Rene Brun writes: > Michael, > > I will investigate this problem once I will be back to CERN on Thursday. > Meanwhile, could you send me a running script showing the problem? > I have some suspicion about your h2D->FindBin(x,y). Hallihallo Rene! Thanks for your answer, but you should have shouted at me much earlier for this stupid mistake :-( When I tried to put all things into a nice macro for better readability, I finally got the point: TF2::Eval(x,y) will calculate the value at EXACT (x,y) TH2F::FindBin(x,y) will give a bin, where (x,y) is inside, but where (x,y) is not necessarily the bin CENTRE. The bin centre one can get with x_=h2D->GetXaxis()->GetBinCenter( h2D->GetXaxis()->FindBin(x) ); y_=h2D->GetYaxis()->GetBinCenter( h2D->GetYaxis()->FindBin(y) ); This bin centre (x_,y_) is used by TH2F::Add(TF2,double): TH2F::GetBinContent(FindBin(x,y)) - TF2::Eval(x_,y_) Of course, root is right ... Sorry, I knew somehow that it was a stupid mistake, but couldn't get a grip on it last week ... But thanks for your encouraging answer nevertheless, Michael > > Rene Brun > > Michael Wiesmann wrote: > > > > Hallihallo Rooters! > > > > I have a TH2F h2D, I check a bin content a la > > h2D->GetBinContent( h2D->FindBin(x,y)); //returns 17 > > > > I have a TF2 fitFcn,I check a value a la > > fitFcn->Eval(x,y) // returns 12.5 > > //same with TF1::EvalPar(x,parameter) > > > > Now I do > > h2D->Add(fitFcn,-1.0); > > h2D->GetBinContent( h2D->FindBin(x,y)); /returns 7.76 > > > > Naively I would expect 17-12.5 to be 4.5. Obviously I am too naive for > > this problem. > > > > I tried to use TH1F::Eval(TF1,"A"), which gave me (17+12.5=) 26.2, > > which is at least consistent. > > > > Sounds like TH2F::Add(TF1) uses another set of parameters. > > > > Any hint what I missunderstood? > > > > Michael > > > > PS: I am using root 3.03/09 on RH Linux 7.2
This archive was generated by hypermail 2b29 : Thu Jan 01 2004 - 17:50:12 MET