| 1 | /* ======================================================================== *\
|
|---|
| 2 | !
|
|---|
| 3 | ! *
|
|---|
| 4 | ! * This file is part of MARS, the MAGIC Analysis and Reconstruction
|
|---|
| 5 | ! * Software. It is distributed to you in the hope that it can be a useful
|
|---|
| 6 | ! * and timesaving tool in analyzing Data of imaging Cerenkov telescopes.
|
|---|
| 7 | ! * It is distributed WITHOUT ANY WARRANTY.
|
|---|
| 8 | ! *
|
|---|
| 9 | ! * Permission to use, copy, modify and distribute this software and its
|
|---|
| 10 | ! * documentation for any purpose is hereby granted without fee,
|
|---|
| 11 | ! * provided that the above copyright notice appear in all copies and
|
|---|
| 12 | ! * that both that copyright notice and this permission notice appear
|
|---|
| 13 | ! * in supporting documentation. It is provided "as is" without express
|
|---|
| 14 | ! * or implied warranty.
|
|---|
| 15 | ! *
|
|---|
| 16 | !
|
|---|
| 17 | ! Author(s): Markus Gaug 05/2004 <mailto:markus@ifae.es>
|
|---|
| 18 | !
|
|---|
| 19 | ! Copyright: MAGIC Software Development, 2002-2004
|
|---|
| 20 | !
|
|---|
| 21 | !
|
|---|
| 22 | \* ======================================================================== */
|
|---|
| 23 |
|
|---|
| 24 | //////////////////////////////////////////////////////////////////////////////
|
|---|
| 25 | //
|
|---|
| 26 | // MExtractTimeFastSpline
|
|---|
| 27 | //
|
|---|
| 28 | // Fast arrival Time extractor using a cubic spline algorithm of Numerical Recipes.
|
|---|
| 29 | // It returns the position of the half maximum between absolute maximum
|
|---|
| 30 | // and pedestal of the spline that interpolates the FADC slices.
|
|---|
| 31 | //
|
|---|
| 32 | // The precision of the half-maximum searches can be chosen by:
|
|---|
| 33 | // SetPrecision().
|
|---|
| 34 | //
|
|---|
| 35 | // The precision of the maximum-finder is fixed to 0.025 FADC units.
|
|---|
| 36 | //
|
|---|
| 37 | //////////////////////////////////////////////////////////////////////////////
|
|---|
| 38 | #include "MExtractTimeFastSpline.h"
|
|---|
| 39 |
|
|---|
| 40 | #include "MPedestalPix.h"
|
|---|
| 41 |
|
|---|
| 42 | #include "MLog.h"
|
|---|
| 43 | #include "MLogManip.h"
|
|---|
| 44 |
|
|---|
| 45 |
|
|---|
| 46 | ClassImp(MExtractTimeFastSpline);
|
|---|
| 47 |
|
|---|
| 48 | using namespace std;
|
|---|
| 49 |
|
|---|
| 50 | const Byte_t MExtractTimeFastSpline::fgHiGainFirst = 2;
|
|---|
| 51 | const Byte_t MExtractTimeFastSpline::fgHiGainLast = 14;
|
|---|
| 52 | const Byte_t MExtractTimeFastSpline::fgLoGainFirst = 3;
|
|---|
| 53 | const Byte_t MExtractTimeFastSpline::fgLoGainLast = 14;
|
|---|
| 54 | const Float_t MExtractTimeFastSpline::fgResolution = 0.003;
|
|---|
| 55 | const Float_t MExtractTimeFastSpline::fgRiseTime = 2.;
|
|---|
| 56 |
|
|---|
| 57 | // --------------------------------------------------------------------------
|
|---|
| 58 | //
|
|---|
| 59 | // Default constructor.
|
|---|
| 60 | //
|
|---|
| 61 | // Calls:
|
|---|
| 62 | // - SetRange(fgHiGainFirst, fgHiGainLast, fgLoGainFirst, fgLoGainLast)
|
|---|
| 63 | //
|
|---|
| 64 | // Initializes:
|
|---|
| 65 | // - fResolution to fgResolution
|
|---|
| 66 | //
|
|---|
| 67 | MExtractTimeFastSpline::MExtractTimeFastSpline(const char *name, const char *title)
|
|---|
| 68 | : fHiGainFirstDeriv(NULL), fLoGainFirstDeriv(NULL),
|
|---|
| 69 | fHiGainSecondDeriv(NULL), fLoGainSecondDeriv(NULL)
|
|---|
| 70 | {
|
|---|
| 71 |
|
|---|
| 72 | fName = name ? name : "MExtractTimeFastSpline";
|
|---|
| 73 | fTitle = title ? title : "Calculate photons arrival time using a fast spline";
|
|---|
| 74 |
|
|---|
| 75 | SetResolution();
|
|---|
| 76 | SetRiseTime ();
|
|---|
| 77 | SetRange(fgHiGainFirst, fgHiGainLast, fgLoGainFirst, fgLoGainLast);
|
|---|
| 78 |
|
|---|
| 79 | }
|
|---|
| 80 |
|
|---|
| 81 | MExtractTimeFastSpline::~MExtractTimeFastSpline()
|
|---|
| 82 | {
|
|---|
| 83 |
|
|---|
| 84 | if (fHiGainFirstDeriv)
|
|---|
| 85 | delete [] fHiGainFirstDeriv;
|
|---|
| 86 | if (fLoGainFirstDeriv)
|
|---|
| 87 | delete [] fLoGainFirstDeriv;
|
|---|
| 88 | if (fHiGainSecondDeriv)
|
|---|
| 89 | delete [] fHiGainSecondDeriv;
|
|---|
| 90 | if (fLoGainSecondDeriv)
|
|---|
| 91 | delete [] fLoGainSecondDeriv;
|
|---|
| 92 |
|
|---|
| 93 | }
|
|---|
| 94 |
|
|---|
| 95 |
|
|---|
| 96 | // --------------------------------------------------------------------------
|
|---|
| 97 | //
|
|---|
| 98 | // SetRange:
|
|---|
| 99 | //
|
|---|
| 100 | // Calls:
|
|---|
| 101 | // - MExtractor::SetRange(hifirst,hilast,lofirst,lolast);
|
|---|
| 102 | // - Deletes x, if not NULL
|
|---|
| 103 | // - Creates x according to the range
|
|---|
| 104 | //
|
|---|
| 105 | void MExtractTimeFastSpline::SetRange(Byte_t hifirst, Byte_t hilast, Byte_t lofirst, Byte_t lolast)
|
|---|
| 106 | {
|
|---|
| 107 |
|
|---|
| 108 | MExtractor::SetRange(hifirst,hilast,lofirst,lolast);
|
|---|
| 109 |
|
|---|
| 110 | if (fHiGainFirstDeriv)
|
|---|
| 111 | delete [] fHiGainFirstDeriv;
|
|---|
| 112 | if (fLoGainFirstDeriv)
|
|---|
| 113 | delete [] fLoGainFirstDeriv;
|
|---|
| 114 | if (fHiGainSecondDeriv)
|
|---|
| 115 | delete [] fHiGainSecondDeriv;
|
|---|
| 116 | if (fLoGainSecondDeriv)
|
|---|
| 117 | delete [] fLoGainSecondDeriv;
|
|---|
| 118 |
|
|---|
| 119 | Int_t range = fHiGainLast - fHiGainFirst + 1;
|
|---|
| 120 |
|
|---|
| 121 | if (range < 2)
|
|---|
| 122 | {
|
|---|
| 123 | *fLog << warn << GetDescriptor()
|
|---|
| 124 | << Form("%s%2i%s%2i%s",": Hi-Gain Extraction range [",(int)fHiGainFirst,","
|
|---|
| 125 | ,fHiGainLast,"] too small, ") << endl;
|
|---|
| 126 | *fLog << warn << GetDescriptor()
|
|---|
| 127 | << " will move higher limit to obtain 4 slices " << endl;
|
|---|
| 128 | SetRange(fHiGainFirst, fHiGainLast+4-range,fLoGainFirst,fLoGainLast);
|
|---|
| 129 | range = fHiGainLast - fHiGainFirst + 1;
|
|---|
| 130 | }
|
|---|
| 131 |
|
|---|
| 132 |
|
|---|
| 133 | fHiGainFirstDeriv = new Float_t[range];
|
|---|
| 134 | memset(fHiGainFirstDeriv,0,range*sizeof(Float_t));
|
|---|
| 135 | fHiGainSecondDeriv = new Float_t[range];
|
|---|
| 136 | memset(fHiGainSecondDeriv,0,range*sizeof(Float_t));
|
|---|
| 137 |
|
|---|
| 138 | range = fLoGainLast - fLoGainFirst + 1;
|
|---|
| 139 |
|
|---|
| 140 | if (range >= 2)
|
|---|
| 141 | {
|
|---|
| 142 |
|
|---|
| 143 | fLoGainFirstDeriv = new Float_t[range];
|
|---|
| 144 | memset(fLoGainFirstDeriv,0,range*sizeof(Float_t));
|
|---|
| 145 | fLoGainSecondDeriv = new Float_t[range];
|
|---|
| 146 | memset(fLoGainSecondDeriv,0,range*sizeof(Float_t));
|
|---|
| 147 |
|
|---|
| 148 | }
|
|---|
| 149 |
|
|---|
| 150 | }
|
|---|
| 151 |
|
|---|
| 152 |
|
|---|
| 153 | // --------------------------------------------------------------------------
|
|---|
| 154 | //
|
|---|
| 155 | // Calculates the arrival time for each pixel
|
|---|
| 156 | //
|
|---|
| 157 | void MExtractTimeFastSpline::FindTimeHiGain(Byte_t *first, Float_t &time, Float_t &dtime,
|
|---|
| 158 | Byte_t &sat, const MPedestalPix &ped) const
|
|---|
| 159 | {
|
|---|
| 160 |
|
|---|
| 161 | const Int_t range = fHiGainLast - fHiGainFirst + 1;
|
|---|
| 162 | const Byte_t *end = first + range;
|
|---|
| 163 | Byte_t *p = first;
|
|---|
| 164 | Byte_t max = 0;
|
|---|
| 165 | Byte_t maxpos = 0;
|
|---|
| 166 |
|
|---|
| 167 | //
|
|---|
| 168 | // Check for saturation in all other slices
|
|---|
| 169 | //
|
|---|
| 170 | while (p<end)
|
|---|
| 171 | {
|
|---|
| 172 | if (*p > max)
|
|---|
| 173 | {
|
|---|
| 174 | max = *p;
|
|---|
| 175 | maxpos = p-first;
|
|---|
| 176 | }
|
|---|
| 177 |
|
|---|
| 178 | if (*p++ >= fSaturationLimit)
|
|---|
| 179 | {
|
|---|
| 180 | sat++;
|
|---|
| 181 | break;
|
|---|
| 182 | }
|
|---|
| 183 | }
|
|---|
| 184 |
|
|---|
| 185 | //
|
|---|
| 186 | // allow one saturated slice
|
|---|
| 187 | //
|
|---|
| 188 | if (sat > 1)
|
|---|
| 189 | return;
|
|---|
| 190 |
|
|---|
| 191 | if (maxpos < 1)
|
|---|
| 192 | {
|
|---|
| 193 | time = -999.;
|
|---|
| 194 | return;
|
|---|
| 195 | }
|
|---|
| 196 |
|
|---|
| 197 | Float_t pp;
|
|---|
| 198 |
|
|---|
| 199 | p = first;
|
|---|
| 200 | fHiGainSecondDeriv[0] = 0.;
|
|---|
| 201 | fHiGainFirstDeriv[0] = 0.;
|
|---|
| 202 |
|
|---|
| 203 | for (Int_t i=1;i<range-1;i++)
|
|---|
| 204 | {
|
|---|
| 205 | p++;
|
|---|
| 206 | pp = fHiGainSecondDeriv[i-1] + 4.;
|
|---|
| 207 | fHiGainSecondDeriv[i] = -1.0/pp;
|
|---|
| 208 | fHiGainFirstDeriv [i] = *(p+1) - 2.* *(p) + *(p-1);
|
|---|
| 209 | fHiGainFirstDeriv [i] = (6.0*fHiGainFirstDeriv[i]-fHiGainFirstDeriv[i-1])/pp;
|
|---|
| 210 | }
|
|---|
| 211 |
|
|---|
| 212 | fHiGainSecondDeriv[range-1] = 0.;
|
|---|
| 213 |
|
|---|
| 214 | for (Int_t k=range-2;k>0;k--)
|
|---|
| 215 | fHiGainSecondDeriv[k] = fHiGainSecondDeriv[k]*fHiGainSecondDeriv[k+1] + fHiGainFirstDeriv[k];
|
|---|
| 216 | for (Int_t k=range-2;k>0;k--)
|
|---|
| 217 | fHiGainSecondDeriv[k] /= 6.;
|
|---|
| 218 |
|
|---|
| 219 | //
|
|---|
| 220 | // Now find the maximum
|
|---|
| 221 | //
|
|---|
| 222 | Float_t step = 0.2; // start with step size of 1ns and loop again with the smaller one
|
|---|
| 223 | Float_t lower = (Float_t)maxpos-1.;
|
|---|
| 224 | Float_t upper = (Float_t)maxpos;
|
|---|
| 225 | Float_t x = lower;
|
|---|
| 226 | Float_t y = 0.;
|
|---|
| 227 | Float_t a = 1.;
|
|---|
| 228 | Float_t b = 0.;
|
|---|
| 229 | Int_t klo = maxpos-1;
|
|---|
| 230 | Int_t khi = maxpos;
|
|---|
| 231 | Float_t klocont = (Float_t)*(first+klo);
|
|---|
| 232 | Float_t khicont = (Float_t)*(first+khi);
|
|---|
| 233 | time = upper;
|
|---|
| 234 | Float_t abmax = khicont;
|
|---|
| 235 |
|
|---|
| 236 | //
|
|---|
| 237 | // Search for the maximum, starting in interval maxpos-1 in steps of 0.2 till maxpos-0.2.
|
|---|
| 238 | // If no maximum is found, go to interval maxpos+1.
|
|---|
| 239 | //
|
|---|
| 240 | Float_t higainklo = fHiGainSecondDeriv[klo];
|
|---|
| 241 | Float_t higainkhi = fHiGainSecondDeriv[khi];
|
|---|
| 242 |
|
|---|
| 243 | while ( x < upper - 0.3 )
|
|---|
| 244 | {
|
|---|
| 245 |
|
|---|
| 246 | x += step;
|
|---|
| 247 | a -= step;
|
|---|
| 248 | b += step;
|
|---|
| 249 |
|
|---|
| 250 | y = a*klocont
|
|---|
| 251 | + b*khicont
|
|---|
| 252 | + (a*a*a-a)*higainklo
|
|---|
| 253 | + (b*b*b-b)*higainkhi;
|
|---|
| 254 |
|
|---|
| 255 | if (y > abmax)
|
|---|
| 256 | {
|
|---|
| 257 | abmax = y;
|
|---|
| 258 | time = x;
|
|---|
| 259 | }
|
|---|
| 260 | }
|
|---|
| 261 |
|
|---|
| 262 | //
|
|---|
| 263 | // Search for the absolute maximum from maxpos to maxpos+1 in steps of 0.2
|
|---|
| 264 | //
|
|---|
| 265 | if (time > upper-0.1)
|
|---|
| 266 | {
|
|---|
| 267 |
|
|---|
| 268 | upper = (Float_t)maxpos+1.;
|
|---|
| 269 | lower = (Float_t)maxpos;
|
|---|
| 270 | x = lower;
|
|---|
| 271 | a = 1.;
|
|---|
| 272 | b = 0.;
|
|---|
| 273 | khi = maxpos+1;
|
|---|
| 274 | klo = maxpos;
|
|---|
| 275 | klocont = (Float_t)*(first+klo);
|
|---|
| 276 | khicont = (Float_t)*(first+khi);
|
|---|
| 277 |
|
|---|
| 278 | higainklo = fHiGainSecondDeriv[klo];
|
|---|
| 279 | higainkhi = fHiGainSecondDeriv[khi];
|
|---|
| 280 |
|
|---|
| 281 | while (x<upper-0.3)
|
|---|
| 282 | {
|
|---|
| 283 |
|
|---|
| 284 | x += step;
|
|---|
| 285 | a -= step;
|
|---|
| 286 | b += step;
|
|---|
| 287 |
|
|---|
| 288 | y = a* klocont
|
|---|
| 289 | + b* khicont
|
|---|
| 290 | + (a*a*a-a)*higainklo
|
|---|
| 291 | + (b*b*b-b)*higainkhi;
|
|---|
| 292 |
|
|---|
| 293 | if (y > abmax)
|
|---|
| 294 | {
|
|---|
| 295 | abmax = y;
|
|---|
| 296 | time = x;
|
|---|
| 297 | }
|
|---|
| 298 | }
|
|---|
| 299 | }
|
|---|
| 300 |
|
|---|
| 301 | //
|
|---|
| 302 | // Now, the time, abmax and khicont and klocont are set correctly within the previous precision.
|
|---|
| 303 | // Try a better precision.
|
|---|
| 304 | //
|
|---|
| 305 | const Float_t up = time+step-0.035;
|
|---|
| 306 | const Float_t lo = time-step+0.035;
|
|---|
| 307 | const Float_t maxpossave = time;
|
|---|
| 308 |
|
|---|
| 309 | x = time;
|
|---|
| 310 | a = upper - x;
|
|---|
| 311 | b = x - lower;
|
|---|
| 312 |
|
|---|
| 313 | step = 0.025; // step size of 83 ps
|
|---|
| 314 |
|
|---|
| 315 | higainklo = fHiGainSecondDeriv[klo];
|
|---|
| 316 | higainkhi = fHiGainSecondDeriv[khi];
|
|---|
| 317 |
|
|---|
| 318 | //
|
|---|
| 319 | // First, try from time up to time+0.2 in steps of 83ps.
|
|---|
| 320 | //
|
|---|
| 321 | while ( x < up )
|
|---|
| 322 | {
|
|---|
| 323 |
|
|---|
| 324 | x += step;
|
|---|
| 325 | a -= step;
|
|---|
| 326 | b += step;
|
|---|
| 327 |
|
|---|
| 328 | y = a* klocont
|
|---|
| 329 | + b* khicont
|
|---|
| 330 | + (a*a*a-a)*higainklo
|
|---|
| 331 | + (b*b*b-b)*higainkhi;
|
|---|
| 332 |
|
|---|
| 333 | if (y > abmax)
|
|---|
| 334 | {
|
|---|
| 335 | abmax = y;
|
|---|
| 336 | time = x;
|
|---|
| 337 | }
|
|---|
| 338 |
|
|---|
| 339 | }
|
|---|
| 340 |
|
|---|
| 341 |
|
|---|
| 342 | //
|
|---|
| 343 | // Second, try from time down to time-0.2 in steps of 0.04.
|
|---|
| 344 | //
|
|---|
| 345 | x = maxpossave;
|
|---|
| 346 | //
|
|---|
| 347 | // Test the possibility that the absolute maximum has not been found between
|
|---|
| 348 | // maxpos and maxpos+0.02, then we have to look between maxpos-0.02 and maxpos
|
|---|
| 349 | // which requires new setting of klocont and khicont
|
|---|
| 350 | //
|
|---|
| 351 | if (x < klo + 0.02)
|
|---|
| 352 | {
|
|---|
| 353 | klo--;
|
|---|
| 354 | khi--;
|
|---|
| 355 | klocont = (Float_t)*(first+klo);
|
|---|
| 356 | khicont = (Float_t)*(first+khi);
|
|---|
| 357 | upper--;
|
|---|
| 358 | lower--;
|
|---|
| 359 | }
|
|---|
| 360 |
|
|---|
| 361 | a = upper - x;
|
|---|
| 362 | b = x - lower;
|
|---|
| 363 |
|
|---|
| 364 | higainklo = fHiGainSecondDeriv[klo];
|
|---|
| 365 | higainkhi = fHiGainSecondDeriv[khi];
|
|---|
| 366 |
|
|---|
| 367 | while ( x > lo )
|
|---|
| 368 | {
|
|---|
| 369 |
|
|---|
| 370 | x -= step;
|
|---|
| 371 | a += step;
|
|---|
| 372 | b -= step;
|
|---|
| 373 |
|
|---|
| 374 | y = a* klocont
|
|---|
| 375 | + b* khicont
|
|---|
| 376 | + (a*a*a-a)*higainklo
|
|---|
| 377 | + (b*b*b-b)*higainkhi;
|
|---|
| 378 |
|
|---|
| 379 | if (y > abmax)
|
|---|
| 380 | {
|
|---|
| 381 | abmax = y;
|
|---|
| 382 | time = x;
|
|---|
| 383 | }
|
|---|
| 384 | }
|
|---|
| 385 |
|
|---|
| 386 | #if 0
|
|---|
| 387 | const Float_t pedes = ped.GetPedestal();
|
|---|
| 388 | const Float_t halfmax = pedes + (abmax - pedes)/2.;
|
|---|
| 389 |
|
|---|
| 390 | //
|
|---|
| 391 | // Now, loop from the maximum bin leftward down in order to find the position of the half maximum.
|
|---|
| 392 | // First, find the right FADC slice:
|
|---|
| 393 | //
|
|---|
| 394 | klo = maxpos;
|
|---|
| 395 | while (klo > maxpos-fStartBeforeMax)
|
|---|
| 396 | {
|
|---|
| 397 | if (*(first+klo) < (Byte_t)halfmax)
|
|---|
| 398 | break;
|
|---|
| 399 | klo--;
|
|---|
| 400 | }
|
|---|
| 401 |
|
|---|
| 402 | //
|
|---|
| 403 | // Loop from the beginning of the slice upwards to reach the halfmax:
|
|---|
| 404 | // With means of bisection:
|
|---|
| 405 | //
|
|---|
| 406 | x = (Float_t)klo;
|
|---|
| 407 | a = 1.;
|
|---|
| 408 | b = 0.;
|
|---|
| 409 | klocont = (Float_t)*(first+klo);
|
|---|
| 410 | khicont = (Float_t)*(first+klo+1);
|
|---|
| 411 |
|
|---|
| 412 | step = 0.5;
|
|---|
| 413 | Bool_t back = kFALSE;
|
|---|
| 414 |
|
|---|
| 415 | while (step > fResolution)
|
|---|
| 416 | {
|
|---|
| 417 |
|
|---|
| 418 | if (back)
|
|---|
| 419 | {
|
|---|
| 420 | x -= step;
|
|---|
| 421 | a += step;
|
|---|
| 422 | b -= step;
|
|---|
| 423 | }
|
|---|
| 424 | else
|
|---|
| 425 | {
|
|---|
| 426 | x += step;
|
|---|
| 427 | a -= step;
|
|---|
| 428 | b += step;
|
|---|
| 429 | }
|
|---|
| 430 |
|
|---|
| 431 | y = a*klocont
|
|---|
| 432 | + b*khicont
|
|---|
| 433 | + (a*a*a-a)*fHiGainSecondDeriv[klo]
|
|---|
| 434 | + (b*b*b-b)*fHiGainSecondDeriv[khi];
|
|---|
| 435 |
|
|---|
| 436 | if (y >= halfmax)
|
|---|
| 437 | back = kTRUE;
|
|---|
| 438 | else
|
|---|
| 439 | back = kFALSE;
|
|---|
| 440 |
|
|---|
| 441 | step /= 2.;
|
|---|
| 442 |
|
|---|
| 443 | }
|
|---|
| 444 | time = (Float_t)fHiGainFirst + x;
|
|---|
| 445 |
|
|---|
| 446 | #endif
|
|---|
| 447 | dtime = 0.035;
|
|---|
| 448 |
|
|---|
| 449 | }
|
|---|
| 450 |
|
|---|
| 451 |
|
|---|
| 452 | // --------------------------------------------------------------------------
|
|---|
| 453 | //
|
|---|
| 454 | // Calculates the arrival time for each pixel
|
|---|
| 455 | //
|
|---|
| 456 | void MExtractTimeFastSpline::FindTimeLoGain(Byte_t *first, Float_t &time, Float_t &dtime,
|
|---|
| 457 | Byte_t &sat, const MPedestalPix &ped) const
|
|---|
| 458 | {
|
|---|
| 459 |
|
|---|
| 460 | const Int_t range = fLoGainLast - fLoGainFirst + 1;
|
|---|
| 461 | const Byte_t *end = first + range;
|
|---|
| 462 | Byte_t *p = first;
|
|---|
| 463 | Byte_t max = 0;
|
|---|
| 464 | Byte_t maxpos = 0;
|
|---|
| 465 |
|
|---|
| 466 | //
|
|---|
| 467 | // Check for saturation in all other slices
|
|---|
| 468 | //
|
|---|
| 469 | while (p<end)
|
|---|
| 470 | {
|
|---|
| 471 | if (*p > max)
|
|---|
| 472 | {
|
|---|
| 473 | max = *p;
|
|---|
| 474 | maxpos = p-first;
|
|---|
| 475 | }
|
|---|
| 476 |
|
|---|
| 477 | if (*p++ >= fSaturationLimit)
|
|---|
| 478 | {
|
|---|
| 479 | sat++;
|
|---|
| 480 | break;
|
|---|
| 481 | }
|
|---|
| 482 | }
|
|---|
| 483 |
|
|---|
| 484 | if (sat)
|
|---|
| 485 | return;
|
|---|
| 486 |
|
|---|
| 487 | if (maxpos < 1)
|
|---|
| 488 | return;
|
|---|
| 489 |
|
|---|
| 490 | Float_t pp;
|
|---|
| 491 |
|
|---|
| 492 | p = first;
|
|---|
| 493 | fLoGainSecondDeriv[0] = 0.;
|
|---|
| 494 | fLoGainFirstDeriv[0] = 0.;
|
|---|
| 495 |
|
|---|
| 496 | for (Int_t i=1;i<range-1;i++)
|
|---|
| 497 | {
|
|---|
| 498 | p++;
|
|---|
| 499 | pp = fLoGainSecondDeriv[i-1] + 4.;
|
|---|
| 500 | fLoGainSecondDeriv[i] = -1.0/pp;
|
|---|
| 501 | fLoGainFirstDeriv [i] = *(p+1) - 2.* *(p) + *(p-1);
|
|---|
| 502 | fLoGainFirstDeriv [i] = (6.0*fLoGainFirstDeriv[i]-fLoGainFirstDeriv[i-1])/pp;
|
|---|
| 503 | }
|
|---|
| 504 |
|
|---|
| 505 | fLoGainSecondDeriv[range-1] = 0.;
|
|---|
| 506 |
|
|---|
| 507 | for (Int_t k=range-2;k>0;k--)
|
|---|
| 508 | fLoGainSecondDeriv[k] = fLoGainSecondDeriv[k]*fLoGainSecondDeriv[k+1] + fLoGainFirstDeriv[k];
|
|---|
| 509 | for (Int_t k=range-2;k>0;k--)
|
|---|
| 510 | fLoGainSecondDeriv[k] /= 6.;
|
|---|
| 511 |
|
|---|
| 512 | //
|
|---|
| 513 | // Now find the maximum
|
|---|
| 514 | //
|
|---|
| 515 | Float_t step = 0.2; // start with step size of 1ns and loop again with the smaller one
|
|---|
| 516 | Float_t lower = (Float_t)maxpos-1.;
|
|---|
| 517 | Float_t upper = (Float_t)maxpos;
|
|---|
| 518 | Float_t x = lower;
|
|---|
| 519 | Float_t y = 0.;
|
|---|
| 520 | Float_t a = 1.;
|
|---|
| 521 | Float_t b = 0.;
|
|---|
| 522 | Int_t klo = maxpos-1;
|
|---|
| 523 | Int_t khi = maxpos;
|
|---|
| 524 | Float_t klocont = (Float_t)*(first+klo);
|
|---|
| 525 | Float_t khicont = (Float_t)*(first+khi);
|
|---|
| 526 | time = upper;
|
|---|
| 527 | Float_t abmax = khicont;
|
|---|
| 528 |
|
|---|
| 529 | //
|
|---|
| 530 | // Search for the maximum, starting in interval maxpos-1. If no maximum is found, go to
|
|---|
| 531 | // interval maxpos+1.
|
|---|
| 532 | //
|
|---|
| 533 | while (x<upper-0.3)
|
|---|
| 534 | {
|
|---|
| 535 |
|
|---|
| 536 | x += step;
|
|---|
| 537 | a -= step;
|
|---|
| 538 | b += step;
|
|---|
| 539 |
|
|---|
| 540 | y = a*klocont
|
|---|
| 541 | + b*khicont
|
|---|
| 542 | + (a*a*a-a)*fLoGainSecondDeriv[klo]
|
|---|
| 543 | + (b*b*b-b)*fLoGainSecondDeriv[khi];
|
|---|
| 544 |
|
|---|
| 545 | if (y > abmax)
|
|---|
| 546 | {
|
|---|
| 547 | abmax = y;
|
|---|
| 548 | time = x;
|
|---|
| 549 | }
|
|---|
| 550 |
|
|---|
| 551 | }
|
|---|
| 552 |
|
|---|
| 553 | if (time > upper-0.1)
|
|---|
| 554 | {
|
|---|
| 555 |
|
|---|
| 556 | upper = (Float_t)maxpos+1.;
|
|---|
| 557 | lower = (Float_t)maxpos;
|
|---|
| 558 | x = lower;
|
|---|
| 559 | a = 1.;
|
|---|
| 560 | b = 0.;
|
|---|
| 561 | khi = maxpos+1;
|
|---|
| 562 | klo = maxpos;
|
|---|
| 563 | klocont = (Float_t)*(first+klo);
|
|---|
| 564 | khicont = (Float_t)*(first+khi);
|
|---|
| 565 |
|
|---|
| 566 | while (x<upper-0.3)
|
|---|
| 567 | {
|
|---|
| 568 |
|
|---|
| 569 | x += step;
|
|---|
| 570 | a -= step;
|
|---|
| 571 | b += step;
|
|---|
| 572 |
|
|---|
| 573 | y = a* klocont
|
|---|
| 574 | + b* khicont
|
|---|
| 575 | + (a*a*a-a)*fLoGainSecondDeriv[klo]
|
|---|
| 576 | + (b*b*b-b)*fLoGainSecondDeriv[khi];
|
|---|
| 577 |
|
|---|
| 578 | if (y > abmax)
|
|---|
| 579 | {
|
|---|
| 580 | abmax = y;
|
|---|
| 581 | time = x;
|
|---|
| 582 | }
|
|---|
| 583 | }
|
|---|
| 584 | }
|
|---|
| 585 |
|
|---|
| 586 | const Float_t up = time+step-0.055;
|
|---|
| 587 | const Float_t lo = time-step+0.055;
|
|---|
| 588 | const Float_t maxpossave = time;
|
|---|
| 589 |
|
|---|
| 590 | x = time;
|
|---|
| 591 | a = upper - x;
|
|---|
| 592 | b = x - lower;
|
|---|
| 593 |
|
|---|
| 594 | step = 0.025; // step size of 165 ps
|
|---|
| 595 |
|
|---|
| 596 | while (x<up)
|
|---|
| 597 | {
|
|---|
| 598 |
|
|---|
| 599 | x += step;
|
|---|
| 600 | a -= step;
|
|---|
| 601 | b += step;
|
|---|
| 602 |
|
|---|
| 603 | y = a* klocont
|
|---|
| 604 | + b* khicont
|
|---|
| 605 | + (a*a*a-a)*fLoGainSecondDeriv[klo]
|
|---|
| 606 | + (b*b*b-b)*fLoGainSecondDeriv[khi];
|
|---|
| 607 |
|
|---|
| 608 | if (y > abmax)
|
|---|
| 609 | {
|
|---|
| 610 | abmax = y;
|
|---|
| 611 | time = x;
|
|---|
| 612 | }
|
|---|
| 613 |
|
|---|
| 614 | }
|
|---|
| 615 |
|
|---|
| 616 | if (time < klo + 0.01)
|
|---|
| 617 | {
|
|---|
| 618 | klo--;
|
|---|
| 619 | khi--;
|
|---|
| 620 | klocont = (Float_t)*(first+klo);
|
|---|
| 621 | khicont = (Float_t)*(first+khi);
|
|---|
| 622 | upper--;
|
|---|
| 623 | lower--;
|
|---|
| 624 | }
|
|---|
| 625 |
|
|---|
| 626 | x = maxpossave;
|
|---|
| 627 | a = upper - x;
|
|---|
| 628 | b = x - lower;
|
|---|
| 629 |
|
|---|
| 630 | while (x>lo)
|
|---|
| 631 | {
|
|---|
| 632 |
|
|---|
| 633 | x -= step;
|
|---|
| 634 | a += step;
|
|---|
| 635 | b -= step;
|
|---|
| 636 |
|
|---|
| 637 | y = a* klocont
|
|---|
| 638 | + b* khicont
|
|---|
| 639 | + (a*a*a-a)*fLoGainSecondDeriv[klo]
|
|---|
| 640 | + (b*b*b-b)*fLoGainSecondDeriv[khi];
|
|---|
| 641 |
|
|---|
| 642 | if (y > abmax)
|
|---|
| 643 | {
|
|---|
| 644 | abmax = y;
|
|---|
| 645 | time = x;
|
|---|
| 646 | }
|
|---|
| 647 |
|
|---|
| 648 | }
|
|---|
| 649 |
|
|---|
| 650 | const Float_t pedes = ped.GetPedestal();
|
|---|
| 651 | const Float_t halfmax = pedes + (abmax - pedes)/2.;
|
|---|
| 652 |
|
|---|
| 653 | //
|
|---|
| 654 | // Now, loop from the maximum bin leftward down in order to find the position of the half maximum.
|
|---|
| 655 | // First, find the right FADC slice:
|
|---|
| 656 | //
|
|---|
| 657 | klo = maxpos;
|
|---|
| 658 | while (klo > maxpos-4)
|
|---|
| 659 | {
|
|---|
| 660 | if (*(first+klo) < (Byte_t)halfmax)
|
|---|
| 661 | break;
|
|---|
| 662 | klo--;
|
|---|
| 663 | }
|
|---|
| 664 |
|
|---|
| 665 | //
|
|---|
| 666 | // Loop from the beginning of the slice upwards to reach the halfmax:
|
|---|
| 667 | // With means of bisection:
|
|---|
| 668 | //
|
|---|
| 669 | x = (Float_t)klo;
|
|---|
| 670 | a = 1.;
|
|---|
| 671 | b = 0.;
|
|---|
| 672 | klocont = (Float_t)*(first+klo);
|
|---|
| 673 | khicont = (Float_t)*(first+klo+1);
|
|---|
| 674 | time = x;
|
|---|
| 675 |
|
|---|
| 676 | step = 0.5;
|
|---|
| 677 | Bool_t back = kFALSE;
|
|---|
| 678 |
|
|---|
| 679 | while (step > fResolution)
|
|---|
| 680 | {
|
|---|
| 681 |
|
|---|
| 682 | if (back)
|
|---|
| 683 | {
|
|---|
| 684 | x -= step;
|
|---|
| 685 | a += step;
|
|---|
| 686 | b -= step;
|
|---|
| 687 | }
|
|---|
| 688 | else
|
|---|
| 689 | {
|
|---|
| 690 | x += step;
|
|---|
| 691 | a -= step;
|
|---|
| 692 | b += step;
|
|---|
| 693 | }
|
|---|
| 694 |
|
|---|
| 695 | y = a*klocont
|
|---|
| 696 | + b*khicont
|
|---|
| 697 | + (a*a*a-a)*fLoGainSecondDeriv[klo]
|
|---|
| 698 | + (b*b*b-b)*fLoGainSecondDeriv[khi];
|
|---|
| 699 |
|
|---|
| 700 | if (y >= halfmax)
|
|---|
| 701 | back = kTRUE;
|
|---|
| 702 | else
|
|---|
| 703 | back = kFALSE;
|
|---|
| 704 |
|
|---|
| 705 | step /= 2.;
|
|---|
| 706 |
|
|---|
| 707 | }
|
|---|
| 708 |
|
|---|
| 709 | time = (Float_t)fLoGainFirst + x;
|
|---|
| 710 | dtime = fResolution;
|
|---|
| 711 | }
|
|---|
| 712 |
|
|---|
| 713 | // --------------------------------------------------------------------------
|
|---|
| 714 | //
|
|---|
| 715 | // In addition to the resources of the base-class MExtractor:
|
|---|
| 716 | // MJPedestal.MExtractor.Resolution: 0.003
|
|---|
| 717 | // MJPedestal.MExtractor.RiseTime: 1.5
|
|---|
| 718 | //
|
|---|
| 719 | Int_t MExtractTimeFastSpline::ReadEnv(const TEnv &env, TString prefix, Bool_t print)
|
|---|
| 720 | {
|
|---|
| 721 | Bool_t rc = kFALSE;
|
|---|
| 722 |
|
|---|
| 723 | if (IsEnvDefined(env, prefix, "HiGainWindowSize", print))
|
|---|
| 724 | {
|
|---|
| 725 | SetResolution(GetEnvValue(env, prefix, "Resolution", fResolution));
|
|---|
| 726 | rc = kTRUE;
|
|---|
| 727 | }
|
|---|
| 728 | if (IsEnvDefined(env, prefix, "LoGainWindowSize", print))
|
|---|
| 729 | {
|
|---|
| 730 | SetRiseTime(GetEnvValue(env, prefix, "RiseTime", fRiseTime));
|
|---|
| 731 | rc = kTRUE;
|
|---|
| 732 | }
|
|---|
| 733 |
|
|---|
| 734 | return MExtractTime::ReadEnv(env, prefix, print) ? kTRUE : rc;
|
|---|
| 735 | }
|
|---|
| 736 |
|
|---|
| 737 | void MExtractTimeFastSpline::Print(Option_t *o) const
|
|---|
| 738 | {
|
|---|
| 739 | *fLog << all;
|
|---|
| 740 | *fLog << GetDescriptor() << ":" << endl;
|
|---|
| 741 | *fLog << " Resolution: " << fResolution << endl;
|
|---|
| 742 | *fLog << " RiseTime: " << fRiseTime << endl;
|
|---|
| 743 | MExtractTime::Print(o);
|
|---|
| 744 | }
|
|---|