The TauAnalysis/MCEmbeddingTools Package

Complete: 3


software administrators: MichalBluj, ChristianVeelken


The TauAnalysis/MCEmbeddingTools package is dedicated to the production of signal template distributions in Z → tau + tau and W → tau + nu analyses by replacing the muons reconstructed in Z → mu + mu or W->munu events selected in data.

The package includes modules to create hybrid events that consist of an overlay of a measured event and some simulated particles. This method is most suitable for the replacement of muons by taus to simulate either Z → tautau or W → taunu decays. However, other decay processes that can be modelled on the basis of Z → mumu and W → munu decays are also thinkable. The overlay itself is done on the basis of the digitized detector output. Hence it is necessary to identify all detector cells that are likely to originate from the muons in the original event to remove them. This step offers the possibility to use a simple reco::RecHit-track association as well as sophisticated methods that take the typical energy deposition of muons into account (as a method to determine the typical energy deposit of a traversing muon in a calorimeter cell is yet not available, this feature is not yet implemented).

The newly created particles and their decay products are put into a noise free detector simulation based on the most current conditions description. The deactivation of electronics noise guarantees that only decay products of the particles are overlaid. Otherwise one would have another layer of electronis noise addition to the one that already exists in the original event. The result of the method is a collection of "digis" that are also the output of the raw2digi step or the first output of real experiment. Subsequently, the user can run all kinds of reconstruction algorithms (for matching on MC truth see remarks).

Please see the following link for a presentation of the technical details of the muon --> tau replacement technique.

In case the tool is used to produce a signal template distribution for W → tau + nu events, the differences in kinematics between Z and W events need to be corrected for. These corrections are described in section 9.1 of the analysis note Towards a measurement of the inclusive W → mu nu and Z → mu + mu cross sections in pp collisions at centre-of-mass energies of 14 TeV.

Get the code

to be checked in

Typical work flow

The following description of a work flow uses exemplarily the transformation of Z → mumu into Z → tautau.

  1. You create a module that selects two muons. You could either use an official select as the EWK group sample or apply individual selection criteria. The two muons are stored in a std::vector<reco::muon> that is passed to the next step. In the same step, the module should determine all detector cells with hits that are associated to the muons.
  2. You call the replacer module which creates you a new event containing of Z → tautau and the taus decay particles. To do so, the module automatically transforms the muon pair into a tau pair while preserving the direction and the dilepton mass, hence, the momentum slightly decreases as taus have a larger rest mass. It creates you a new HepMC::GenEvent containing the Z boson, the taus and their decay products. Of course, you can configure Tauola and Pythia for the tau decay as usual, e.g. impose restrictions on the decay modes.
  3. You run the full simulation on this product, but without any detector noise.
  4. You call the mixing module to overlay the original residual event and the separately simulated tau pair decay (the module uses a std::vector with raw cell IDs created in step 1 for the original event).
  5. You run the standard reconstruction on the result of step 4 and yield a hybrid event that can be used for subsequent analyses.



Only muons that are detected can act as a basis for a replacement. If a muon falls into a gap of the detector, it can't be reconstructed (of course, there are plenty other reasons why a muon cannot be detected). A tau in the same direction with the same momentum might have been detected as it is either measured in other detector components (e.g. hadronic tau decays) or a descending muon (in tau → mununu decay) has a different curvature. This effect should be considered in all analyses that depend on efficiencies. This effect can only be studied with the Monte Carlo truth. The option replacementMethod=0 is suited well for such a task.

Matching to MC truth

There are some algorithms that match reconstructed objects on simulated objects (e.g. references of recHits to simHits). These references are not available for the hybrid event. Unless there is some need for this, this situation will not change. If you think that this feature should be implemented (and you want to volunteer) contact us wink

Multiple instances

In general, it is not a good idea to use the replacer module more than once in one configuration file since there are is a problem with multiple initializations of the generators. If you really can't avoid it, you have to set the noInitialisation option for the second module.

Review status

Responsible: ToBeFound

Reviewer/Editor and Date (copy from screen) Comments
Last reviewed by: -- KatiLassilaPerini - 17 Oct 2006 created template page
Last reviewed by: -- ChristianVeelken - 01 Jul 2009 created initial version

Edit | Attach | Watch | Print version | History: r2 < r1 | Backlinks | Raw View | WYSIWYG | More topic actions
Topic revision: r2 - 2009-07-01 - ManuelZeise
    • Cern Search Icon Cern Search
    • TWiki Search Icon TWiki Search
    • Google Search Icon Google Search

    CMSPublic All webs login

This site is powered by the TWiki collaboration platform Powered by PerlCopyright & 2008-2019 by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding TWiki? Send feedback