TPC PID update:

[u/mrichter/AliRoot.git] / EMCAL / doc / simulation.tex

\section{Simulation code}

The class AliSimulation manages this part. 
Have a look to the macro ``\$ALICE\_ROOT/EMCAL/
macros/TestEMCALSimulation.C''. The simulation
consists of different steps: geometry and event definition, particle
generation, transport of the particle in the material (GEANT) and
finally digitization. Note that the final output from the digitization
process is not the same as the experimental Raw Data. The process
of converting the digitized data to Raw Data is discussed in Sec.~1.4.
In Sec.~1.5, I give the recipe to do all the steps.


\subsection{Event Generation and particle transport: Hits}


Once the generator is executed, the generated particles are transported
in the detector material with the Monte Carlo code, GEANT3 by default. Other options are 
GEANT4 or FLUKA (some license problems  with FLUKA right now so not in use?).
All the generated particles are kept in a file called \textbf{Kinematics.root}.
After the particle transport is executed, the objects \textbf{Hits}
are created. They contain the energy deposited in the sensitive material
of the detector by the generated particle, their position, impact
time (after collision) and the identity of the original particle.
Hits are stored in a file called \textbf{DETECTOR.Hits.root}, in the
calorimeter case: \textbf{EMCAL.Hits.root.}


\subsection{Digitization: SDigits and Digits - Evi}

We want to generate events which look like the real data collected
by the experiment. In the end, we want to have an amplitude in ADC
counts and a time (when particle traverse a cell) per each cell (tower)
of the calorimeter. In the code for calorimeters, it is done in the
following steps: 1st) \textbf{SDigit} objects are created, they consist
of the sum of deposited energy by all Hits in a cell (a particle can
create Hits in different cells but only one in a single cell), so
there is only one SDigit per fired cell; 2nd) \textbf{Digit} objects
are created, they are like the SDigits but the energy in the cell
is transformed into the ADC amplitude units, the electronic noise
is added and Digits whose energy does not pass an energy threshold
(3 ADC counts) are eliminated. SDigits and Digits are stored in the files
\textbf{EMCAL.SDigits.root} and \textbf{EMCAL.Digits.root}, respectively.

\subsection{Raw data - David}

What we will get directly from the experiment are not Digits but a
time samples of ADC counts per each cell. These samples are called
\textbf{Raw Data}. The samples have a shape, more complicated than
a Gaussian distribution, which is fitted offline. With real data,
Digits amplitude is just the maximum of the distribution obtained
with the fit to the sample. The Digit time (defined by a time the
particle hit the active volume of the detector) is the time bin when
the signal begins to rise. There is a method to pass from Digits to
Raw and vice versa in the class AliEMCALRawUtils: Raw2Digits and Digits2Raw,
respectively. For the reconstruction step we need the Digits. The
generation of Raw Data is optional during simulations, we can reconstruct
data generating directly Digits, but Raw data will be the initial
step when reconstructing real data.


\subsection{How to make a simulation}

TestEMCALSimulation.C is a very simple macro where we specify all the simulation parameters
and execute the simulation, here I put a similar but a bit more elaborated macro:

\begin{singlespace}
\texttt{void TestEMCALSimulation() \{ }

\texttt{TString detector=\char`\"{}EMCAL TPC\char`\"{}; // Define
in this variable the detectors }

\texttt{//that you want to be included in the simulation for the digitization. }

\texttt{//They can be less detectors than the detectors defined in
the Config.C }

\texttt{//file, imagine that you want all the detectors in front of
EMCal present}

\texttt{//to consider the conversion of particles but you are not
really }

\texttt{//interested in the output from these detectors. Option detector=\char`\"{}ALL\char`\"{} }

\texttt{//makes all detectors. }

\texttt{AliSimulation sim ; //Create simulation object }

\texttt{// Generation and simulation }

\texttt{sim.SetRunGeneration(kTRUE) ; //Default value is kTRUE, make
generation }

\texttt{//For some reason we may want to redo the Digitization, without
redoing }

\texttt{//the generation, in this case it must set to kFALSE }

\texttt{// Making SDigits }

\texttt{sim.SetMakeSDigits(detector) ; //We want to make SDigits}

\texttt{// set no detectors if SDigits are already made}

\texttt{// Making Digits }

\texttt{sim.SetMakeDigits(detector) ; //We want to make Digits }

\texttt{// set no detectors if SDigits are already made}

\texttt{//Merging }

\texttt{//sim.MergeWith(\char`\"{}bgrd/galice.root\char`\"{}) ; //If
we want to merge a signal }

\texttt{//and a background, the merging is done at the SDigit level. The }

\texttt{//background must be located in the repertory defined in the
method. }

\texttt{//Write Raw Data, make Raw data from digits }

\texttt{//sim.SetWriteRawData(detector) ;}

\texttt{//sim.SetConfigFile(\char`\"{}somewhere/ConfigXXX.C\char`\"{});//Default
is Config.C}

\texttt{//Make the simulation }

\texttt{sim.Run(3) ; // Run the simulation and make 3 events }

\texttt{\} }
\end{singlespace}