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 analysing 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 | !
|
---|
18 | ! Author(s): Hendrik Bartko, 09/2004 <mailto:hbartko@mppmu.mpg.de>
|
---|
19 | ! Author(s): Markus Gaug, 05/2004 <mailto:markus@ifae.es>
|
---|
20 | ! Author(s): Diego Tescaro, 05/2004 <mailto:tescaro@pd.infn.it>
|
---|
21 | !
|
---|
22 | ! Copyright: MAGIC Software Development, 2000-2004
|
---|
23 | !
|
---|
24 | !
|
---|
25 | \* ======================================================================== */
|
---|
26 |
|
---|
27 | //////////////////////////////////////////////////////////////////////////////
|
---|
28 | //
|
---|
29 | // MExtractTimeAndChargeDigitalFilter
|
---|
30 | //
|
---|
31 | // Hendrik has promised to write more documentation
|
---|
32 | //
|
---|
33 | // The following variables have to be set by the derived class and
|
---|
34 | // do not have defaults:
|
---|
35 | // - fNumHiGainSamples
|
---|
36 | // - fNumLoGainSamples
|
---|
37 | // - fSqrtHiGainSamples
|
---|
38 | // - fSqrtLoGainSamples
|
---|
39 | //
|
---|
40 | // The reading of automatic weights files (color, type) can be switched
|
---|
41 | // off using EnableAutomaticWeights(kFALSE).
|
---|
42 | //
|
---|
43 | // An empty name or "-" as the weights file name is a synonym for
|
---|
44 | // setting all weights to 1
|
---|
45 | //
|
---|
46 | // Input Containers:
|
---|
47 | // MRawEvtData
|
---|
48 | // MRawRunHeader
|
---|
49 | // MPedestalCam
|
---|
50 | // [MCalibrationPattern]
|
---|
51 | //
|
---|
52 | // Output Containers:
|
---|
53 | // MArrivalTimeCam
|
---|
54 | // MExtractedSignalCam
|
---|
55 | //
|
---|
56 | //////////////////////////////////////////////////////////////////////////////
|
---|
57 | #include "MExtractTimeAndChargeDigitalFilter.h"
|
---|
58 |
|
---|
59 | #include <errno.h>
|
---|
60 | #include <fstream>
|
---|
61 |
|
---|
62 | #include <TH1.h>
|
---|
63 | #include <TH2.h>
|
---|
64 | #include <TMatrix.h>
|
---|
65 |
|
---|
66 | #include "MLog.h"
|
---|
67 | #include "MLogManip.h"
|
---|
68 |
|
---|
69 | #include "MParList.h"
|
---|
70 |
|
---|
71 | #include "MRawRunHeader.h"
|
---|
72 | #include "MCalibrationPattern.h"
|
---|
73 | #include "MExtractedSignalCam.h"
|
---|
74 |
|
---|
75 | #include "MPedestalPix.h"
|
---|
76 |
|
---|
77 | ClassImp(MExtractTimeAndChargeDigitalFilter);
|
---|
78 |
|
---|
79 | using namespace std;
|
---|
80 |
|
---|
81 | const Byte_t MExtractTimeAndChargeDigitalFilter::fgHiGainFirst = 0;
|
---|
82 | const Byte_t MExtractTimeAndChargeDigitalFilter::fgHiGainLast = 15;
|
---|
83 | const Byte_t MExtractTimeAndChargeDigitalFilter::fgLoGainFirst = 1;
|
---|
84 | const Byte_t MExtractTimeAndChargeDigitalFilter::fgLoGainLast = 14;
|
---|
85 | const Int_t MExtractTimeAndChargeDigitalFilter::fgWindowSizeHiGain = 4;
|
---|
86 | const Int_t MExtractTimeAndChargeDigitalFilter::fgWindowSizeLoGain = 6;
|
---|
87 | const Int_t MExtractTimeAndChargeDigitalFilter::fgBinningResolutionHiGain = 10;
|
---|
88 | const Int_t MExtractTimeAndChargeDigitalFilter::fgBinningResolutionLoGain = 10;
|
---|
89 | const Int_t MExtractTimeAndChargeDigitalFilter::fgSignalStartBinHiGain = 4;
|
---|
90 | const Int_t MExtractTimeAndChargeDigitalFilter::fgSignalStartBinLoGain = 4;
|
---|
91 | const Float_t MExtractTimeAndChargeDigitalFilter::fgOffsetLoGain = 0.95;
|
---|
92 | const Float_t MExtractTimeAndChargeDigitalFilter::fgLoGainStartShift = -1.8;
|
---|
93 |
|
---|
94 | // --------------------------------------------------------------------------
|
---|
95 | //
|
---|
96 | // Default constructor.
|
---|
97 | //
|
---|
98 | // Calls:
|
---|
99 | // - SetWindowSize();
|
---|
100 | // - SetRange(fgHiGainFirst, fgHiGainLast, fgLoGainFirst, fgLoGainLast)
|
---|
101 | // - SetBinningResolution();
|
---|
102 | //
|
---|
103 | // Sets all weights to 1.
|
---|
104 | //
|
---|
105 | MExtractTimeAndChargeDigitalFilter::MExtractTimeAndChargeDigitalFilter(const char *name, const char *title)
|
---|
106 | : fTimeShiftHiGain(0.), fTimeShiftLoGain(0.), fAutomaticWeights(kTRUE), fRandomIter(0)
|
---|
107 | {
|
---|
108 | fName = name ? name : "MExtractTimeAndChargeDigitalFilter";
|
---|
109 | fTitle = title ? title : "Digital Filter";
|
---|
110 |
|
---|
111 | SetRange(fgHiGainFirst, fgHiGainLast, fgLoGainFirst, fgLoGainLast);
|
---|
112 | SetWindowSize();
|
---|
113 | SetBinningResolution();
|
---|
114 | SetSignalStartBin();
|
---|
115 |
|
---|
116 | SetOffsetLoGain(fgOffsetLoGain); // Order between both
|
---|
117 | SetLoGainStartShift(fgLoGainStartShift); // is important
|
---|
118 | }
|
---|
119 |
|
---|
120 | // ---------------------------------------------------------------------------------------
|
---|
121 | //
|
---|
122 | // Checks:
|
---|
123 | // - if a window is bigger than the one defined by the ranges, set it to the available range
|
---|
124 | //
|
---|
125 | // Sets:
|
---|
126 | // - fNumHiGainSamples to: (Float_t)fWindowSizeHiGain
|
---|
127 | // - fNumLoGainSamples to: (Float_t)fWindowSizeLoGain
|
---|
128 | //
|
---|
129 | void MExtractTimeAndChargeDigitalFilter::SetWindowSize(Int_t windowh, Int_t windowl)
|
---|
130 | {
|
---|
131 |
|
---|
132 | if (windowh != fgWindowSizeHiGain)
|
---|
133 | *fLog << warn << GetDescriptor()
|
---|
134 | << ": ATTENTION!!! If you are not Hendrik Bartko, do NOT use a different window size than the default." << endl;
|
---|
135 | if (windowl != fgWindowSizeLoGain)
|
---|
136 | *fLog << warn << GetDescriptor()
|
---|
137 | << ": ATTENTION!!! If you are not Hendrik Bartko, do NOT use a different window size than the default" << endl;
|
---|
138 |
|
---|
139 | fWindowSizeHiGain = windowh;
|
---|
140 | fWindowSizeLoGain = windowl;
|
---|
141 |
|
---|
142 | const Int_t availhirange = (Int_t)(fHiGainLast-fHiGainFirst+1);
|
---|
143 |
|
---|
144 | if (fWindowSizeHiGain > availhirange)
|
---|
145 | {
|
---|
146 | *fLog << warn << GetDescriptor() << ": Hi Gain window size: " << Form("%2i",fWindowSizeHiGain);
|
---|
147 | *fLog << " is bigger than available range: [" << Form("%2i", (int)fHiGainFirst);
|
---|
148 | *fLog << "," << Form("%21", (int)fHiGainLast) << "]" << endl;
|
---|
149 |
|
---|
150 | fHiGainLast = fHiGainFirst + fWindowSizeHiGain;
|
---|
151 |
|
---|
152 | *fLog << warn << GetDescriptor() << ": Will set the upper range to: " << (int)fHiGainLast << endl;
|
---|
153 | }
|
---|
154 |
|
---|
155 | if (fWindowSizeHiGain < 2)
|
---|
156 | {
|
---|
157 | fWindowSizeHiGain = 2;
|
---|
158 | *fLog << warn << GetDescriptor() << ": High Gain window size set to two samples" << endl;
|
---|
159 | }
|
---|
160 |
|
---|
161 | if (fLoGainLast != 0 && fWindowSizeLoGain != 0)
|
---|
162 | {
|
---|
163 | const Int_t availlorange = (Int_t)(fLoGainLast-fLoGainFirst+1);
|
---|
164 |
|
---|
165 | if (fWindowSizeLoGain > availlorange)
|
---|
166 | {
|
---|
167 | *fLog << warn << GetDescriptor() << ": Lo Gain window size: " << Form("%2i",fWindowSizeLoGain);
|
---|
168 | *fLog << " is bigger than available range: [" << Form("%2i", (int)fLoGainFirst);
|
---|
169 | *fLog << "," << Form("%21", (int)fLoGainLast) << "]" << endl;
|
---|
170 |
|
---|
171 | fLoGainLast = fLoGainFirst + fWindowSizeLoGain;
|
---|
172 |
|
---|
173 | *fLog << warn << GetDescriptor() << ": Will set the upper range to: " << (int)fLoGainLast << endl;
|
---|
174 | }
|
---|
175 |
|
---|
176 | if (fWindowSizeLoGain<2)
|
---|
177 | {
|
---|
178 | fWindowSizeLoGain = 2;
|
---|
179 | *fLog << warn << GetDescriptor() << ": Low Gain window size set to two samples" << endl;
|
---|
180 | }
|
---|
181 | }
|
---|
182 | //
|
---|
183 | // We need here the effective number of samples which is about 2.5 in the case of a window
|
---|
184 | // size of 6. The exact numbers have to be found still.
|
---|
185 | //
|
---|
186 | fNumHiGainSamples = (Float_t)fWindowSizeHiGain/2.4;
|
---|
187 | fNumLoGainSamples = (Float_t)fWindowSizeLoGain/2.4;
|
---|
188 | fSqrtHiGainSamples = TMath::Sqrt(fNumHiGainSamples);
|
---|
189 | fSqrtLoGainSamples = TMath::Sqrt(fNumLoGainSamples);
|
---|
190 |
|
---|
191 | }
|
---|
192 |
|
---|
193 | // --------------------------------------------------------------------------
|
---|
194 | //
|
---|
195 | // Executing MExtractTimeAndCharge::PreProcess and searching for
|
---|
196 | // MCalibrationPattern
|
---|
197 | //
|
---|
198 | Int_t MExtractTimeAndChargeDigitalFilter::PreProcess(MParList *pList)
|
---|
199 | {
|
---|
200 | if (!MExtractTimeAndCharge::PreProcess(pList))
|
---|
201 | return kFALSE;
|
---|
202 |
|
---|
203 | fCalibPattern = (MCalibrationPattern*)pList->FindObject("MCalibrationPattern");
|
---|
204 | return kTRUE;
|
---|
205 | }
|
---|
206 |
|
---|
207 | // --------------------------------------------------------------------------
|
---|
208 | //
|
---|
209 | // The weights are determined using GetAutimaticWeights().
|
---|
210 | //
|
---|
211 | // kFALSE is returned if it returned an error.
|
---|
212 | // kTRUE is returned if no new weights were set.
|
---|
213 | //
|
---|
214 | // If new weights are set
|
---|
215 | // fTimeShiftHiGain
|
---|
216 | // fTimeShiftLoGain
|
---|
217 | // fNumHiGainSamples
|
---|
218 | // fNumLoGainSamples
|
---|
219 | // fSqrtHiGainSamples
|
---|
220 | // fSqrtLoGainSamples
|
---|
221 | // and
|
---|
222 | // fSignals->SetUsedFADCSlices(...)
|
---|
223 | // is updated accordingly.
|
---|
224 | //
|
---|
225 | Bool_t MExtractTimeAndChargeDigitalFilter::GetWeights()
|
---|
226 | {
|
---|
227 | switch (GetAutomaticWeights())
|
---|
228 | {
|
---|
229 | case kERROR: // An error occured
|
---|
230 | return kFALSE;
|
---|
231 | case kFALSE: // No new weights set
|
---|
232 | return kTRUE;
|
---|
233 | }
|
---|
234 |
|
---|
235 | //
|
---|
236 | // shift the times back to the right reference (start counting from 0)
|
---|
237 | //
|
---|
238 | // The high-gain has the extraction offset (fHiGainFirst) already included
|
---|
239 | // here for speed reason. The low-gain has a changing extraction offset,
|
---|
240 | // so it will be added at every event (in FindTimeAndChargeLoGain)
|
---|
241 | fTimeShiftHiGain = 0.5 + 1./fBinningResolutionHiGain;
|
---|
242 | fTimeShiftLoGain = 0.5 + 1./fBinningResolutionLoGain;
|
---|
243 |
|
---|
244 | //
|
---|
245 | // We need here the effective number of samples which is about 2.5 in the case of a window
|
---|
246 | // size of 6. The exact numbers have to be found still.
|
---|
247 | //
|
---|
248 | fNumHiGainSamples = (Float_t)fWindowSizeHiGain/2.4;
|
---|
249 | fNumLoGainSamples = (Float_t)fWindowSizeLoGain/2.4;
|
---|
250 | fSqrtHiGainSamples = TMath::Sqrt(fNumHiGainSamples);
|
---|
251 | fSqrtLoGainSamples = TMath::Sqrt(fNumLoGainSamples);
|
---|
252 |
|
---|
253 | // From MExtractTimeAndCharge::ReInit
|
---|
254 | if (fSignals)
|
---|
255 | fSignals->SetUsedFADCSlices(fHiGainFirst, fHiGainLast+fHiLoLast, fNumHiGainSamples,
|
---|
256 | fLoGainFirst, fLoGainLast, fNumLoGainSamples);
|
---|
257 | return kTRUE;
|
---|
258 | }
|
---|
259 |
|
---|
260 | // --------------------------------------------------------------------------
|
---|
261 | //
|
---|
262 | // InitArrays
|
---|
263 | //
|
---|
264 | // Gets called in the ReInit() and initialized the arrays
|
---|
265 | //
|
---|
266 | Bool_t MExtractTimeAndChargeDigitalFilter::InitArrays()
|
---|
267 | {
|
---|
268 | if (!fRunHeader)
|
---|
269 | return kFALSE;
|
---|
270 |
|
---|
271 | const Int_t rangehi = (Int_t)(fHiGainLast - fHiGainFirst + 1 + fHiLoLast);
|
---|
272 | const Int_t rangelo = (Int_t)(fLoGainLast - fLoGainFirst + 1);
|
---|
273 |
|
---|
274 | fHiGainSignal.Set(rangehi);
|
---|
275 | fLoGainSignal.Set(rangelo);
|
---|
276 |
|
---|
277 | return GetWeights();
|
---|
278 | }
|
---|
279 |
|
---|
280 | // --------------------------------------------------------------------------
|
---|
281 | //
|
---|
282 | // Check if reading a new weights file is necessary because the calibration
|
---|
283 | // pattern has changed. (Cannot be done in ReInit, because at this time
|
---|
284 | // the calibration pattern is not available.
|
---|
285 | // Then process the event.
|
---|
286 | //
|
---|
287 | Int_t MExtractTimeAndChargeDigitalFilter::Process()
|
---|
288 | {
|
---|
289 | // Change Weights if the calibration patter changes
|
---|
290 | if (!GetWeights())
|
---|
291 | return kERROR;
|
---|
292 |
|
---|
293 | // Process event
|
---|
294 | return MExtractTimeAndCharge::Process();
|
---|
295 | }
|
---|
296 |
|
---|
297 | void MExtractTimeAndChargeDigitalFilter::CalcBinningResArrays()
|
---|
298 | {
|
---|
299 |
|
---|
300 | fArrBinningResHiGain.Set(fWindowSizeHiGain);
|
---|
301 | fArrBinningResHalfHiGain.Set(fWindowSizeHiGain);
|
---|
302 |
|
---|
303 | for (int i=0; i<fWindowSizeHiGain; i++)
|
---|
304 | {
|
---|
305 | fArrBinningResHiGain[i] = fBinningResolutionHiGain*i;
|
---|
306 | fArrBinningResHalfHiGain[i] = fArrBinningResHiGain[i] + fBinningResolutionHalfHiGain;
|
---|
307 | }
|
---|
308 |
|
---|
309 | fArrBinningResLoGain.Set(fWindowSizeLoGain);
|
---|
310 | fArrBinningResHalfLoGain.Set(fWindowSizeLoGain);
|
---|
311 |
|
---|
312 | for (int i=0; i<fWindowSizeLoGain; i++)
|
---|
313 | {
|
---|
314 | fArrBinningResLoGain[i] = fBinningResolutionLoGain*i;
|
---|
315 | fArrBinningResHalfLoGain[i] = fArrBinningResLoGain[i] + fBinningResolutionHalfLoGain;
|
---|
316 | }
|
---|
317 | }
|
---|
318 |
|
---|
319 | // --------------------------------------------------------------------------
|
---|
320 | //
|
---|
321 | // Apply the digital filter algorithm to the high-gain slices.
|
---|
322 | //
|
---|
323 | void MExtractTimeAndChargeDigitalFilter::FindTimeAndChargeHiGain(Byte_t *ptr, Byte_t *logain, Float_t &sum, Float_t &dsum,
|
---|
324 | Float_t &time, Float_t &dtime,
|
---|
325 | Byte_t &sat, const MPedestalPix &ped, const Bool_t abflag)
|
---|
326 | {
|
---|
327 |
|
---|
328 | Int_t range = fHiGainLast - fHiGainFirst + 1;
|
---|
329 |
|
---|
330 | const Byte_t *end = ptr + range;
|
---|
331 | Byte_t *p = ptr;
|
---|
332 |
|
---|
333 | //
|
---|
334 | // Preparations for the pedestal subtraction (with AB-noise correction)
|
---|
335 | //
|
---|
336 | const Float_t pedes = ped.GetPedestal();
|
---|
337 | const Float_t ABoffs = ped.GetPedestalABoffset();
|
---|
338 |
|
---|
339 | const Float_t pedmean[2] = { pedes + ABoffs, pedes - ABoffs };
|
---|
340 |
|
---|
341 | range += fHiLoLast;
|
---|
342 | fMaxBinContent = 0;
|
---|
343 | //
|
---|
344 | // Check for saturation in all other slices
|
---|
345 | //
|
---|
346 | Int_t ids = fHiGainFirst;
|
---|
347 | Float_t *sample = fHiGainSignal.GetArray();
|
---|
348 | while (p<end)
|
---|
349 | {
|
---|
350 | *sample++ = (Float_t)*p - pedmean[(ids++ + abflag) & 0x1];
|
---|
351 |
|
---|
352 | if (*p > fMaxBinContent)
|
---|
353 | {
|
---|
354 | Byte_t maxpos = p-ptr;
|
---|
355 |
|
---|
356 | // range-fWindowSizeHiGain+1 == fHiLoLast isn't it?
|
---|
357 | if (maxpos > 1 && maxpos < (range - fWindowSizeHiGain + 1))
|
---|
358 | fMaxBinContent = *p;
|
---|
359 | }
|
---|
360 |
|
---|
361 | if (*p++ >= fSaturationLimit)
|
---|
362 | if (!sat)
|
---|
363 | sat = ids-fHiGainFirst;
|
---|
364 | }
|
---|
365 |
|
---|
366 | //
|
---|
367 | // Slide with a window of size fWindowSizeHiGain over the sample
|
---|
368 | // and multiply the entries with the corresponding weights
|
---|
369 | //
|
---|
370 | if (IsNoiseCalculation())
|
---|
371 | {
|
---|
372 | if (fRandomIter == fBinningResolutionHiGain)
|
---|
373 | fRandomIter = 0;
|
---|
374 | for (Int_t ids=0; ids < fWindowSizeHiGain; ids++)
|
---|
375 | {
|
---|
376 | const Int_t idx = fArrBinningResHiGain[ids] + fRandomIter;
|
---|
377 | sum += fAmpWeightsHiGain [idx]*fHiGainSignal[ids];
|
---|
378 | }
|
---|
379 | fRandomIter++;
|
---|
380 | return;
|
---|
381 | }
|
---|
382 |
|
---|
383 | if (fHiLoLast != 0)
|
---|
384 | {
|
---|
385 |
|
---|
386 | end = logain + fHiLoLast;
|
---|
387 |
|
---|
388 | while (logain<end)
|
---|
389 | {
|
---|
390 |
|
---|
391 | *sample++ = (Float_t)*logain - pedmean[(ids++ + abflag) & 0x1];
|
---|
392 |
|
---|
393 | if (*logain++ >= fSaturationLimit)
|
---|
394 | if (!sat)
|
---|
395 | sat = ids-fHiGainFirst;
|
---|
396 | }
|
---|
397 | }
|
---|
398 |
|
---|
399 | //
|
---|
400 | // allow no saturated slice
|
---|
401 | //
|
---|
402 | // if (sat > 0)
|
---|
403 | // return;
|
---|
404 |
|
---|
405 | Float_t time_sum = 0.;
|
---|
406 | Float_t fmax = -FLT_MAX;
|
---|
407 | Float_t ftime_max = 0.;
|
---|
408 | Int_t max_p = -1;
|
---|
409 |
|
---|
410 | //
|
---|
411 | // Calculate the sum of the first fWindowSize slices
|
---|
412 | //
|
---|
413 | for (Int_t i=0;i<range-fWindowSizeHiGain+1;i++)
|
---|
414 | {
|
---|
415 | sum = 0.;
|
---|
416 | time_sum = 0.;
|
---|
417 |
|
---|
418 | //
|
---|
419 | // Slide with a window of size fWindowSizeHiGain over the sample
|
---|
420 | // and multiply the entries with the corresponding weights
|
---|
421 | //
|
---|
422 | for (Int_t sample=0; sample < fWindowSizeHiGain; sample++)
|
---|
423 | {
|
---|
424 | const Int_t idx = fBinningResolutionHiGain*sample+fBinningResolutionHalfHiGain;
|
---|
425 | const Float_t pex = fHiGainSignal[sample+i];
|
---|
426 | sum += fAmpWeightsHiGain [idx]*pex;
|
---|
427 | time_sum += fTimeWeightsHiGain[idx]*pex;
|
---|
428 | }
|
---|
429 |
|
---|
430 | if (sum>fmax)
|
---|
431 | {
|
---|
432 | fmax = sum;
|
---|
433 | ftime_max = time_sum;
|
---|
434 | max_p = i;
|
---|
435 | }
|
---|
436 | } /* for (Int_t i=0;i<range-fWindowSizeHiGain+1;i++) */
|
---|
437 |
|
---|
438 | if (max_p<0)
|
---|
439 | {
|
---|
440 | sum = 0;
|
---|
441 | time = -1;
|
---|
442 | dsum = -1;
|
---|
443 | dtime = -1;
|
---|
444 | return;
|
---|
445 | }
|
---|
446 |
|
---|
447 | if (fmax==0)
|
---|
448 | {
|
---|
449 | sum = 0;
|
---|
450 | time = -1;
|
---|
451 | dsum = 0;
|
---|
452 | dtime = 0;
|
---|
453 | return;
|
---|
454 | }
|
---|
455 |
|
---|
456 | ftime_max /= fmax;
|
---|
457 | Int_t t_iter = Int_t(ftime_max*fBinningResolutionHiGain);
|
---|
458 | Int_t sample_iter = 0;
|
---|
459 |
|
---|
460 | while ( t_iter > fBinningResolutionHalfHiGain-1 || t_iter < -fBinningResolutionHalfHiGain )
|
---|
461 | {
|
---|
462 | if (t_iter > fBinningResolutionHalfHiGain-1)
|
---|
463 | {
|
---|
464 | t_iter -= fBinningResolutionHiGain;
|
---|
465 | max_p--;
|
---|
466 | sample_iter--;
|
---|
467 | }
|
---|
468 | if (t_iter < -fBinningResolutionHalfHiGain)
|
---|
469 | {
|
---|
470 | t_iter += fBinningResolutionHiGain;
|
---|
471 | max_p++;
|
---|
472 | sample_iter++;
|
---|
473 | }
|
---|
474 | }
|
---|
475 |
|
---|
476 | sum = 0.;
|
---|
477 | time_sum = 0.;
|
---|
478 | if (max_p < 0)
|
---|
479 | max_p = 0;
|
---|
480 | if (max_p+fWindowSizeHiGain > range)
|
---|
481 | max_p = range-fWindowSizeHiGain;
|
---|
482 | //
|
---|
483 | // Slide with a window of size fWindowSizeHiGain over the sample
|
---|
484 | // and multiply the entries with the corresponding weights
|
---|
485 | //
|
---|
486 | for (Int_t sample=0; sample < fWindowSizeHiGain; sample++)
|
---|
487 | {
|
---|
488 | const Int_t idx = fArrBinningResHalfHiGain[sample] + t_iter;
|
---|
489 | const Int_t ids = max_p + sample;
|
---|
490 | const Float_t pex = fHiGainSignal[ids];
|
---|
491 | sum += fAmpWeightsHiGain [idx]*pex;
|
---|
492 | time_sum += fTimeWeightsHiGain[idx]*pex;
|
---|
493 | }
|
---|
494 |
|
---|
495 | if (sum == 0)
|
---|
496 | {
|
---|
497 | time = -1;
|
---|
498 | dsum = 0;
|
---|
499 | dtime = 0;
|
---|
500 | return;
|
---|
501 | }
|
---|
502 |
|
---|
503 | // here, the first high-gain slice is already included in the fTimeShiftHiGain
|
---|
504 | // time = fTimeShiftHiGain + max_p - Float_t(t_iter)/fBinningResolutionHiGain;
|
---|
505 | time = max_p + fTimeShiftHiGain + (Float_t)fHiGainFirst /* this shifts the time to the start of the rising edge */
|
---|
506 | - ((Float_t)t_iter)/fBinningResolutionHiGain;
|
---|
507 |
|
---|
508 | const Float_t timefineadjust = time_sum/sum;
|
---|
509 |
|
---|
510 | if (TMath::Abs(timefineadjust) < 4./fBinningResolutionHiGain)
|
---|
511 | time -= timefineadjust;
|
---|
512 |
|
---|
513 | }
|
---|
514 |
|
---|
515 | // --------------------------------------------------------------------------
|
---|
516 | //
|
---|
517 | // Apply the digital filter algorithm to the low-gain slices.
|
---|
518 | //
|
---|
519 | void MExtractTimeAndChargeDigitalFilter::FindTimeAndChargeLoGain(Byte_t *ptr, Float_t &sum, Float_t &dsum,
|
---|
520 | Float_t &time, Float_t &dtime,
|
---|
521 | Byte_t &sat, const MPedestalPix &ped, const Bool_t abflag)
|
---|
522 | {
|
---|
523 |
|
---|
524 | const Int_t range = fLoGainLast - fLoGainFirst + 1;
|
---|
525 |
|
---|
526 | const Byte_t *end = ptr + range;
|
---|
527 | Byte_t *p = ptr;
|
---|
528 | //
|
---|
529 | // Prepare the low-gain pedestal
|
---|
530 | //
|
---|
531 | const Float_t pedes = ped.GetPedestal();
|
---|
532 | const Float_t ABoffs = ped.GetPedestalABoffset();
|
---|
533 |
|
---|
534 | const Float_t pedmean[2] = { pedes + ABoffs, pedes - ABoffs };
|
---|
535 |
|
---|
536 | //
|
---|
537 | // Check for saturation in all other slices
|
---|
538 | //
|
---|
539 | Float_t *sample = fLoGainSignal.GetArray();
|
---|
540 | Int_t ids = fLoGainFirst;
|
---|
541 | while (p<end)
|
---|
542 | {
|
---|
543 | *sample++ = (Float_t)*p - pedmean[(ids++ + abflag) & 0x1];
|
---|
544 |
|
---|
545 | if (*p++ >= fSaturationLimit)
|
---|
546 | sat++;
|
---|
547 | }
|
---|
548 |
|
---|
549 | //
|
---|
550 | // Slide with a window of size fWindowSizeLoGain over the sample
|
---|
551 | // and multiply the entries with the corresponding weights
|
---|
552 | //
|
---|
553 | if (IsNoiseCalculation())
|
---|
554 | {
|
---|
555 | if (fRandomIter == fBinningResolutionLoGain)
|
---|
556 | fRandomIter = 0;
|
---|
557 | for (Int_t ids=0; ids < fWindowSizeLoGain; ids++)
|
---|
558 | {
|
---|
559 | const Int_t idx = fArrBinningResLoGain[ids] + fRandomIter;
|
---|
560 | sum += fAmpWeightsLoGain [idx]*fLoGainSignal[ids];
|
---|
561 | }
|
---|
562 | return;
|
---|
563 | }
|
---|
564 |
|
---|
565 | Float_t time_sum = 0.;
|
---|
566 | Float_t fmax = -FLT_MAX;
|
---|
567 | Float_t ftime_max = 0.;
|
---|
568 | Int_t max_p = -1;
|
---|
569 |
|
---|
570 | //
|
---|
571 | // Calculate the sum of the first fWindowSize slices
|
---|
572 | //
|
---|
573 | for (Int_t i=0;i<range-fWindowSizeLoGain+1;i++)
|
---|
574 | {
|
---|
575 | sum = 0.;
|
---|
576 | time_sum = 0.;
|
---|
577 |
|
---|
578 | //
|
---|
579 | // Slide with a window of size fWindowSizeLoGain over the sample
|
---|
580 | // and multiply the entries with the corresponding weights
|
---|
581 | //
|
---|
582 | for (Int_t sample=0; sample < fWindowSizeLoGain; sample++)
|
---|
583 | {
|
---|
584 | const Int_t idx = fArrBinningResHalfLoGain[sample];
|
---|
585 | const Float_t pex = fLoGainSignal[sample+i];
|
---|
586 | sum += fAmpWeightsLoGain [idx]*pex;
|
---|
587 | time_sum += fTimeWeightsLoGain[idx]*pex;
|
---|
588 | }
|
---|
589 |
|
---|
590 | if (sum>fmax)
|
---|
591 | {
|
---|
592 | fmax = sum;
|
---|
593 | ftime_max = time_sum;
|
---|
594 | max_p = i;
|
---|
595 | }
|
---|
596 | } /* for (Int_t i=0;i<range-fWindowSizeLoGain+1;i++) */
|
---|
597 |
|
---|
598 | if (max_p<0)
|
---|
599 | {
|
---|
600 | sum = 0;
|
---|
601 | time = -1;
|
---|
602 | dsum = -1;
|
---|
603 | dtime = -1;
|
---|
604 | return;
|
---|
605 | }
|
---|
606 |
|
---|
607 | if (fmax==0)
|
---|
608 | {
|
---|
609 | sum = 0;
|
---|
610 | time = -1;
|
---|
611 | dsum = 0;
|
---|
612 | dtime = 0;
|
---|
613 | return;
|
---|
614 | }
|
---|
615 |
|
---|
616 | ftime_max /= fmax;
|
---|
617 | Int_t t_iter = Int_t(ftime_max*fBinningResolutionLoGain);
|
---|
618 | Int_t sample_iter = 0;
|
---|
619 |
|
---|
620 | while ( t_iter > fBinningResolutionHalfLoGain-1 || t_iter < -fBinningResolutionHalfLoGain )
|
---|
621 | {
|
---|
622 | if (t_iter > fBinningResolutionHalfLoGain-1)
|
---|
623 | {
|
---|
624 | t_iter -= fBinningResolutionLoGain;
|
---|
625 | max_p--;
|
---|
626 | sample_iter--;
|
---|
627 | }
|
---|
628 | if (t_iter < -fBinningResolutionHalfLoGain)
|
---|
629 | {
|
---|
630 | t_iter += fBinningResolutionLoGain;
|
---|
631 | max_p++;
|
---|
632 | sample_iter++;
|
---|
633 | }
|
---|
634 | }
|
---|
635 |
|
---|
636 | sum = 0.;
|
---|
637 | time_sum = 0.;
|
---|
638 |
|
---|
639 | //
|
---|
640 | // Slide with a window of size fWindowSizeLoGain over the sample
|
---|
641 | // and multiply the entries with the corresponding weights
|
---|
642 | //
|
---|
643 | for (Int_t sample=0; sample < fWindowSizeLoGain; sample++)
|
---|
644 | {
|
---|
645 | const Int_t idx = fArrBinningResHalfLoGain[sample] + t_iter;
|
---|
646 | const Int_t ids = max_p + sample;
|
---|
647 | const Float_t pex = ids < 0 ? 0. : ( ids >= range ? 0. : fLoGainSignal[ids]);
|
---|
648 | sum += fAmpWeightsLoGain [idx]*pex;
|
---|
649 | time_sum += fTimeWeightsLoGain[idx]*pex;
|
---|
650 | }
|
---|
651 |
|
---|
652 | if (sum == 0)
|
---|
653 | {
|
---|
654 | time = -1;
|
---|
655 | dsum = 0;
|
---|
656 | dtime = 0;
|
---|
657 | return;
|
---|
658 | }
|
---|
659 |
|
---|
660 | time = max_p + fTimeShiftLoGain + (Float_t)fLoGainFirst /* this shifts the time to the start of the rising edge */
|
---|
661 | - ((Float_t)t_iter)/fBinningResolutionLoGain;
|
---|
662 |
|
---|
663 | const Float_t timefineadjust = time_sum/sum;
|
---|
664 |
|
---|
665 | if (TMath::Abs(timefineadjust) < 4./fBinningResolutionLoGain)
|
---|
666 | time -= timefineadjust;
|
---|
667 |
|
---|
668 | }
|
---|
669 |
|
---|
670 | // --------------------------------------------------------------------------
|
---|
671 | //
|
---|
672 | // Read the setup from a TEnv, eg:
|
---|
673 | // MJPedestal.MExtractor.WindowSizeHiGain: 6
|
---|
674 | // MJPedestal.MExtractor.WindowSizeLoGain: 6
|
---|
675 | // MJPedestal.MExtractor.BinningResolutionHiGain: 10
|
---|
676 | // MJPedestal.MExtractor.BinningResolutionLoGain: 10
|
---|
677 | // MJPedestal.MExtractor.WeightsFile: filename
|
---|
678 | // MJPedestal.MExtractor.AutomaticWeights: off
|
---|
679 | //
|
---|
680 | Int_t MExtractTimeAndChargeDigitalFilter::ReadEnv(const TEnv &env, TString prefix, Bool_t print)
|
---|
681 | {
|
---|
682 |
|
---|
683 | Byte_t hw = fWindowSizeHiGain;
|
---|
684 | Byte_t lw = fWindowSizeLoGain;
|
---|
685 | Bool_t rc = kFALSE;
|
---|
686 |
|
---|
687 | if (IsEnvDefined(env, prefix, "AutomaticWeights", print))
|
---|
688 | {
|
---|
689 | EnableAutomaticWeights(GetEnvValue(env, prefix, "AutomaticWeights", fAutomaticWeights));
|
---|
690 | rc = kTRUE;
|
---|
691 | }
|
---|
692 |
|
---|
693 | if (IsEnvDefined(env, prefix, "WindowSizeHiGain", print))
|
---|
694 | {
|
---|
695 | hw = GetEnvValue(env, prefix, "WindowSizeHiGain", hw);
|
---|
696 | rc = kTRUE;
|
---|
697 | }
|
---|
698 | if (IsEnvDefined(env, prefix, "WindowSizeLoGain", print))
|
---|
699 | {
|
---|
700 | lw = GetEnvValue(env, prefix, "WindowSizeLoGain", lw);
|
---|
701 | rc = kTRUE;
|
---|
702 | }
|
---|
703 |
|
---|
704 | if (rc)
|
---|
705 | SetWindowSize(hw, lw);
|
---|
706 |
|
---|
707 | Bool_t rc2 = kFALSE;
|
---|
708 | Int_t brh = fBinningResolutionHiGain;
|
---|
709 | Int_t brl = fBinningResolutionLoGain;
|
---|
710 |
|
---|
711 | if (IsEnvDefined(env, prefix, "BinningResolutionHiGain", print))
|
---|
712 | {
|
---|
713 | brh = GetEnvValue(env, prefix, brh);
|
---|
714 | rc2 = kTRUE;
|
---|
715 | }
|
---|
716 | if (IsEnvDefined(env, prefix, "BinningResolutionLoGain", print))
|
---|
717 | {
|
---|
718 | brl = GetEnvValue(env, prefix, brl);
|
---|
719 | rc2 = kTRUE;
|
---|
720 | }
|
---|
721 |
|
---|
722 | if (rc2)
|
---|
723 | {
|
---|
724 | SetBinningResolution(brh, brl);
|
---|
725 | rc = kTRUE;
|
---|
726 | }
|
---|
727 |
|
---|
728 | if (IsEnvDefined(env, prefix, "WeightsFile", print))
|
---|
729 | {
|
---|
730 | SetNameWeightsFile(GetEnvValue(env, prefix, "WeightsFile", ""));
|
---|
731 | rc = kTRUE;
|
---|
732 | }
|
---|
733 |
|
---|
734 | return MExtractTimeAndCharge::ReadEnv(env, prefix, print) ? kTRUE : rc;
|
---|
735 | }
|
---|
736 |
|
---|
737 | //----------------------------------------------------------------------------
|
---|
738 | //
|
---|
739 | // If automatic weights are requested, no default weights (name.IsNull())
|
---|
740 | // are requested, fRunHeader is available and fRunHeader->IsMonteCarloRun()
|
---|
741 | // is true prepend "MC_" in from of the name.
|
---|
742 | //
|
---|
743 | // return poth+name;
|
---|
744 | //
|
---|
745 | TString MExtractTimeAndChargeDigitalFilter::CompileWeightFileName(TString path, const TString &name) const
|
---|
746 | {
|
---|
747 | if (fAutomaticWeights && !name.IsNull() && fRunHeader && fRunHeader->IsMonteCarloRun())
|
---|
748 | path += "MC_";
|
---|
749 |
|
---|
750 | path += name;
|
---|
751 |
|
---|
752 | return path;
|
---|
753 | }
|
---|
754 |
|
---|
755 | //----------------------------------------------------------------------------
|
---|
756 | //
|
---|
757 | // Read a pre-defined weights file into the class.
|
---|
758 | // This is mandatory for the extraction
|
---|
759 | //
|
---|
760 | // If filenname is empty, then all weights will be set to 1.
|
---|
761 | //
|
---|
762 | // Returns:
|
---|
763 | // kTRUE: new weights set
|
---|
764 | // kFALSE: no weights set
|
---|
765 | // kERROR: error
|
---|
766 | //
|
---|
767 | Int_t MExtractTimeAndChargeDigitalFilter::ReadWeightsFile(TString filename, TString path)
|
---|
768 | {
|
---|
769 | if (filename.IsNull())
|
---|
770 | {
|
---|
771 | fAmpWeightsHiGain .Set(fBinningResolutionHiGain*fWindowSizeHiGain);
|
---|
772 | fAmpWeightsLoGain .Set(fBinningResolutionLoGain*fWindowSizeLoGain);
|
---|
773 | fTimeWeightsHiGain.Set(fBinningResolutionHiGain*fWindowSizeHiGain);
|
---|
774 | fTimeWeightsLoGain.Set(fBinningResolutionLoGain*fWindowSizeLoGain);
|
---|
775 |
|
---|
776 | fAmpWeightsHiGain.Reset(1);
|
---|
777 | fTimeWeightsHiGain.Reset(1);
|
---|
778 | fAmpWeightsLoGain.Reset(1);
|
---|
779 | fTimeWeightsLoGain.Reset(1);
|
---|
780 | return kTRUE;
|
---|
781 | }
|
---|
782 |
|
---|
783 | // Add "MC_" in front of the filename if necessary
|
---|
784 | filename = CompileWeightFileName(path, filename);
|
---|
785 |
|
---|
786 | //filename = MJob::ExpandPath(filename);
|
---|
787 |
|
---|
788 | if (fNameWeightsFileSet==filename)
|
---|
789 | return kFALSE; // No file read
|
---|
790 |
|
---|
791 | ifstream fin(filename.Data());
|
---|
792 | if (!fin)
|
---|
793 | {
|
---|
794 | *fLog << err << GetDescriptor() << ": ERROR - Cannot open file " << filename << ": ";
|
---|
795 | *fLog << strerror(errno) << endl;
|
---|
796 | return kERROR;
|
---|
797 | }
|
---|
798 |
|
---|
799 | *fLog << all << GetDescriptor() << ": Reading weights in " << filename << "..." << flush;
|
---|
800 |
|
---|
801 | Int_t len = 0;
|
---|
802 | Int_t cnt = 0;
|
---|
803 | Int_t line = 0;
|
---|
804 | Bool_t hi = kFALSE;
|
---|
805 | Bool_t lo = kFALSE;
|
---|
806 |
|
---|
807 | TString str;
|
---|
808 |
|
---|
809 | while (1)
|
---|
810 | {
|
---|
811 | str.ReadLine(fin);
|
---|
812 | if (!fin)
|
---|
813 | break;
|
---|
814 |
|
---|
815 | line++;
|
---|
816 |
|
---|
817 | if (str.Contains("# High Gain Weights:"))
|
---|
818 | {
|
---|
819 | if (hi)
|
---|
820 | {
|
---|
821 | *fLog << err << "ERROR - 'High Gain Weights' found twice in line #" << line << "." << endl;
|
---|
822 | return kERROR;
|
---|
823 | }
|
---|
824 |
|
---|
825 | if (2!=sscanf(str.Data(), "# High Gain Weights:%2i %2i", &fWindowSizeHiGain, &fBinningResolutionHiGain))
|
---|
826 | {
|
---|
827 | *fLog << err << "ERROR - Wrong number of arguments in line #" << line << ":" << endl;
|
---|
828 | *fLog << str << endl;
|
---|
829 | return kERROR;
|
---|
830 | }
|
---|
831 |
|
---|
832 | len = fBinningResolutionHiGain*fWindowSizeHiGain;
|
---|
833 | fAmpWeightsHiGain .Set(len);
|
---|
834 | fTimeWeightsHiGain.Set(len);
|
---|
835 | hi = kTRUE;
|
---|
836 | continue;
|
---|
837 | }
|
---|
838 |
|
---|
839 | if (str.Contains("# Low Gain Weights:"))
|
---|
840 | {
|
---|
841 | if (lo)
|
---|
842 | {
|
---|
843 | *fLog << err << "ERROR - 'Lo Gain Weights' found twice in line #" << line << "." << endl;
|
---|
844 | return kERROR;
|
---|
845 | }
|
---|
846 |
|
---|
847 | if (2!=sscanf(str.Data(),"# Low Gain Weights:%2i %2i", &fWindowSizeLoGain, &fBinningResolutionLoGain))
|
---|
848 | {
|
---|
849 | *fLog << err << "ERROR - Wrong number of arguments in line #" << line << ":" << endl;
|
---|
850 | *fLog << str << endl;
|
---|
851 | return kERROR;
|
---|
852 | }
|
---|
853 |
|
---|
854 | len = fBinningResolutionLoGain*fWindowSizeLoGain;
|
---|
855 | fAmpWeightsLoGain .Set(len);
|
---|
856 | fTimeWeightsLoGain.Set(len);
|
---|
857 | lo = kTRUE;
|
---|
858 | continue;
|
---|
859 | }
|
---|
860 |
|
---|
861 | // Handle lines with comments
|
---|
862 | if (str.Contains("#"))
|
---|
863 | continue;
|
---|
864 |
|
---|
865 | // Nothing found so far
|
---|
866 | if (len == 0)
|
---|
867 | continue;
|
---|
868 |
|
---|
869 | if (2!=sscanf(str.Data(), "%f %f",
|
---|
870 | lo ? &fAmpWeightsLoGain [cnt] : &fAmpWeightsHiGain [cnt],
|
---|
871 | lo ? &fTimeWeightsLoGain[cnt] : &fTimeWeightsHiGain[cnt]))
|
---|
872 | {
|
---|
873 | *fLog << err << "ERROR - Wrong number of arguments in line #" << line << ":" << endl;
|
---|
874 | *fLog << str << endl;
|
---|
875 | return kERROR;
|
---|
876 | }
|
---|
877 |
|
---|
878 | if (++cnt == len)
|
---|
879 | {
|
---|
880 | len = 0;
|
---|
881 | cnt = 0;
|
---|
882 | }
|
---|
883 | }
|
---|
884 |
|
---|
885 | if (cnt != len)
|
---|
886 | {
|
---|
887 | *fLog << err << "ERROR - Size mismatch in weights file " << filename << endl;
|
---|
888 | return kERROR;
|
---|
889 | }
|
---|
890 |
|
---|
891 | if (!hi)
|
---|
892 | {
|
---|
893 | *fLog << err << "ERROR - No correct header found in weights file " << filename << endl;
|
---|
894 | return kERROR;
|
---|
895 | }
|
---|
896 |
|
---|
897 | *fLog << "done." << endl;
|
---|
898 |
|
---|
899 | *fLog << inf << " File contains " << fWindowSizeHiGain << " hi-gain slices ";
|
---|
900 | *fLog << "with a resolution of " << fBinningResolutionHiGain << endl;
|
---|
901 |
|
---|
902 | *fLog << inf << " File contains " << fWindowSizeLoGain << " lo-gain slices ";
|
---|
903 | *fLog << "with a resolution of " << fBinningResolutionLoGain << endl;
|
---|
904 |
|
---|
905 | CalcBinningResArrays();
|
---|
906 |
|
---|
907 | switch (fWindowSizeHiGain)
|
---|
908 | {
|
---|
909 | case 4:
|
---|
910 | SetResolutionPerPheHiGain(0.036);
|
---|
911 | break;
|
---|
912 | case 6:
|
---|
913 | SetResolutionPerPheHiGain(0.021);
|
---|
914 | break;
|
---|
915 | default:
|
---|
916 | *fLog << warn << "Could not set the high-gain extractor resolution per phe for window size "
|
---|
917 | << fWindowSizeHiGain << endl;
|
---|
918 | }
|
---|
919 |
|
---|
920 | switch (fWindowSizeLoGain)
|
---|
921 | {
|
---|
922 | case 4:
|
---|
923 | SetResolutionPerPheLoGain(0.005);
|
---|
924 | break;
|
---|
925 | case 6:
|
---|
926 | SetResolutionPerPheLoGain(0.004);
|
---|
927 | break;
|
---|
928 | default:
|
---|
929 | *fLog << warn << "Could not set the low-gain extractor resolution per phe for window size "
|
---|
930 | << fWindowSizeLoGain << endl;
|
---|
931 | }
|
---|
932 |
|
---|
933 | fNameWeightsFileSet = filename;
|
---|
934 |
|
---|
935 | return kTRUE;
|
---|
936 | }
|
---|
937 |
|
---|
938 |
|
---|
939 | //----------------------------------------------------------------------------
|
---|
940 | //
|
---|
941 | // The default (+ prepending possible "MC_") is read for:
|
---|
942 | //
|
---|
943 | // - RunType: Pedestal (independant of fAutomaticWeights)
|
---|
944 | // - fAutomaticWeights disabled
|
---|
945 | //
|
---|
946 | // if fAutomaticWeights enabled:
|
---|
947 | // - fNameWeightsFile.IsNull()
|
---|
948 | // - !fCalibPattern
|
---|
949 | // - fCalibPattern->GetPulserColor()==MCalibrationCam::kNONE
|
---|
950 | //
|
---|
951 | // If automatic weights are enabled, the case above didn't take place and
|
---|
952 | // fNameWeightsFile starts with "calibration_weights_"
|
---|
953 | // - the color (blue, UV) is replaced by the appropriate one
|
---|
954 | // taken from the calibration pattern
|
---|
955 | //
|
---|
956 | // In most cases a debug output is printed. Further output about the color
|
---|
957 | // determination can be switched on with debug level > 5;
|
---|
958 | //
|
---|
959 | // Returns:
|
---|
960 | // kFALSE: No new weights set
|
---|
961 | // kTRUE: New weights set
|
---|
962 | // kERROR: Error
|
---|
963 | //
|
---|
964 | Int_t MExtractTimeAndChargeDigitalFilter::GetAutomaticWeights()
|
---|
965 | {
|
---|
966 | const Ssiz_t pos = fNameWeightsFile.Last('/')+1;
|
---|
967 | const Ssiz_t len = fNameWeightsFile.Length();
|
---|
968 |
|
---|
969 | // Split file name in path and name
|
---|
970 | TString path = fNameWeightsFile(0, pos>=0?pos:len);
|
---|
971 | TString name = fNameWeightsFile(pos>=0?pos:0, len);
|
---|
972 |
|
---|
973 | // Remove trailing "MC_" for automatic weights
|
---|
974 | if (fAutomaticWeights && name.BeginsWith("MC_"))
|
---|
975 | name.Remove(0, 3);
|
---|
976 |
|
---|
977 | // In case of a pedetsal run no calibration pattern can be available
|
---|
978 | // the default weights are always used.
|
---|
979 | if (fRunHeader->GetRunType()==MRawRunHeader::kRTPedestal)
|
---|
980 | {
|
---|
981 | *fLog << dbg << "Pedestal file... using default weights: " << fNameWeightsFile << endl;
|
---|
982 | return ReadWeightsFile(name, path);
|
---|
983 | }
|
---|
984 |
|
---|
985 | // If automatic weights are switched off use default weights
|
---|
986 | if (!fAutomaticWeights)
|
---|
987 | {
|
---|
988 | *fLog << dbg << "Automatic weights switched off... using default weights: " << fNameWeightsFile << endl;
|
---|
989 | return ReadWeightsFile(name, path);
|
---|
990 | }
|
---|
991 |
|
---|
992 | // If automatic weights are switched on but no filename is given raise error
|
---|
993 | if (fNameWeightsFile.IsNull())
|
---|
994 | {
|
---|
995 | *fLog << err << "ERROR - Cannot get automatic weights without default filename." << endl;
|
---|
996 | return kERROR;
|
---|
997 | }
|
---|
998 |
|
---|
999 | // If this is no pedestal run, automatic weights are requested and a
|
---|
1000 | // filename for the weights file is given pedestal-extraction from
|
---|
1001 | // cosmics data is assumed.
|
---|
1002 | if (!fCalibPattern)
|
---|
1003 | {
|
---|
1004 | *fLog << dbg << "No decoded calibration pattern available... using default weights: " << fNameWeightsFile << endl;
|
---|
1005 | return ReadWeightsFile(name, path);
|
---|
1006 | }
|
---|
1007 |
|
---|
1008 | const Bool_t debug = gLog.GetDebugLevel()>5;
|
---|
1009 |
|
---|
1010 | // If no calibration pattern is available do not change the
|
---|
1011 | // current weighs or current weights file name.
|
---|
1012 | if (fCalibPattern->GetPulserColor()==MCalibrationCam::kNONE)
|
---|
1013 | {
|
---|
1014 | // If we are extracting data and the calibration pattern is kNONE
|
---|
1015 | // we assume that it is a data file without interleaved events
|
---|
1016 | // and calibration pattern information available.
|
---|
1017 | if ((fRunHeader->GetRunType()!=MRawRunHeader::kRTData && !fRunHeader->IsMonteCarloRun()) || debug)
|
---|
1018 | *fLog << dbg << "No calibration color set so far... guessing default: " << fNameWeightsFile << endl;
|
---|
1019 |
|
---|
1020 | return ReadWeightsFile(name, path);
|
---|
1021 | }
|
---|
1022 |
|
---|
1023 | if (debug)
|
---|
1024 | {
|
---|
1025 | *fLog << dbg << endl;
|
---|
1026 | *fLog << underline << GetDescriptor() << endl;
|
---|
1027 | *fLog << " Trying to get automatic weight for " << fNameWeightsFile << endl;
|
---|
1028 | *fLog << " Run type: ";
|
---|
1029 | }
|
---|
1030 |
|
---|
1031 | if (name.BeginsWith("calibration_weights_") && fCalibPattern)
|
---|
1032 | {
|
---|
1033 | if (debug)
|
---|
1034 | *fLog << " Calibration with color " << fCalibPattern->GetPulserColorStr() << ", setting ";
|
---|
1035 | switch (fCalibPattern->GetPulserColor())
|
---|
1036 | {
|
---|
1037 | case MCalibrationCam::kBLUE: // 2
|
---|
1038 | case MCalibrationCam::kGREEN: // 1
|
---|
1039 | if (debug)
|
---|
1040 | *fLog << "blue/green, ";
|
---|
1041 | name.ReplaceAll("UV", "blue");
|
---|
1042 | break;
|
---|
1043 |
|
---|
1044 | case MCalibrationCam::kUV: // 3
|
---|
1045 | case MCalibrationCam::kCT1: // 0
|
---|
1046 | if (debug)
|
---|
1047 | *fLog << "UV/CT1, ";
|
---|
1048 | name.ReplaceAll("blue", "UV");
|
---|
1049 | break;
|
---|
1050 | case MCalibrationCam::kNONE:
|
---|
1051 | break;
|
---|
1052 | default: // kNone + etc
|
---|
1053 | *fLog << err << "ERROR - Cannot get automatic weights for " << fCalibPattern->GetPulserColorStr() << endl;
|
---|
1054 | return kERROR;
|
---|
1055 | }
|
---|
1056 | }
|
---|
1057 |
|
---|
1058 | return ReadWeightsFile(name, path);
|
---|
1059 | }
|
---|
1060 |
|
---|
1061 | //----------------------------------------------------------------------------
|
---|
1062 | //
|
---|
1063 | // Create the weights file
|
---|
1064 | // Beware that the shape-histogram has to contain the pulse starting at bin 1
|
---|
1065 | //
|
---|
1066 | Bool_t MExtractTimeAndChargeDigitalFilter::WriteWeightsFile(TString filename, TH1F *shapehi, TH2F *autocorrhi,
|
---|
1067 | TH1F *shapelo, TH2F *autocorrlo )
|
---|
1068 | {
|
---|
1069 |
|
---|
1070 | const Int_t nbinshi = shapehi->GetNbinsX();
|
---|
1071 | Float_t binwidth = shapehi->GetBinWidth(1);
|
---|
1072 |
|
---|
1073 | TH1F *derivativehi = new TH1F(Form("%s%s",shapehi->GetName(),"_der"),
|
---|
1074 | Form("%s%s",shapehi->GetTitle()," derivative"),
|
---|
1075 | nbinshi,
|
---|
1076 | shapehi->GetBinLowEdge(1),
|
---|
1077 | shapehi->GetBinLowEdge(nbinshi)+binwidth);
|
---|
1078 |
|
---|
1079 | //
|
---|
1080 | // Calculate the derivative of shapehi
|
---|
1081 | //
|
---|
1082 | for (Int_t i = 1; i<nbinshi+1;i++)
|
---|
1083 | {
|
---|
1084 | derivativehi->SetBinContent(i,
|
---|
1085 | ((shapehi->GetBinContent(i+1)-shapehi->GetBinContent(i-1))/2./binwidth));
|
---|
1086 | derivativehi->SetBinError(i,
|
---|
1087 | (sqrt(shapehi->GetBinError(i+1)*shapehi->GetBinError(i+1)
|
---|
1088 | +shapehi->GetBinError(i-1)*shapehi->GetBinError(i-1))/2./binwidth));
|
---|
1089 | }
|
---|
1090 |
|
---|
1091 | //
|
---|
1092 | // normalize the shapehi, such that the integral for fWindowSize slices is one!
|
---|
1093 | //
|
---|
1094 | Float_t sum = 0;
|
---|
1095 | Int_t lasttemp = fBinningResolutionHiGain * (fSignalStartBinHiGain + fWindowSizeHiGain);
|
---|
1096 | lasttemp = lasttemp > nbinshi ? nbinshi : lasttemp;
|
---|
1097 |
|
---|
1098 | for (Int_t i=fBinningResolutionHiGain*fSignalStartBinHiGain; i<lasttemp; i++) {
|
---|
1099 | sum += shapehi->GetBinContent(i);
|
---|
1100 | }
|
---|
1101 | sum /= fBinningResolutionHiGain;
|
---|
1102 |
|
---|
1103 | shapehi->Scale(1./sum);
|
---|
1104 | derivativehi->Scale(1./sum);
|
---|
1105 |
|
---|
1106 | //
|
---|
1107 | // read in the noise auto-correlation function:
|
---|
1108 | //
|
---|
1109 | TMatrix Bhi(fWindowSizeHiGain,fWindowSizeHiGain);
|
---|
1110 |
|
---|
1111 | for (Int_t i=0; i<fWindowSizeHiGain; i++){
|
---|
1112 | for (Int_t j=0; j<fWindowSizeHiGain; j++){
|
---|
1113 | Bhi[i][j]=autocorrhi->GetBinContent(i+1,j+1); //+fSignalStartBinHiGain +fSignalStartBinHiGain
|
---|
1114 | }
|
---|
1115 | }
|
---|
1116 | Bhi.Invert();
|
---|
1117 |
|
---|
1118 | const Int_t nsizehi = fWindowSizeHiGain*fBinningResolutionHiGain;
|
---|
1119 | fAmpWeightsHiGain.Set(nsizehi);
|
---|
1120 | fTimeWeightsHiGain.Set(nsizehi);
|
---|
1121 |
|
---|
1122 | //
|
---|
1123 | // Loop over relative time in one BinningResolution interval
|
---|
1124 | //
|
---|
1125 | Int_t start = fBinningResolutionHiGain*(fSignalStartBinHiGain + 1);
|
---|
1126 |
|
---|
1127 | for (Int_t i = -fBinningResolutionHalfHiGain+1; i<=fBinningResolutionHalfHiGain; i++)
|
---|
1128 | {
|
---|
1129 |
|
---|
1130 | TMatrix g(fWindowSizeHiGain,1);
|
---|
1131 | TMatrix gT(1,fWindowSizeHiGain);
|
---|
1132 | TMatrix d(fWindowSizeHiGain,1);
|
---|
1133 | TMatrix dT(1,fWindowSizeHiGain);
|
---|
1134 |
|
---|
1135 | for (Int_t count=0; count < fWindowSizeHiGain; count++){
|
---|
1136 |
|
---|
1137 | g[count][0]=shapehi->GetBinContent(start
|
---|
1138 | +fBinningResolutionHiGain*count+i);
|
---|
1139 | gT[0][count]=shapehi->GetBinContent(start
|
---|
1140 | +fBinningResolutionHiGain*count+i);
|
---|
1141 | d[count][0]=derivativehi->GetBinContent(start
|
---|
1142 | +fBinningResolutionHiGain*count+i);
|
---|
1143 | dT[0][count]=derivativehi->GetBinContent(start
|
---|
1144 | +fBinningResolutionHiGain*count+i);
|
---|
1145 | }
|
---|
1146 |
|
---|
1147 | TMatrix m_denom = (gT*(Bhi*g))*(dT*(Bhi*d)) - (dT*(Bhi*g))*(dT*(Bhi*g));
|
---|
1148 | Float_t denom = m_denom[0][0]; // ROOT thinks, m_denom is still a matrix
|
---|
1149 |
|
---|
1150 | TMatrix m_first = dT*(Bhi*d); // ROOT thinks, m_first is still a matrix
|
---|
1151 | Float_t first = m_first[0][0]/denom;
|
---|
1152 |
|
---|
1153 | TMatrix m_last = gT*(Bhi*d); // ROOT thinks, m_last is still a matrix
|
---|
1154 | Float_t last = m_last[0][0]/denom;
|
---|
1155 |
|
---|
1156 | TMatrix m1 = gT*Bhi;
|
---|
1157 | m1 *= first;
|
---|
1158 |
|
---|
1159 | TMatrix m2 = dT*Bhi;
|
---|
1160 | m2 *=last;
|
---|
1161 |
|
---|
1162 | TMatrix w_amp = m1 - m2;
|
---|
1163 |
|
---|
1164 | TMatrix m_first1 = gT*(Bhi*g);
|
---|
1165 | Float_t first1 = m_first1[0][0]/denom;
|
---|
1166 |
|
---|
1167 | TMatrix m_last1 = gT*(Bhi*d);
|
---|
1168 | Float_t last1 = m_last1 [0][0]/denom;
|
---|
1169 |
|
---|
1170 | TMatrix m11 = dT*Bhi;
|
---|
1171 | m11 *=first1;
|
---|
1172 |
|
---|
1173 | TMatrix m21 = gT*Bhi;
|
---|
1174 | m21 *=last1;
|
---|
1175 |
|
---|
1176 | TMatrix w_time= m11 - m21;
|
---|
1177 |
|
---|
1178 | for (Int_t count=0; count < fWindowSizeHiGain; count++)
|
---|
1179 | {
|
---|
1180 | const Int_t idx = i+fBinningResolutionHalfHiGain+fBinningResolutionHiGain*count-1;
|
---|
1181 | fAmpWeightsHiGain [idx] = w_amp [0][count];
|
---|
1182 | fTimeWeightsHiGain[idx] = w_time[0][count];
|
---|
1183 | }
|
---|
1184 |
|
---|
1185 | } // end loop over i
|
---|
1186 |
|
---|
1187 | //
|
---|
1188 | // Low Gain histograms
|
---|
1189 | //
|
---|
1190 | TH1F *derivativelo = NULL;
|
---|
1191 | if (shapelo)
|
---|
1192 | {
|
---|
1193 | const Int_t nbinslo = shapelo->GetNbinsX();
|
---|
1194 | binwidth = shapelo->GetBinWidth(1);
|
---|
1195 |
|
---|
1196 | derivativelo = new TH1F(Form("%s%s",shapelo->GetName(),"_der"),
|
---|
1197 | Form("%s%s",shapelo->GetTitle()," derivative"),
|
---|
1198 | nbinslo,
|
---|
1199 | shapelo->GetBinLowEdge(1),
|
---|
1200 | shapelo->GetBinLowEdge(nbinslo)+binwidth);
|
---|
1201 |
|
---|
1202 | //
|
---|
1203 | // Calculate the derivative of shapelo
|
---|
1204 | //
|
---|
1205 | for (Int_t i = 1; i<nbinslo+1;i++)
|
---|
1206 | {
|
---|
1207 | derivativelo->SetBinContent(i,
|
---|
1208 | ((shapelo->GetBinContent(i+1)-shapelo->GetBinContent(i-1))/2./binwidth));
|
---|
1209 | derivativelo->SetBinError(i,
|
---|
1210 | (sqrt(shapelo->GetBinError(i+1)*shapelo->GetBinError(i+1)
|
---|
1211 | +shapelo->GetBinError(i-1)*shapelo->GetBinError(i-1))/2./binwidth));
|
---|
1212 | }
|
---|
1213 |
|
---|
1214 | //
|
---|
1215 | // normalize the shapelo, such that the integral for fWindowSize slices is one!
|
---|
1216 | //
|
---|
1217 | sum = 0;
|
---|
1218 | lasttemp = fBinningResolutionLoGain * (fSignalStartBinLoGain + fWindowSizeLoGain);
|
---|
1219 | lasttemp = lasttemp > nbinslo ? nbinslo : lasttemp;
|
---|
1220 |
|
---|
1221 | for (Int_t i=fBinningResolutionLoGain*fSignalStartBinLoGain; i<lasttemp; i++)
|
---|
1222 | sum += shapelo->GetBinContent(i);
|
---|
1223 |
|
---|
1224 | sum /= fBinningResolutionLoGain;
|
---|
1225 |
|
---|
1226 | shapelo->Scale(1./sum);
|
---|
1227 | derivativelo->Scale(1./sum);
|
---|
1228 |
|
---|
1229 | //
|
---|
1230 | // read in the noise auto-correlation function:
|
---|
1231 | //
|
---|
1232 | TMatrix Blo(fWindowSizeLoGain,fWindowSizeLoGain);
|
---|
1233 |
|
---|
1234 | for (Int_t i=0; i<fWindowSizeLoGain; i++){
|
---|
1235 | for (Int_t j=0; j<fWindowSizeLoGain; j++){
|
---|
1236 | Blo[i][j]=autocorrlo->GetBinContent(i+1+fSignalStartBinLoGain,j+1+fSignalStartBinLoGain);
|
---|
1237 | }
|
---|
1238 | }
|
---|
1239 | Blo.Invert();
|
---|
1240 |
|
---|
1241 | const Int_t nsizelo = fWindowSizeLoGain*fBinningResolutionLoGain;
|
---|
1242 | fAmpWeightsLoGain.Set(nsizelo);
|
---|
1243 | fTimeWeightsLoGain.Set(nsizelo);
|
---|
1244 |
|
---|
1245 | //
|
---|
1246 | // Loop over relative time in one BinningResolution interval
|
---|
1247 | //
|
---|
1248 | Int_t start = fBinningResolutionLoGain*fSignalStartBinLoGain + fBinningResolutionHalfLoGain;
|
---|
1249 |
|
---|
1250 | for (Int_t i = -fBinningResolutionHalfLoGain+1; i<=fBinningResolutionHalfLoGain; i++)
|
---|
1251 | {
|
---|
1252 |
|
---|
1253 | TMatrix g(fWindowSizeLoGain,1);
|
---|
1254 | TMatrix gT(1,fWindowSizeLoGain);
|
---|
1255 | TMatrix d(fWindowSizeLoGain,1);
|
---|
1256 | TMatrix dT(1,fWindowSizeLoGain);
|
---|
1257 |
|
---|
1258 | for (Int_t count=0; count < fWindowSizeLoGain; count++){
|
---|
1259 |
|
---|
1260 | g[count][0] = shapelo->GetBinContent(start
|
---|
1261 | +fBinningResolutionLoGain*count+i);
|
---|
1262 | gT[0][count]= shapelo->GetBinContent(start
|
---|
1263 | +fBinningResolutionLoGain*count+i);
|
---|
1264 | d[count][0] = derivativelo->GetBinContent(start
|
---|
1265 | +fBinningResolutionLoGain*count+i);
|
---|
1266 | dT[0][count]= derivativelo->GetBinContent(start
|
---|
1267 | +fBinningResolutionLoGain*count+i);
|
---|
1268 | }
|
---|
1269 |
|
---|
1270 | TMatrix m_denom = (gT*(Blo*g))*(dT*(Blo*d)) - (dT*(Blo*g))*(dT*(Blo*g));
|
---|
1271 | Float_t denom = m_denom[0][0]; // ROOT thinks, m_denom is still a matrix
|
---|
1272 |
|
---|
1273 | TMatrix m_first = dT*(Blo*d); // ROOT thinks, m_first is still a matrix
|
---|
1274 | Float_t first = m_first[0][0]/denom;
|
---|
1275 |
|
---|
1276 | TMatrix m_last = gT*(Blo*d); // ROOT thinks, m_last is still a matrix
|
---|
1277 | Float_t last = m_last[0][0]/denom;
|
---|
1278 |
|
---|
1279 | TMatrix m1 = gT*Blo;
|
---|
1280 | m1 *= first;
|
---|
1281 |
|
---|
1282 | TMatrix m2 = dT*Blo;
|
---|
1283 | m2 *=last;
|
---|
1284 |
|
---|
1285 | TMatrix w_amp = m1 - m2;
|
---|
1286 |
|
---|
1287 | TMatrix m_first1 = gT*(Blo*g);
|
---|
1288 | Float_t first1 = m_first1[0][0]/denom;
|
---|
1289 |
|
---|
1290 | TMatrix m_last1 = gT*(Blo*d);
|
---|
1291 | Float_t last1 = m_last1 [0][0]/denom;
|
---|
1292 |
|
---|
1293 | TMatrix m11 = dT*Blo;
|
---|
1294 | m11 *=first1;
|
---|
1295 |
|
---|
1296 | TMatrix m21 = gT*Blo;
|
---|
1297 | m21 *=last1;
|
---|
1298 |
|
---|
1299 | TMatrix w_time= m11 - m21;
|
---|
1300 |
|
---|
1301 | for (Int_t count=0; count < fWindowSizeLoGain; count++)
|
---|
1302 | {
|
---|
1303 | const Int_t idx = i+fBinningResolutionHalfLoGain+fBinningResolutionLoGain*count-1;
|
---|
1304 | fAmpWeightsLoGain [idx] = w_amp [0][count];
|
---|
1305 | fTimeWeightsLoGain[idx] = w_time[0][count];
|
---|
1306 | }
|
---|
1307 |
|
---|
1308 | } // end loop over i
|
---|
1309 | }
|
---|
1310 |
|
---|
1311 | ofstream fn(filename.Data());
|
---|
1312 |
|
---|
1313 | fn << "# High Gain Weights: " << fWindowSizeHiGain << " " << fBinningResolutionHiGain << endl;
|
---|
1314 | fn << "# (Amplitude) (Time) " << endl;
|
---|
1315 |
|
---|
1316 | for (Int_t i=0; i<nsizehi; i++)
|
---|
1317 | fn << "\t" << fAmpWeightsHiGain[i] << "\t" << fTimeWeightsHiGain[i] << endl;
|
---|
1318 |
|
---|
1319 | fn << "# Low Gain Weights: " << fWindowSizeLoGain << " " << fBinningResolutionLoGain << endl;
|
---|
1320 | fn << "# (Amplitude) (Time) " << endl;
|
---|
1321 |
|
---|
1322 | for (Int_t i=0; i<nsizehi; i++)
|
---|
1323 | fn << "\t" << fAmpWeightsLoGain[i] << "\t" << fTimeWeightsLoGain[i] << endl;
|
---|
1324 |
|
---|
1325 | delete derivativehi;
|
---|
1326 | if (derivativelo)
|
---|
1327 | delete derivativelo;
|
---|
1328 |
|
---|
1329 | return kTRUE;
|
---|
1330 | }
|
---|
1331 |
|
---|
1332 | void MExtractTimeAndChargeDigitalFilter::Print(Option_t *o) const
|
---|
1333 | {
|
---|
1334 | if (IsA()==Class())
|
---|
1335 | *fLog << GetDescriptor() << ":" << endl;
|
---|
1336 |
|
---|
1337 | MExtractTimeAndCharge::Print(o);
|
---|
1338 | *fLog << " Time Shift HiGain: " << fTimeShiftHiGain << " LoGain: " << fTimeShiftLoGain << endl;
|
---|
1339 | *fLog << " Window Size HiGain: " << fWindowSizeHiGain << " LoGain: " << fWindowSizeLoGain << endl;
|
---|
1340 | *fLog << " Binning Res HiGain: " << fBinningResolutionHiGain << " LoGain: " << fBinningResolutionHiGain << endl;
|
---|
1341 | *fLog << " Weights File desired: " << (fNameWeightsFile.IsNull()?"-":fNameWeightsFile.Data()) << endl;
|
---|
1342 | if (!fNameWeightsFileSet.IsNull())
|
---|
1343 | *fLog << " Weights File set: " << fNameWeightsFileSet << endl;
|
---|
1344 |
|
---|
1345 | TString opt(o);
|
---|
1346 | if (!opt.Contains("weights"))
|
---|
1347 | return;
|
---|
1348 |
|
---|
1349 | *fLog << endl;
|
---|
1350 | *fLog << inf << "Using the following weights: " << endl;
|
---|
1351 | *fLog << "Hi-Gain:" << endl;
|
---|
1352 | for (Int_t i=0; i<fBinningResolutionHiGain*fWindowSizeHiGain; i++)
|
---|
1353 | *fLog << " " << fAmpWeightsHiGain[i] << " \t " << fTimeWeightsHiGain[i] << endl;
|
---|
1354 |
|
---|
1355 | *fLog << "Lo-Gain:" << endl;
|
---|
1356 | for (Int_t i=0; i<fBinningResolutionLoGain*fWindowSizeLoGain; i++)
|
---|
1357 | *fLog << " " << fAmpWeightsLoGain[i] << " \t " << fTimeWeightsLoGain[i] << endl;
|
---|
1358 | }
|
---|