Per pulse pedestal subtraction

• It has been observed in p4455, that there are per pulse pedestal fluctuations relative. Currently only a single pedestal for all pulses is assumed.
• This MR fixes that.

https://max-jhub.desy.de/hub/user-redirect/lab/tree/GPFS/exfel/data/scratch/danilo/notebooks/ForLukasSep19_p4455r141.ipynb

assigned to @danilo

added 1 commit

• cc45e8f2 - Subtract pedestal per pulse.

Compare with previous version

changed title from Draft: Fix/per pulse pedestal subtraction to Draft: Per pulse pedestal subtraction

requested review from @schmidtp

marked this merge request as ready

Did this improve the predictions?

Implementation-wise LGTM.

Two questions:

• Since you still have a global pedestal, you could somehow expose the relationship between global and pulse-wise pedestals
• I see you're taking the pedestals after averaging on line 514. How many traces do you average here? The ADQs typically suffer from common mode effects: