fPhotWei [i] = 0;
}
//hidden algorithm
- fMipX=fMipY=fThTrkFit=fPhTrkFit=fCkovFit=-999;
+ fMipX=fMipY=fThTrkFit=fPhTrkFit=fCkovFit=fMipQ=fRadX=fRadY=-999;
fIdxMip=fNClu=0;
- for (Int_t i=0; i<1000; i++) {
+ fCkovSig2=0;
+ for (Int_t i=0; i<100; i++) {
fXClu[i] = fYClu[i] = 0;
+ fClCk[i] = kTRUE;
}
}
//++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
TVector3 v(-999,-999,-999);
Double_t trkBeta = 1./(TMath::Cos(ckovTh)*fRadNmean);
+
+ if(trkBeta > 1) trkBeta = 1; //protection against bad measured thetaCer
+ if(trkBeta < 0) trkBeta = 0.0001; //
v.SetX(SigLoc (ckovTh,ckovPh,trkBeta));
v.SetY(SigGeom(ckovTh,ckovPh,trkBeta));
// dip and azimuthal angles for MIP taken at the entrance to radiator, [radians]
// MIP beta
// Returns: absolute error on Cerenkov angle, [radians]
+
Double_t phiDelta = phiC - fTrkDir.Phi();
- Double_t alpha =TMath::Cos(fTrkDir.Theta())-TMath::Tan(thetaC)*TMath::Cos(phiDelta)*TMath::Sin(fTrkDir.Theta());
- Double_t k = 1.-fRadNmean*fRadNmean+alpha*alpha/(betaM*betaM);
+ Double_t sint = TMath::Sin(fTrkDir.Theta());
+ Double_t cost = TMath::Cos(fTrkDir.Theta());
+ Double_t sinf = TMath::Sin(fTrkDir.Phi());
+ Double_t cosf = TMath::Cos(fTrkDir.Phi());
+ Double_t sinfd = TMath::Sin(phiDelta);
+ Double_t cosfd = TMath::Cos(phiDelta);
+ Double_t tantheta = TMath::Tan(thetaC);
+
+ Double_t alpha =cost-tantheta*cosfd*sint; // formula (11)
+ Double_t k = 1.-fRadNmean*fRadNmean+alpha*alpha/(betaM*betaM); // formula (after 8 in the text)
if (k<0) return 1e10;
+ Double_t mu =sint*sinf+tantheta*(cost*cosfd*sinf+sinfd*cosf); // formula (10)
+ Double_t e =sint*cosf+tantheta*(cost*cosfd*cosf-sinfd*sinf); // formula (9)
- Double_t mu =TMath::Sin(fTrkDir.Theta())*TMath::Sin(fTrkDir.Phi())+TMath::Tan(thetaC)*(TMath::Cos(fTrkDir.Theta())*TMath::Cos(phiDelta)*TMath::Sin(fTrkDir.Phi())+TMath::Sin(phiDelta)*TMath::Cos(fTrkDir.Phi()));
- Double_t e =TMath::Sin(fTrkDir.Theta())*TMath::Cos(fTrkDir.Phi())+TMath::Tan(thetaC)*(TMath::Cos(fTrkDir.Theta())*TMath::Cos(phiDelta)*TMath::Cos(fTrkDir.Phi())-TMath::Sin(phiDelta)*TMath::Sin(fTrkDir.Phi()));
+ Double_t kk = betaM*TMath::Sqrt(k)/(fgkGapThick*alpha); // formula (6) and (7)
+ Double_t dtdxc = kk*(k*(cosfd*cosf-cost*sinfd*sinf)-(alpha*mu/(betaM*betaM))*sint*sinfd); // formula (6)
+ Double_t dtdyc = kk*(k*(cosfd*sinf+cost*sinfd*cosf)+(alpha* e/(betaM*betaM))*sint*sinfd); // formula (7) pag.4
- Double_t kk = betaM*TMath::Sqrt(k)/(8*alpha);
- Double_t dtdxc = kk*(k*(TMath::Cos(phiDelta)*TMath::Cos(fTrkDir.Phi())-TMath::Cos(fTrkDir.Theta())*TMath::Sin(phiDelta)*TMath::Sin(fTrkDir.Phi()))-(alpha*mu/(betaM*betaM))*TMath::Sin(fTrkDir.Theta())*TMath::Sin(phiDelta));
- Double_t dtdyc = kk*(k*(TMath::Cos(phiDelta)*TMath::Sin(fTrkDir.Phi())+TMath::Cos(fTrkDir.Theta())*TMath::Sin(phiDelta)*TMath::Cos(fTrkDir.Phi()))+(alpha* e/(betaM*betaM))*TMath::Sin(fTrkDir.Theta())*TMath::Sin(phiDelta));
-
- return TMath::Sqrt(0.2*0.2*dtdxc*dtdxc + 0.25*0.25*dtdyc*dtdyc);
+ Double_t errX = 0.2,errY=0.25; //end of page 7
+ return TMath::Sqrt(errX*errX*dtdxc*dtdxc + errY*errY*dtdyc*dtdyc);
}
//++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
Double_t AliHMPIDRecon::SigCrom(Double_t thetaC, Double_t phiC,Double_t betaM)const
// dip and azimuthal angles for MIP taken at the entrance to radiator, [radians]
// MIP beta
// Returns: absolute error on Cerenkov angle, [radians]
+
Double_t phiDelta = phiC - fTrkDir.Phi();
- Double_t alpha =TMath::Cos(fTrkDir.Theta())-TMath::Tan(thetaC)*TMath::Cos(phiDelta)*TMath::Sin(fTrkDir.Theta());
- Double_t dtdn = TMath::Cos(fTrkDir.Theta())*fRadNmean*betaM*betaM/(alpha*TMath::Tan(thetaC));
+ Double_t sint = TMath::Sin(fTrkDir.Theta());
+ Double_t cost = TMath::Cos(fTrkDir.Theta());
+ Double_t cosfd = TMath::Cos(phiDelta);
+ Double_t tantheta = TMath::Tan(thetaC);
+
+ Double_t alpha =cost-tantheta*cosfd*sint; // formula (11)
+ Double_t dtdn = cost*fRadNmean*betaM*betaM/(alpha*tantheta); // formula (12)
- Double_t f = 0.00928*(7.75-5.635)/TMath::Sqrt(12.);
+// Double_t f = 0.00928*(7.75-5.635)/TMath::Sqrt(12.);
+ Double_t f = 0.0172*(7.75-5.635)/TMath::Sqrt(24.);
return f*dtdn;
}//SigCrom()
// Returns: absolute error on Cerenkov angle, [radians]
Double_t phiDelta = phiC - fTrkDir.Phi();
- Double_t alpha =TMath::Cos(fTrkDir.Theta())-TMath::Tan(thetaC)*TMath::Cos(phiDelta)*TMath::Sin(fTrkDir.Theta());
- Double_t k = 1.-fRadNmean*fRadNmean+alpha*alpha/(betaM*betaM);
+ Double_t sint = TMath::Sin(fTrkDir.Theta());
+ Double_t cost = TMath::Cos(fTrkDir.Theta());
+ Double_t sinf = TMath::Sin(fTrkDir.Phi());
+ Double_t cosfd = TMath::Cos(phiDelta);
+ Double_t costheta = TMath::Cos(thetaC);
+ Double_t tantheta = TMath::Tan(thetaC);
+
+ Double_t alpha =cost-tantheta*cosfd*sint; // formula (11)
+
+ Double_t k = 1.-fRadNmean*fRadNmean+alpha*alpha/(betaM*betaM); // formula (after 8 in the text)
if (k<0) return 1e10;
- Double_t eTr = 0.5*1.5*betaM*TMath::Sqrt(k)/(8*alpha);
- Double_t lambda = 1.-TMath::Sin(fTrkDir.Theta())*TMath::Sin(fTrkDir.Theta())*TMath::Sin(phiC)*TMath::Sin(phiC);
-
- Double_t c = 1./(1.+ eTr*k/(alpha*alpha*TMath::Cos(thetaC)*TMath::Cos(thetaC)));
- Double_t i = betaM*TMath::Tan(thetaC)*lambda*TMath::Power(k,1.5);
- Double_t ii = 1.+eTr*betaM*i;
+ Double_t eTr = 0.5*fgkRadThick*betaM*TMath::Sqrt(k)/(fgkGapThick*alpha); // formula (14)
+ Double_t lambda = 1.-sint*sint*sinf*sinf; // formula (15)
- Double_t err = c * (i/(alpha*alpha*8) + ii*(1.-lambda) / ( alpha*alpha*8*betaM*(1.+eTr)) );
- Double_t trErr = 1.5/(TMath::Sqrt(12.)*TMath::Cos(fTrkDir.Theta()));
+ Double_t c1 = 1./(1.+ eTr*k/(alpha*alpha*costheta*costheta)); // formula (13.a)
+ Double_t c2 = betaM*TMath::Power(k,1.5)*tantheta*lambda/(fgkGapThick*alpha*alpha); // formula (13.b)
+ Double_t c3 = (1.+eTr*k*betaM*betaM)/((1+eTr)*alpha*alpha); // formula (13.c)
+ Double_t c4 = TMath::Sqrt(k)*tantheta*(1-lambda)/(fgkGapThick*betaM); // formula (13.d)
+ Double_t dtdT = c1 * (c2+c3*c4);
+ Double_t trErr = fgkRadThick/(TMath::Sqrt(12.)*cost);
- return trErr*err;
+ return trErr*dtdT;
}//SigGeom()
//++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
//
// From here HTA....
//
//++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
-Int_t AliHMPIDRecon::CkovHiddenTrk(AliESDtrack *pTrk,TClonesArray *pCluLst,Double_t nmean)
+Bool_t AliHMPIDRecon::CkovHiddenTrk(AliESDtrack *pTrk,TClonesArray *pCluLst,Double_t nmean)
{
// Pattern recognition method without any infos from tracking:HTA (Hidden Track Algorithm)...
// The method finds in the chmber the cluster with the highest charge
fRadNmean=nmean;
+ if(pCluLst->GetEntriesFast()>100) return kFALSE; //boundary check for CluX,CluY...
Float_t mipX=-1,mipY=-1;Int_t mipId=-1,mipQ=-1;
- fPhotCnt=0;
Double_t qRef = 0;
- fNClu = pCluLst->GetEntriesFast();
- for (Int_t iClu=0;iClu<fNClu;iClu++){ //clusters loop
+ Int_t nCh=0;
+ for (Int_t iClu=0;iClu<pCluLst->GetEntriesFast();iClu++){ //clusters loop
AliHMPIDCluster *pClu=(AliHMPIDCluster*)pCluLst->UncheckedAt(iClu); //get pointer to current cluster
+ nCh = pClu->Ch();
fXClu[iClu] = pClu->X();fYClu[iClu] = pClu->Y(); //store x,y for fitting procedure
+ fClCk[iClu] = kTRUE; //all cluster are accepted at this stage to be reconstructed
if(pClu->Q()>qRef){ //searching the highest charge to select a MIP
qRef = pClu->Q();
mipId=iClu; mipX=pClu->X();mipY=pClu->Y();mipQ=(Int_t)pClu->Q();
}
}//clusters loop
- if(qRef>pParam->QCut()){ //charge compartible with MIP clusters
+ fNClu = pCluLst->GetEntriesFast();
+ if(qRef>pParam->QCut()){ //charge compartible with MIP clusters
fIdxMip = mipId;
+ fClCk[mipId] = kFALSE;
fMipX = mipX; fMipY=mipY; fMipQ = qRef;
- if(!DoRecHiddenTrk()) return 1; //Do track and ring reconstruction,if problems returns 1
- pTrk->SetHMPIDtrk(fRadX,fRadY,fThTrkFit,fPhTrkFit); //store track intersection info
- pTrk->SetHMPIDmip(fMipX,fMipY,(Int_t)fMipQ,fNClu); //store mip info
- pTrk->SetHMPIDcluIdx(pCluLst->GetUniqueID(),fIdxMip); //set cham number and index of cluster
- pTrk->SetHMPIDsignal(fCkovFit); //find best Theta ckov for ring i.e. track
+ if(!DoRecHiddenTrk(pCluLst)) {
+ pTrk->SetHMPIDsignal(kNoPhotAccept);
+ return kFALSE;
+ } //Do track and ring reconstruction,if problems returns 1
+ pTrk->SetHMPIDtrk(fRadX,fRadY,fThTrkFit,fPhTrkFit); //store track intersection info
+ pTrk->SetHMPIDmip(fMipX,fMipY,(Int_t)fMipQ,fNClu); //store mip info
+ pTrk->SetHMPIDcluIdx(nCh,fIdxMip); //set cham number and index of cluster
+ pTrk->SetHMPIDsignal(fCkovFit); //find best Theta ckov for ring i.e. track
+ pTrk->SetHMPIDchi2(fCkovSig2); //errors squared
+// Printf(" n clusters tot %i accepted %i",pCluLst->GetEntriesFast(),fNClu);
+ return kTRUE;
}
- return 0;
+
+ return kFALSE;
}//CkovHiddenTrk()
//++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
-Bool_t AliHMPIDRecon::DoRecHiddenTrk()
+Bool_t AliHMPIDRecon::DoRecHiddenTrk(TClonesArray *pCluLst)
{
// Pattern recognition method without any infos from tracking...
// First a preclustering filter to avoid part of the noise
// Arguments: none
// Returns: none
Double_t phiRec;
- CluPreFilter();
+ if(!CluPreFilter(pCluLst)) {return kFALSE;}
if(!FitEllipse(phiRec)) {return kFALSE;}
- return FitFree(phiRec);
-}
+ Int_t nClTmp1 = pCluLst->GetEntriesFast()-1; //minus MIP...
+ Int_t nClTmp2 = 0;
+ while(nClTmp1 != nClTmp2){
+ SetNClu(pCluLst->GetEntriesFast());
+ if(!FitFree(phiRec)) {return kFALSE;}
+ nClTmp2 = NClu();
+ if(nClTmp2!=nClTmp1) {nClTmp1=nClTmp2;nClTmp2=0;}
+ }
+ fNClu = nClTmp2;
+ return kTRUE;
+}//DoRecHiddenTrk()
//++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
-void AliHMPIDRecon::CluPreFilter()
+Bool_t AliHMPIDRecon::CluPreFilter(TClonesArray *pCluLst)
{
// Filter of bkg clusters
// based on elliptical-shapes...
//
- ;
+ if(pCluLst->GetEntriesFast()>50||pCluLst->GetEntriesFast()<4) return kFALSE;
+ else return kTRUE;
}
//++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
Bool_t AliHMPIDRecon::FitEllipse(Double_t &phiRec)
// HG - AF
// b = --------
// AB - H^2
-
Double_t cA,cB,cF,cG,cH;
Double_t aArg=-1; Int_t iErrFlg; //tmp vars for TMinuit
gMinuit->mnparm(3," G ",1,0.01,0,0,iErrFlg);
gMinuit->mnparm(4," F ",1,0.01,0,0,iErrFlg);
- gMinuit->mnexcm("SIMPLEX" ,&aArg,0,iErrFlg);
+ gMinuit->mnexcm("SIMPLEX",&aArg,0,iErrFlg);
gMinuit->mnexcm("MIGRAD" ,&aArg,0,iErrFlg);
gMinuit->mnpout(0,sName,cA,d1,d2,d3,iErrFlg);
gMinuit->mnpout(1,sName,cB,d1,d2,d3,iErrFlg);
delete gMinuit;
Double_t i2 = cA*cB-cH*cH; //quartic invariant : i2 > 0 ellipse, i2 < 0 hyperbola
+ if(i2<=0) return kFALSE;
Double_t aX = (cH*cF-cB*cG)/i2; //x centre of the canonical section
Double_t bY = (cH*cG-cA*cF)/i2; //y centre of the canonical section
Double_t alfa1 = TMath::ATan(2*cH/(cA-cB)); //alpha = angle of rotation of the conical section
if(alfa1<0) alfa1+=TMath::Pi();
alfa1*=0.5;
- Double_t alfa2 = alfa1+TMath::Pi();
- Double_t phiref = TMath::ATan2(bY-fMipY,aX-fMipX); //evaluate in a unique way the angle of rotation comapring it
- if(phiref<0) phiref+=TMath::TwoPi(); //with the vector that poinst to the centre from the mip
+// Double_t alfa2 = alfa1+TMath::Pi();
+ Double_t phiref = TMath::ATan2(bY-fMipY,aX-fMipX); //evaluate in a unique way the angle of rotation comparing it
+ if(phiref<0) phiref+=TMath::TwoPi(); //with the vector that points to the centre from the mip
if(i2<0) phiref+=TMath::Pi();
if(phiref>TMath::TwoPi()) phiref-=TMath::TwoPi();
// Printf(" alfa1 %f",alfa1*TMath::RadToDeg());
// Printf(" alfa2 %f",alfa2*TMath::RadToDeg());
// Printf(" firef %f",phiref*TMath::RadToDeg());
- if(TMath::Abs(alfa1-phiref)<TMath::Abs(alfa2-phiref)) phiRec = alfa1; else phiRec = alfa2;
+// if(TMath::Abs(alfa1-phiref)<TMath::Abs(alfa2-phiref)) phiRec = alfa1; else phiRec = alfa2;
-// cout << " phi reconstructed " << phiRec*TMath::RadToDeg() << endl;
- return (i2>0);
+// Printf("FitEllipse: phi reconstructed %f",phiRec*TMath::RadToDeg());
+ phiRec=phiref;
+ return kTRUE;
//
}
//++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
TString sName;
Double_t th,ph;
- gMinuit->mnexcm("SET PRI",&aArg,1,iErrFlg); //suspend all printout from TMinuit
- gMinuit->mnexcm("SET NOW",&aArg,0,iErrFlg);
-
gMinuit->mnparm(0," theta ", 0.01,0.01,0,TMath::PiOver2(),iErrFlg);
gMinuit->mnparm(1," phi ",phiRec,0.01,0,TMath::TwoPi() ,iErrFlg);
pRec->SetRadXY(xrad,yrad);
pRec->SetTrack(xrad,yrad,thTrk,phTrk);
- Double_t meanCkov=0;
+ Double_t meanCkov =0;
Double_t meanCkov2=0;
Double_t thetaCer,phiCer;
- for(Int_t i=0;i<pRec->NClu();i++) {
+ Int_t nClAcc = 0;
+ Int_t nClTot=pRec->NClu();
+
+ for(Int_t i=0;i<nClTot;i++) {
+ if(!(pRec->ClCk(i))) continue;
pRec->FindPhotCkov(pRec->XClu(i),pRec->YClu(i),thetaCer,phiCer);
-
meanCkov += thetaCer;
meanCkov2 += thetaCer*thetaCer;
+ nClAcc++;
}
- meanCkov/=pRec->NClu();
- Double_t rms = TMath::Sqrt(meanCkov2/pRec->NClu() - meanCkov*meanCkov);
- f = rms/TMath::Sqrt(pRec->NClu());
- Printf(" mean %f rms/sqrt(n) %f",meanCkov,f);
- if(iflag==3) pRec->SetCkovFit(meanCkov);
+ if(nClAcc==0) {f=999;return;}
+ meanCkov/=nClAcc;
+ Double_t rms = (meanCkov2 - meanCkov*meanCkov*nClAcc)/nClAcc;
+ if(rms<0) Printf(" rms2 = %f, strange!!!",rms);
+ rms = TMath::Sqrt(rms);
+ f = rms/TMath::Sqrt(nClAcc);
+
+ if(iflag==3) {
+ Printf("FunMinPhot before: photons candidates %i used %i",nClTot,nClAcc);
+ nClAcc = 0;
+ Double_t meanCkov1=0;
+ Double_t meanCkov2=0;
+ for(Int_t i=0;i<nClTot;i++) {
+ if(!(pRec->ClCk(i))) continue;
+ pRec->FindPhotCkov(pRec->XClu(i),pRec->YClu(i),thetaCer,phiCer);
+ if(TMath::Abs(thetaCer-meanCkov)<2*rms) {
+ meanCkov1 += thetaCer;
+ meanCkov2 += thetaCer*thetaCer;
+ nClAcc++;
+ } else pRec->SetClCk(i,kFALSE);
+ }
+ meanCkov1/=nClAcc;
+ Double_t rms2 = (meanCkov2 - meanCkov*meanCkov*nClAcc)/nClAcc;
+ Printf("FunMinPhot after: photons candidates %i used %i thetaCer %f",nClTot,nClAcc,meanCkov1);
+ pRec->SetCkovFit(meanCkov1);
+ pRec->SetCkovSig2(rms2);
+ pRec->SetNClu(nClAcc);
+ }
}//FunMinPhot()
//++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
//