<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=us-ascii">
<style type="text/css" style="display:none;"> P {margin-top:0;margin-bottom:0;} </style>
</head>
<body dir="ltr">
<div style="font-family: Calibri, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
Elton,</div>
<div style="font-family: Calibri, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
<br>
</div>
<div style="font-family: Calibri, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
Colin is using the same procedure as described by Matt i.e. he is starting from the previous gains and energy dependence correction, iterate the gains until the pi0 resolution does not change, and then changes the energy dependence correction. Because the pi0
fitting (which is of poor quality) is systematically biased by 1 to 2 MeV when he doing the energy dependence correction the pi0 mass is shifted to lower mass and falls within 2-3% around the pi0 PDG mass and stops there. If one wants to achieve a +/- 1% mass
calibration precision, one needs to iterate again (in this procedure) until all calibrations converge or fall within 1% of the PDG mass.
<br>
</div>
<div style="font-family: Calibri, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
<br>
</div>
<div style="font-family: Calibri, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
Note that in my case or in the BCAL method i.e. do the gains calibration while the energy dependence correction is turned off and then do the energy dependence correction. The pi0 mass is also shifted to lower mass after application of the energy dependence
correction which I did not expect. It either means I am using wrong functions or the photon energy range is so large that it is unavoidable (which could well be).
<br>
</div>
<div style="font-family: Calibri, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
<br>
</div>
<div style="font-family: Calibri, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
tks ig.<br>
</div>
<div id="appendonsend"></div>
<hr style="display:inline-block;width:98%" tabindex="-1">
<div id="divRplyFwdMsg" dir="ltr"><font face="Calibri, sans-serif" style="font-size:11pt" color="#000000"><b>From:</b> Halld-cal <halld-cal-bounces@jlab.org> on behalf of Shepherd, Matthew <mashephe@indiana.edu><br>
<b>Sent:</b> Friday, May 1, 2020 7:44 AM<br>
<b>To:</b> Elton Smith <elton@jlab.org><br>
<b>Cc:</b> Hall-D Calorimetry <halld-cal@jlab.org><br>
<b>Subject:</b> Re: [Halld-cal] [EXTERNAL] FCAL calibration follow-up</font>
<div> </div>
</div>
<div class="BodyFragment"><font size="2"><span style="font-size:11pt;">
<div class="PlainText"><br>
Hi Elton,<br>
<br>
> On May 1, 2020, at 7:27 AM, Elton Smith <elton@jlab.org> wrote:<br>
> <br>
> Hi Matt,<br>
> <br>
> The reason to use non-linear corrections during the gain matching would be to minimize incorporating non-linear effects into the determined gains, as you point out. However, using non-linear corrections builds in the existing non-linearity corrections, which
may be wrong, into the gain factors. In either case, one will be sensitive to the non-linearities at some level. Iterating over applied corrections is somewhat tricky, which argues in favor of Collin's approach to do the gain calibration with raw inputs,
and then make the energy correction.<br>
<br>
Colin should clarify, but I don't think Colin uses raw inputs. At least that is not how the code is setup. You can check line #135 here:<br>
<br>
<a href="https://urldefense.proofpoint.com/v2/url?u=https-3A__github.com_JeffersonLab_halld-5Frecon_blob_master_src_plugins_Calibration_FCAL-5FPi0HFA_JEventProcessor-5FFCAL-5FPi0HFA.cc&d=DwIFAg&c=CJqEzB1piLOyyvZjb8YUQw&r=qOykDDW07dxNIz2gqzRW_0sRJeff1E30DnojDqga1Jo&m=ZSxvCFbmrwhwoEAuVlSkSYV-YsHgPOxW3OtXD1CV2GI&s=iKV9-Afb0D88NMero_q6cJU4EW5TLb5vEGcALtMouGY&e=">https://urldefense.proofpoint.com/v2/url?u=https-3A__github.com_JeffersonLab_halld-5Frecon_blob_master_src_plugins_Calibration_FCAL-5FPi0HFA_JEventProcessor-5FFCAL-5FPi0HFA.cc&d=DwIFAg&c=CJqEzB1piLOyyvZjb8YUQw&r=qOykDDW07dxNIz2gqzRW_0sRJeff1E30DnojDqga1Jo&m=ZSxvCFbmrwhwoEAuVlSkSYV-YsHgPOxW3OtXD1CV2GI&s=iKV9-Afb0D88NMero_q6cJU4EW5TLb5vEGcALtMouGY&e=</a>
<br>
<br>
Note that the pointers photon1 and photon2 are the corrected showers returned by the high-level analysis objects. Unless Colin has rewritten the DB with a set of nonlinear correction constants to effectively disable the non-linear correction, it will be applied.<br>
<br>
This procedure, as evidenced by years of experience now, seems to have worked and stabilized. I'm just raising this as potential mechanism to generate an apparent position dependence when, in fact, it one just has a uniform energy dependent response.
<br>
<br>
Maybe one could argue that the calibration mechanism now would take a true position dependence (e.g., like on that arises from Cherenkov angular distributions) and shoehorn it into an energy nonlinearity.
<br>
<br>
My point is more that we should be aware of and look for such effects if we desire to devise a new strategy for calibration as Igal is proposing.<br>
<br>
Matt<br>
<br>
<br>
_______________________________________________<br>
Halld-cal mailing list<br>
Halld-cal@jlab.org<br>
<a href="https://mailman.jlab.org/mailman/listinfo/halld-cal">https://mailman.jlab.org/mailman/listinfo/halld-cal</a><br>
</div>
</span></font></div>
</body>
</html>