Notes on BaseQA with HFT hits
Updated on Tue, 2015-09-15 11:24. Originally created by jwebb on 2015-09-15 11:18.
Goal: compare TPC only tracks which would have been reconstructed with HFT hits to corresponding HFT tracks.
Method: Reconstruct events with HFT in tracking. Create a TTree associating ID truth with number of reco hits in each layer. Process TPC only tracking sample and apply cuts based on ID truth / TTree lookup.
Figure 1 -- Shows corrspondence between the number of hits in the TPC (0-49), SST( 50-54), IST (55-59) and PXL (60-65).
z-axis is cut... < 100 entries are eliminated. Note that the z-axis is scaled by 1000. Looks like a good (enough) correlation to proceed.
NOTE -- in debugging these plots, I found that the base QA code has an issue with respect to clone tracks. TPC only tracks are not considered when counting clones -- i.e. a track may be split into two. One is HFT, one is TPC only, and the base QA doesn't register this as a clone. The reason for this is that TPC only tracks are eliminated from study before the clone classification is made. Will need to revisit this.
Method: Reconstruct events with HFT in tracking. Create a TTree associating ID truth with number of reco hits in each layer. Process TPC only tracking sample and apply cuts based on ID truth / TTree lookup.
Figure 1 -- Shows corrspondence between the number of hits in the TPC (0-49), SST( 50-54), IST (55-59) and PXL (60-65).
z-axis is cut... < 100 entries are eliminated. Note that the z-axis is scaled by 1000. Looks like a good (enough) correlation to proceed.
NOTE -- in debugging these plots, I found that the base QA code has an issue with respect to clone tracks. TPC only tracks are not considered when counting clones -- i.e. a track may be split into two. One is HFT, one is TPC only, and the base QA doesn't register this as a clone. The reason for this is that TPC only tracks are eliminated from study before the clone classification is made. Will need to revisit this.
»
- jwebb's blog
- Login or register to post comments