Zoom session link:
Meeting ID: 940 3272 7802
Maarten: misreconstruction of energy is related to ommitance of delta-ray contributions. Not sure how this will change the plot on slide 23 of Lodewijk, but it has a sizeable effect.
The delta-ray contribution has a logarithmic dependence on the energy of the muon.
Brían's combined fit will bypass the current PID.
Could be interesting to have a continuous PID and use it as another fit parameter in MONA.
Maarten: Can we define the analytical derivatives for the oscillation probabilities and use it to speed up MONA?
Noone has investigated the input parameters for JMuon so far.
I have looked at the time-window before.
The roadwidth is what I studied next.
There is a sweet spot between 50 and 80.
I have also studied combinations of input parameters, for example combinations of the roadwidth and time-window.
The E_max does not make a big difference.
I have also looked at the 'resolution' parameter, which is the final step size in your minimization procedure.
There is also the parameter zmin, which I have not checked yet. At the moment it is set to -3.
Maarten: Would really be in favor of moving the roadwidth from 50m to 80 m.
In addition to the work above, I have one more plot on the combined likelihoods.
I can now look at every single event.
Yesterday I told that I found a discrepancy between the PDF and mc-truth of 2.0.'
This issue has been resolved.
Thinking of now moving to higher energy MC, for example to MuPage files.
I would also like to move from MC-hits to real hits, but I am not yet sure how my study transfers.
Maarten: look at the hits, sample from the PDF and subtract the contribution of a MIP to determine the position of Brehmsstrahlung-showers.
Timon: What I am doing at the moment is discretizing the track longitudinal profile in steps of 5m and assume a shower in each bin and evaluate the corresponding hit probabilities. Should see a clear dip at the position of a bremsstrahlung shower.
Maarten: Would advize to use smaller step size than 5m, but this sounds like a good approach.
Timon: Is there already a way to implement current fitting algorithms?
Brían: Not yet.
Jordan: I can show you how
The toy-MC is being pushed to AAnet.
It is contained in the rec2-folder.
Maarten: Just for my understanding: this is a toy-MC that generates hits, correct?
Jordan: Correct. There is the possibility to generate hits corresponding to a certain event file. But there is also the possibility to generate random hits.
Reco for JSirene still breaks down.
Not that much to say.
Working on the energy reconstruction.
Was also looking at a way to convert a ToT to a number of photo-electrons.
Maarten: in the new productions for each hit there is already a field A, which contains the number of photo-electrons (ToT -> npe conversion is already done for you in the new productions).
We are also working on including the dynamic position calibration in the hits.
How prepared are we for rerunning separate reconstruction-stages on processed data?
Valentin: I am currently working with the offline file-formats (evt files).