sPHENIX tracking discussion

[Christof Roland <Christof.Roland AT cern.ch>]

Hi Ben,

I agree with Ross, A plot of mean and variance across the accessible x-y coordinates of the detector would be nice. 

Also I think it would be good to see the time sequence for on matching position. Is the (supstantial) spread in r and phi this a random resolution effect or do we really see fluctuations...

Currently the spread in R and phi is on the order of a few millimeters and R at least is also a strong function of threshold. 

On the scale of O(150um) we want to achieve we need to understand this better. 

Cheers

   Christof 

On 8. Aug 2024, at 21:53, Ross Corliss <sphenix-tracking-l AT lists.bnl.gov> wrote:

Thanks for sharing this! 

So is the dr dphi spread a real effect of different places in the detector? Your spot-check of #1600 certainly hints that could be the case. A plot of distribution mean, and one of variance, as a function of (r, phi) position would give a compact way to visualize whether you are seeing weird artifacts, or a smooth map of the distortion. 

- Ross

On Thu, Aug 8, 2024, 14:20 "Kimelman, Benjamin" <sphenix-tracking-l AT lists.bnl.gov> wrote:

Hi all,

I've put together some slides following up on the figures I showed today about changing to a nearest neighbor matching scheme. I've zoomed in as Christof suggested and also selected on just a single truth position as Ross suggested (I selected on truthIndex 1600, which is petal 0, row 16, phi 0). These clearly show that the NN scheme is more robust and gives better performance.

My slides don't include every histogram from all pedestal subtraction thresholds, so I'm including files that have those as well.

Let me know if you have any questions or feedback.

Benjamin Kimelman Postdoctoral Scholar, Dept. of Physics & Astronomy Vanderbilt University benjamin.kimelman AT vanderbilt.edu Pronouns: He/him/his