *filter)
‍**************************************‍/
Definition: KVDigitalFilter.cpp:1036
KVDigitalFilter::CombineStagesMany
static KVDigitalFilter CombineStagesMany(const KVDigitalFilter *f1, const KVDigitalFilter *f2, const KVDigitalFilter *f3=NULL, const KVDigitalFilter *f4=NULL, const KVDigitalFilter *f5=NULL, const KVDigitalFilter *f6=NULL, const KVDigitalFilter *f7=NULL, const KVDigitalFilter *f8=NULL, const KVDigitalFilter *f9=NULL, const KVDigitalFilter *f10=NULL)
se ne devi combinare + di 1 IN CASCATA!
Definition: KVDigitalFilter.cpp:636