Choose timezone
Your profile timezone:
Reconstruction Group Meeting
→
US/Eastern
Description
Zoom coordinates:
- URL: https://lbnl.zoom.us/j/94910574347?pwd=c01YQlg5bG04dFR6UWs3NjZkS3ZSdz09
- Meeting ID: 949 1057 4347
- Password: recon
Meeting Notes
- See the google doc
Reconstruction WG Meeting Minutes
1. hpDIRC sensitive-volume PR
The PR makes the hpDIRC bars sensitive, providing hits for downstream geometric reconstruction.
Dmitrii asked that compatibility with the current simulation be checked, particularly whether making the bars sensitive also records absorbed Cherenkov photons. He explained that charged-particle hits would supply the particle angle for geometric reconstruction; reconstructed track projections should provide this information in the final reconstruction.
Chandradoy questioned the reliance on truth information and the effect on simulation savings. After Dmitrii explained that applying quantum efficiency early still saves photon transport and storage, Chandradoy withdrew the latter concern.
Follow-up: Check simulation compatibility and clarify with Bill whether the truth information is temporary or needed for building the geometric-reconstruction lookup tables. Chandradoy will review the PR.
2. PID outputs and replacement of existing lookup tables
Chandradoy requested a clearer assessment of progress toward replacing the existing PID lookup tables with reconstruction outputs usable by analyzers. He highlighted the lack of a consistent interface across detectors, including likelihoods, chi-square values, and their representation in EDM4eic.
Dmitrii and Chandradoy favored holding the broader discussion once hpDIRC reconstruction is working, with Bill involved. Derek emphasized the Reconstruction WG’s role in defining interfaces and identifying missing connections in the reconstruction chain.
Follow-up: Revisit available outputs and remaining gaps with the relevant PID developers and detector representatives. The lookup tables used internally by hpDIRC geometric reconstruction are distinct from the existing PID tables discussed for replacement.
3. Track covariance and secondary vertexing — PR 3024
Austin explained that the PR fills previously unimplemented position-covariance information when converting ACTS tracks to the EIC track object. It uses ACTS Jacobians to transform the covariance into global coordinates for positional errors and chi-square calculations in the secondary-vertex helix algorithm.
The PR appeared reasonable, with another review round suggested.
Austin also reported zero NDF values from AMVF. Discussion pointed to fractional NDF values and the need to agree on their meaning and storage; a data-model change is likely.
Follow-up: Austin will raise the NDF issue at the next Heavy Flavor group meeting. The vertexing experts should agree on a convention and bring any proposed data-model change back for discussion.
4. Magnetic-field contribution to dRICH angular resolution
Chandradoy wants to isolate the contribution of track bending within the radiator to the reconstructed Cherenkov-angle spread. Earlier studies using projection planes produced unexpectedly large uncertainties and unexplained covariance behavior.
Shujie noted that propagation to material-free surfaces is an appropriate approach, provided the magnetic field and material information are handled correctly. PR 3024 concerns a covariance-coordinate transformation and does not resolve this separate question.
Follow-up: Review the propagation implementation in a dedicated discussion. Chandradoy will try to document the implementation and an example of the excessive uncertainties in an issue around the end of November.
5. Timing in CKF — PR 3012 and related work
Shujie reviewed work to include timing from AC-LGAD and MPGD measurements while excluding SVT hit timing. The separate PR disabling SVT timing looked reasonable to her.
Two questions remained:
-
Seed timing: The current setup uses an artificial 10 ns initial timing uncertainty. An unconstrained seed was suggested, while Dmitrii argued that Kalman-filter initialization requires a covariance and that a large prior uncertainty is standard. No choice was settled.
-
Detector timing resolution: Dmitrii distinguished digitization smearing from the uncertainty supplied to reconstruction. He suggested sharing a configuration constant where the values should match. Shujie will inspect the implementation further.
Inspecting individual Kalman-filter updates was suggested to assess timing effects on chi-square and covariance.
Follow-up: Continue the discussion on Thursday with more participants; John may have results to present.
There are minutes attached to this event.
Show them.