Dear Djamel, I think that's exactly what we need to improve global tracking efficiency in bfs. Today we had a discussion, here in Krakow, concerning this issue, tomorrow will continue. As I understand selecting BrBfsTrackingModule::SetCleanUp(BrBfsTrackingModule::kCUT2, kTRUE); will keep all T3-D3-T4 and only these T2-D3-T4 which do not share the same T4 track with any of T3-D3-T4. Is it right? Regards Pawel. Djamel Ouerdane wrote: > Hi all, > > I discovered recently some odd behaviour while I was trying to improve the > BFS tracking and found out that (at least in the data I was checking, > containing local DC tracks), the local track IDs were not set. > > The reason why we haven't seen some weird e.g. PID is because there's on > average only one track found in each DC. But of course, it happens that > more than one track are recontructed in each DC. In this case, the track > IDs are random, sometimes this random number is the same for all tracks. > When it comes to the matching between say T3 and T4, you can guess what > happens then (this is extremely important when D3 matched tracks are > combined with D4 matched tracks to get BfsTracks!) > > I have fixed the problem in BrDCTrackingModule (well, I hope, since I'm > not the author of this class) and introduced a couple of more changes in > BrBfsTrackingModule: > > * removed everything dealing with the primary vertex in TrackToVertex > when FFS and BFS are aligned. The global track member fTrackVertex in this > case is the intersection of the TOF1 plane. The path length calculation > stops also there. So we need an extra module to combine FFS and BFS (very > easy when T2 - T4 are combined) for specific purpose e.g. TOF2 time > calibration. > > * added options so that: > 1- T3 - T4 are combined if you call > BrBfsTrackingModule::SetFrontMode(BrBfsTrackingModule::kT3, kTRUE); > > 2- T2 - T4 are combined if you call > BrBfsTrackingModule::SetFrontMode(BrBfsTrackingModule::kT2, kTRUE); > > The default of fFrontMode is both. If you call set T2 to true in case the > BFS faces the beamline directly, it is disabled and use T3 instead. > The BrBfsTrack class has already a boolean member to know if T2 was used > or not. > > 3- now, if both T2 - T4 and T3 - T4 were combined, you end up with > matched tracks sharing the same T4 track. It is ok for checking data and > later on for combining T2 - T3 (for e.g. consistency check) but you can > also choose to clean up the duplicates, either the ones made with T3 or > the ones using T2. This is done by calling: > > BrBfsTrackingModule::SetCleanUp(BrBfsTrackingModule::kCUT2, kTRUE); > > OR > > BrBfsTrackingModule::SetCleanUp(BrBfsTrackingModule::kCUT3, kTRUE); > > Of course, you cannot clean up both! :) The default is no clean up. > > I checked that the stuff works and give the desirable results. > > * the old method SetCombineMode(bool) has been renamed SetExtendD3Match(bool) > (This method allows you to enable the swim of D3 matched tracks > through D4 - in case there were no D4 matched tracks to be combined with - > and the creation of a virtual D4 matched track, useful when the > efficiency of T5 is low) > > This is it :) > > Ciao > Djam > > -- > Djamel Ouerdane ------------------------------------------o > | Niels Bohr Institute | Home: | > | Blegdamsvej 17, DK-2100 Ø | Jagtvej 141 2D, | > | Fax: +45 35 32 50 16 | DK-2200 Copenhagen N | > | Tel: +45 35 32 52 69 | +45 35 86 19 74 | > | http://www.nbi.dk/~ouerdane | > | ouerdane@nbi.dk | > o---------------------------------------------------------o -- ---------------------------------------------------------------------- | Pawel Staszel | | Niels Bohr Institute Tb 8 email: staszel@nbi.dk | | Blegdamsvej 17 phone: (+45) 35 32 53 51 | | København Ø FAX: (+45) 35 32 50 16 | | Danmark | ----------------------------------------------------------------------
This archive was generated by hypermail 2b30 : Thu Apr 04 2002 - 10:57:39 EST