Difference between revisions of "Mattione Update 08212013"

From GlueXWiki
Jump to: navigation, search
(Accuracy)
Line 27: Line 27:
 
| [[Image:Mattione_Update_08202013_NumTimeBased_Current.png|thumb|left|500px|Current (08/20/13) Code]]
 
| [[Image:Mattione_Update_08202013_NumTimeBased_Current.png|thumb|left|500px|Current (08/20/13) Code]]
 
| [[Image:Mattione_Update_08202013_NumTimeBased_Spiral.png|thumb|left|500px|Spiral Code]]
 
| [[Image:Mattione_Update_08202013_NumTimeBased_Spiral.png|thumb|left|500px|Spiral Code]]
|}
 
 
* # Time-Based Tracks (with PID = &pi;<sup>+</sup>) with Hit Selector DISABLED
 
** Hit Selector:
 
*** Runs just prior wire-based & time-based tracking.
 
*** Re-selects which hits to use for track fitting.
 
*** In theory this improves things, because with better guesses for the track momentum and position, you should be able to better select the hits along that trajectory.
 
*** In practice, this can pick up many extra, unwanted hits for spiraling tracks
 
*** Recommendation: Test with the hit selector set to ONLY reject hits, NOT to acquire new ones.
 
{| class="wikitable"
 
| [[Image:Mattione_Update_08202013_NumTimeBased_Current_NoSelector.png|thumb|left|500px|Current (08/20/13) Code, No Hit Selector]]
 
| [[Image:Mattione_Update_08202013_NumTimeBased_Spiral_NoSelector.png|thumb|left|500px|Spiral Code, No Hit Selector]]
 
 
|}
 
|}
  
Line 71: Line 59:
 
|}
 
|}
  
 +
* % of events with at least one track that is "roughly close" to the thrown parameters:
 +
** "Roughly Close": p within 20%, &theta; within 15 degrees, &phi; within 30 degrees, Vertex-Z within 10cm
 +
{| border="1" cellpadding="2" align="left" valign="top" style="text-align:center;"
 +
!width="100"| Stage
 +
!width="200"| Current (08/20/13) Code
 +
!width="200"| Spiral Code
 +
|-
 +
! Track Candidates
 +
| 28.9% || 33.0%
 +
|-
 +
! Time-Based Tracks
 +
| 5.7% || 5.8%
 +
|}
 +
 +
== Hit Selector ==
 +
=== Multiplicity ===
 +
* # Time-Based Tracks (with PID = &pi;<sup>+</sup>) with Hit Selector DISABLED
 +
** Hit Selector:
 +
*** Runs just prior wire-based & time-based tracking.
 +
*** Re-selects which hits to use for track fitting.
 +
*** In theory this improves things, because with better guesses for the track momentum and position, you should be able to better select the hits along that trajectory.
 +
*** In practice, this can pick up many extra, unwanted hits for spiraling tracks
 +
*** Recommendation: Test with the hit selector set to ONLY reject hits, NOT to acquire new ones.
 +
{| class="wikitable"
 +
| [[Image:Mattione_Update_08202013_NumTimeBased_Current_NoSelector.png|thumb|left|500px|Current (08/20/13) Code, No Hit Selector]]
 +
| [[Image:Mattione_Update_08202013_NumTimeBased_Spiral_NoSelector.png|thumb|left|500px|Spiral Code, No Hit Selector]]
 +
|}
 +
 +
=== Accuracy ===
 
* % of events with at least one track that is "roughly close" to the thrown parameters:
 
* % of events with at least one track that is "roughly close" to the thrown parameters:
 
** "Roughly Close": p within 20%, &theta; within 15 degrees, &phi; within 30 degrees, Vertex-Z within 10cm
 
** "Roughly Close": p within 20%, &theta; within 15 degrees, &phi; within 30 degrees, Vertex-Z within 10cm

Revision as of 00:44, 21 August 2013

Track Candidate Reconstruction Sampling

  • All events have one π+ thrown at p = 150 MeV/c, θ = 90 degrees, φ = random.
Event # Truth Hits Reconstructed Hits Current (08/20/13) Code Track Candidates Spiral Code Track Candidates
Event 1
Mattione Update 08202013 SlowPiPlus Event1 TruthHits.png
Mattione Update 08202013 SlowPiPlus Event1 ReconHits.png
Mattione Update 08202013 SlowPiPlus Event1 Current.png
Mattione Update 08202013 SlowPiPlus Event1 Spiral.png


Performance

  • 1000 events, all with one π+ thrown at p = 150 MeV/c, θ = 90 degrees, φ = random.

Multiplicity

  • # Track Candidates
Current (08/20/13) Code
Spiral Code
  • # Time-Based Tracks (with PID = π+)
Current (08/20/13) Code
Spiral Code
  • Average Event Reconstruction Rate:
    • Current (08/20/13) Code: 7.5Hz
    • Spiral Code: 17.5Hz

Accuracy

  • Track Candidates: Δp/p
Current (08/20/13) Code
Spiral Code (large spike at p = 0 and NaN...)
  • Track Candidates: Δθ
Current (08/20/13) Code
Spiral Code
  • Track Candidates: Δφ
Current (08/20/13) Code
Spiral Code
  • Track Candidates: ΔVertex-Z
Current (08/20/13) Code
Spiral Code
  •  % of events with at least one track that is "roughly close" to the thrown parameters:
    • "Roughly Close": p within 20%, θ within 15 degrees, φ within 30 degrees, Vertex-Z within 10cm
Stage Current (08/20/13) Code Spiral Code
Track Candidates 28.9% 33.0%
Time-Based Tracks 5.7% 5.8%

Hit Selector

Multiplicity

  • # Time-Based Tracks (with PID = π+) with Hit Selector DISABLED
    • Hit Selector:
      • Runs just prior wire-based & time-based tracking.
      • Re-selects which hits to use for track fitting.
      • In theory this improves things, because with better guesses for the track momentum and position, you should be able to better select the hits along that trajectory.
      • In practice, this can pick up many extra, unwanted hits for spiraling tracks
      • Recommendation: Test with the hit selector set to ONLY reject hits, NOT to acquire new ones.
Current (08/20/13) Code, No Hit Selector
Spiral Code, No Hit Selector

Accuracy

  •  % of events with at least one track that is "roughly close" to the thrown parameters:
    • "Roughly Close": p within 20%, θ within 15 degrees, φ within 30 degrees, Vertex-Z within 10cm
Stage Current (08/20/13) Code Spiral Code Current (08/20/13) Code, No Hit Selector Spiral Code, No Hit Selector
Track Candidates 28.9% 33.0% NA NA
Time-Based Tracks 5.7% 5.8% 11.3% 7.5%