Address simulated ECAL energy resolution

[baltzell]

For now, just moving to uncorrelated noise on each sample, while still applying the photoelectron/statistical resolution separately on the pulse integral. Previously the incorrect treatment of the noise effect rendered it zero. Effect should be largest at lower energies, but no feeling for scale.

Need to get some simulations run to assess. Can anyone help with this?

[baltzell]

What all shall we look at to assess this and #978? Of course FEEs, just because they're easy/free. Two-cluster energy sum for Mollers would get sensitive to lower energies (where the problem and effect of this change should be larger) and still give an easily interpretable resolution number. And I think people are also interested in the trigger turn-on, but I haven't thought about how to assess that, other than just plotting cluster energy distribution around the trigger threshold.

[cbravo135]

You can look at the the momentum sum of tracks in V0s with an electron positron pair in opposite volumes. You could also look at the positron momentum distribution turn on comparing data and MC.

[baltzell]

I don't think we want to be looking at tracking for this.

[normangraf]

I don't think we want to be looking at tracking for this.

I concur.

[normangraf]

What all shall we look at to assess this and #978? Of course FEEs, just because they're easy/free. Two-cluster energy sum for Mollers would get sensitive to lower energies (where the problem and effect of this change should be larger) and still give an easily interpretable resolution number. And I think people are also interested in the trigger turn-on, but I haven't thought about how to assess that, other than just plotting cluster energy distribution around the trigger threshold.

I don't think the FEE trigger is really sensitive to this resolution, since the energy threshold for the trigger was well below the peak. I would look at the low end of the two-cluster esum plot in trident production since what we really want to probe is the positron turn-on curve.

[cbravo135]

We absolutely want to look at the tracks as a metric for this, they are how we noticed there is an issue in the first place. The shape of the threshold in tracks is an independent way to validate the trigger resolution. Using methods of validation like this will be critical to fixing the issue at hand. It is analytically a more robust method of validation, since the track momentum is independent of the information used to generate the trigger, we should not just use something fundamentally derived from the same information as a sole source of validation.

[baltzell]

Yeah, I agree, ultimately we'll need tracks for this (but first FEEs and Mollers would be good without tracking), can you set up some metric for this?

[baltzell]

Things are looking better now, see screenshot. Time for some more simulations to look at simple Mollers plus the pair/positron trigger turn on curve.

[bloodyyugo]

Looks pretty great...thanks for looking at this.

[tongtongcao]

The update for DigitizationWithPulserDataMergingReadoutDriver is similar to the update for DigitizationReadoutDriver. For pulser data overlay, there are two difference cases: 1) if a ADC sample is directly from random data, or merged with random data, we do not need to do anything, such as smearing hit energy, adding noise, and adding pedestal. 2) if a ADC sample is from MC signal, no merging with random data, we should do the same thing as we did for MC signal + MC beam as background.

[tongtongcao]

Analysis group will test effects of Ecal smearing updates. Let's wait for their feedbacks.