Tune magic nsb

[jsitarek]

a script to derive the missing NSB noise in MAGIC part of the simulations and the modifications to the calibration to allow adding the noise also for MAGIC

[Elisa-Visentin]

Hi @jsitarek, I was checking the PR. Just a naïve question: why do we need two different NSB corrections for the 2 MAGIC? It is due to a different level of simulated (DL0) background, to the different features of the telescopes or to the (slightly) different FOV?

[jsitarek]

Hi @Elisa-Visentin
The code is done LST-style, i.e. you make a distribution of noise in MC and in data and using Gaussian approximations you add the difference. When you do the same procedure for M1 and M2 you would always get somewhat different results. You are right that in principle the difference should be small, but the two telescopes are not the same, in particular the afterpulse distribution in PMTs is different for M1 and M2, plus the simulation of electronic noise and NSB has some simplifications. From the tests that I was running I saw that the extra_noise parameters, both for dim and bright pixels end up very similar for M1 and M2, but there is a difference in added bias, it is very small for M2, but there is some bias to be added in M1.

[Elisa-Visentin]

Ok thank you!

[Elisa-Visentin]

There seems to be a failing test due to no M1 M2 stereo events in the training sample. I will check it on my laptop asap just to be sure that everything is fine (i.e., only statistics)

[Elisa-Visentin]

Tests to be fixed before merging

[codecov]

Codecov Report

Attention: Patch coverage is 92.85714% with 1 line in your changes missing coverage. Please review.

Project coverage is 77.23%. Comparing base (2c7d9b4) to head (29129f3).
Report is 1 commits behind head on master.

Files	Patch %	Lines
magicctapipe/image/calib.py	92.30%	1 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##           master     #236      +/-   ##
==========================================
+ Coverage   77.19%   77.23%   +0.04%     
==========================================
  Files          21       21              
  Lines        2605     2614       +9     
==========================================
+ Hits         2011     2019       +8     
- Misses        594      595       +1     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.