[G12] GFlux
Eugene Pasyuk
pasyuk at jlab.org
Sat Mar 29 09:10:32 EDT 2014
Excellent!
If you used exactly 6 MeV binning it looks exactly as expected. This is very good illustration of what happens when discrete variable is treated as continuous. All these spikes and dips are artefacts of your bins not being exact multiples of original photon energy bins.
In g12 experiment the width of E-bin is NOT exactly 6 MeV. Moreover, it is NOT a constant. This effect is very similar to the beating of two frequencies in acoustics.
There are two ways around this problem.
1. Use binning in exact multiples of E-bins. This will lead to non-equidistant binning in photon energy. But there is nothing wrong with it. Having one bin 50.3 MeV and the other 51.2 MeV is perfectly fine. Bins don't have to be "round" numbers.
2. Convert photon energy back from discrete to continuous using randomization within E-bin.
Whichever approach you choose it must be applied to both yield flux binning the same way.
Another remark related to binned representation of variables. Most often people attribute the value of the variable to the center of the bin. This is correct for more or less flat dependence. In general case one needs to atribut bin value to the weighted average of the variable within this bin. This is particularly important for rapidly changing variables.
-Eugene
----- Original Message -----
> From: "Michael C. Kunkel" <mkunkel at jlab.org>
> To: "Eugene Pasyuk" <pasyuk at jlab.org>
> Cc: "g12" <g12 at jlab.org>, "Johann Goetz" <theodore.goetz at gmail.com>
> Sent: Friday, March 28, 2014 5:15:05 PM
> Subject: Re: [G12] GFlux
> Greetings,
> If I understood Eugene correctly, then what I did was plot the beam
> energy in 6 MeV bins for all events from the pi0.
> I also did this for the Glux, by creating a histogram of the output
> of Johann's glfux calculator.
> I am attaching the results, but saving judgment until my mistakes are
> pointed out.
> BR
> MK
> On 3/28/14 4:27 PM, Eugene Pasyuk wrote:
> > reaction does not matter, pi0 or anything else
>
> > -Eugene
>
> > ----- Original Message -----
>
> > > From: "Michael C. Kunkel" <mkunkel at jlab.org>
> >
>
> > > To: "Eugene Pasyuk" <pasyuk at jlab.org>
> >
>
> > > Cc: "g12" <g12 at jlab.org> , "Johann Goetz"
> > > <theodore.goetz at gmail.com>
> >
>
> > > Sent: Friday, March 28, 2014 4:22:52 PM
> >
>
> > > Subject: Re: [G12] GFlux
> >
>
> > > Greetings,
> >
>
> > > I am wondering if Johann's trip program takes into account the
> > > tagger
> > > hit status to be 7 or 15 as well as does the GFlux calculator
> > > when
> > > using the "good" option take this into account?
> >
>
> > > Also, Eugene, when you wrote to look at the flux and yield and
> > > normalized yield per E-bin, would the yield of pi0 suffice?
> >
>
> > > BR
> >
>
> > > MK
> >
>
> > > On 3/28/14 4:13 PM, Eugene Pasyuk wrote:
> >
>
> > > > The other thing is to make sure when you select your events and
> > > > match
> > > > them to the tagger you use tagger hits with status 7 or 15
> > > > only.
> > >
> >
>
> > > > -Eugene
> > >
> >
>
> > > > ----- Original Message -----
> > >
> >
>
> > > > > From: "Eugene Pasyuk" <pasyuk at jlab.org>
> > > >
> > >
> >
>
> > > > > To: "Michael C. Kunkel" <mkunkel at jlab.org>
> > > >
> > >
> >
>
> > > > > Cc: "g12" <g12 at jlab.org> , "Johann Goetz"
> > > > > <theodore.goetz at gmail.com>
> > > >
> > >
> >
>
> > > > > Sent: Friday, March 28, 2014 4:04:41 PM
> > > >
> > >
> >
>
> > > > > Subject: Re: [G12] GFlux
> > > >
> > >
> >
>
> > > > > The way to figure out is to look at the flux and yield and
> > > > > normalized
> > > > > yield per E-bin, that is the finest binning. If they are
> > > > > consistent,
> > > > > the the problem is in re-binning procedure. If they are not,
> > > > > then
> > > > > we
> > > > > have to look further. Technically speaking there is not
> > > > > unusual
> > > > > that
> > > > > you may have holes on photon spectrum, that would just
> > > > > indicate
> > > > > that
> > > > > some of the counters are dead.
> > > >
> > >
> >
>
> > > > > -Eugene
> > > >
> > >
> >
>
> > > > > ----- Original Message -----
> > > >
> > >
> >
>
> > > > > > From: "Michael C. Kunkel" <mkunkel at jlab.org>
> > > > >
> > > >
> > >
> >
>
> > > > > > To: "Eugene Pasyuk" <pasyuk at jlab.org> , "Johann Goetz"
> > > > > > <theodore.goetz at gmail.com>
> > > > >
> > > >
> > >
> >
>
> > > > > > Cc: "g12" <g12 at jlab.org>
> > > > >
> > > >
> > >
> >
>
> > > > > > Sent: Friday, March 28, 2014 3:55:45 PM
> > > > >
> > > >
> > >
> >
>
> > > > > > Subject: Re: [G12] GFlux
> > > > >
> > > >
> > >
> >
>
> > > > > > Greetings,
> > > > >
> > > >
> > >
> >
>
> > > > > > I am wondering where we are at with the flux issue?
> > > > >
> > > >
> > >
> >
>
> > > > > > Meaning, what is going on with the flux in the 3.5 GeV
> > > > > > region?
> > > > >
> > > >
> > >
> >
>
> > > > > > I feel that the flux is in error for a few reasons.
> > > > >
> > > >
> > >
> >
>
> > > > > > First when I bin flux in 15 MeV, I see 0 flux at 3.515 -
> > > > > > 3.53.
> > > > >
> > > >
> > >
> >
>
> > > > > > Second, when I bin flux in 50 MeV binning I see a drop in
> > > > > > flux
> > > > > > at
> > > > > > 3.5
> > > > > > MeV
> > > > >
> > > >
> > >
> >
>
> > > > > > Here is the output
> > > > >
> > > >
> > >
> >
>
> > > > > > //
> > > > >
> > > >
> > >
> >
>
> > > > > > 3.45 3.5 1.06825209003e+12
> > > > >
> > > >
> > >
> >
>
> > > > > > 3.5 3.55 470202243457.0
> > > > >
> > > >
> > >
> >
>
> > > > > > 3.55 3.6 1.69234003548e+12
> > > > >
> > > >
> > >
> >
>
> > > > > > //
> > > > >
> > > >
> > >
> >
>
> > > > > > Third reason is that if you look at my attached pdf, I show
> > > > > > the
> > > > > > pi0
> > > > > > cross section for various energies, if you were to go to
> > > > > > compare
> > > > > > the
> > > > > > cross sections around 3.5 GeV to the neighboring Egamma
> > > > > > bins,
> > > > > > you
> > > > > > will notice a sizable discrepancy. See G12_XSection_ALL.pdf
> > > > >
> > > >
> > >
> >
>
> > > > > > Fourth reason, if I were to compare the pi0 cross section
> > > > > > in
> > > > > > unit
> > > > > > of
> > > > > > dsigma/dt and compare it other experiments at 4 and 5 GeV,
> > > > > > our
> > > > > > cross
> > > > > > section is off by a factor of 2, see
> > > > > > t_vs_dsigdt_Compare.pdf
> > > > >
> > > >
> > >
> >
>
> > > > > > In order for any g12 papers to be published, we need
> > > > > > verification
> > > > > > of
> > > > > > cross sections, which as of this moment I do not believe we
> > > > > > have.
> > > > >
> > > >
> > >
> >
>
> > > > > > BR
> > > > >
> > > >
> > >
> >
>
> > > > > > MK
> > > > >
> > > >
> > >
> >
>
> > > > > > On 3/25/14 10:16 PM, Eugene Pasyuk wrote:
> > > > >
> > > >
> > >
> >
>
> > > > > > > This is not quite correct, especially when bins are small
> > > > > > > and
> > > > > > > comparable with the width of E-bin.
> > > > > >
> > > > >
> > > >
> > >
> >
>
> > > > > > > Just imagine, You energy bin is 3.000-3.015 GeV. E-bin
> > > > > > > let's
> > > > > > > say
> > > > > > > covers 3.0121-3.0181 GeV. The center of E-bin will be
> > > > > > > 3.01505
> > > > > > > GeV.
> > > > > > > When you enter this into the histogram it will fall to
> > > > > > > 3.015-3.030
> > > > > > > bin then. However, this E-bin should be split between two
> > > > > > > adjacent
> > > > > > > bins in new binning scheme. There are two possible ways
> > > > > > > of
> > > > > > > doing
> > > > > > > this. Either randomize energy within the width of the
> > > > > > > E-bin
> > > > > > > before
> > > > > > > you enter it to th ehistogram, or split it between
> > > > > > > adjacent
> > > > > > > bins
> > > > > > > using weight which depends on how the bin boundary splits
> > > > > > > E-bin
> > > > > > > and
> > > > > > > have two entries to the histogram with sum of the weights
> > > > > > > equal
> > > > > > > one.
> > > > > > > You can imagine, if the bin width is chosen to be less
> > > > > > > than
> > > > > > > the
> > > > > > > widths of E-bin, then you will get completely empty bins
> > > > > > > in
> > > > > > > resulting histogram.
> > > > > >
> > > > >
> > > >
> > >
> >
>
> > > > > > > -Eugene
> > > > > >
> > > > >
> > > >
> > >
> >
>
> > > > > > > ----- Original Message -----
> > > > > >
> > > > >
> > > >
> > >
> >
>
> > > > > > > > From: "Johann Goetz" <theodore.goetz at gmail.com>
> > > > > > >
> > > > > >
> > > > >
> > > >
> > >
> >
>
> > > > > > > > To: "Eugene Pasyuk" <pasyuk at jlab.org>
> > > > > > >
> > > > > >
> > > > >
> > > >
> > >
> >
>
> > > > > > > > Sent: Tuesday, March 25, 2014 9:12:49 PM
> > > > > > >
> > > > > >
> > > > >
> > > >
> > >
> >
>
> > > > > > > > Subject: Re: [G12] GFlux
> > > > > > >
> > > > > >
> > > > >
> > > >
> > >
> >
>
> > > > > > > > I consider the beam energy to be the center of the
> > > > > > > > energy
> > > > > > > > paddle
> > > > > > > > and
> > > > > > > > fill the histogram at that energy with the number if
> > > > > > > > counts
> > > > > > > > in
> > > > > > > > the
> > > > > > > > gflux file. Is that wrong?
> > > > > > >
> > > > > >
> > > > >
> > > >
> > >
> >
>
> > > > > > > > On Mar 25, 2014 5:59 PM, "Eugene Pasyuk" <
> > > > > > > > pasyuk at jlab.org
> > > > > > > > >
> > > > > > > > wrote:
> > > > > > >
> > > > > >
> > > > >
> > > >
> > >
> >
>
> > > > > > > > > I see four numbers in each row in this file. I
> > > > > > > > > figured
> > > > > > > > > the
> > > > > > > > > first
> > > > > > > > > one
> > > > > > > > > is E-bin number. What are the other three? I guess
> > > > > > > > > the
> > > > > > > > > second
> > > > > > > > > and
> > > > > > > > > the fourth are some energies. What is the third
> > > > > > > > > number
> > > > > > > > > which
> > > > > > > > > is
> > > > > > > > > integer and varies a lot.
> > > > > > > >
> > > > > > >
> > > > > >
> > > > >
> > > >
> > >
> >
>
> > > > > > > > > When you say you "histogram" does it mean you take
> > > > > > > > > one
> > > > > > > > > fixed
> > > > > > > > > number
> > > > > > > > > for each E-bin (beam centroid?) and put it in the
> > > > > > > > > histogram,
> > > > > > > > > or
> > > > > > > > > something else?
> > > > > > > >
> > > > > > >
> > > > > >
> > > > >
> > > >
> > >
> >
>
> > > > > > > > > -Eugene
> > > > > > > >
> > > > > > >
> > > > > >
> > > > >
> > > >
> > >
> >
>
> > > > > > > > > > From: "Johann Goetz" < theodore.goetz at gmail.com >
> > > > > > > > >
> > > > > > > >
> > > > > > >
> > > > > >
> > > > >
> > > >
> > >
> >
>
> > > > > > > > > > To: "g12" < g12 at jlab.org >
> > > > > > > > >
> > > > > > > >
> > > > > > >
> > > > > >
> > > > >
> > > >
> > >
> >
>
> > > > > > > > > > Sent: Tuesday, March 25, 2014 5:40:42 PM
> > > > > > > > >
> > > > > > > >
> > > > > > >
> > > > > >
> > > > >
> > > >
> > >
> >
>
> > > > > > > > > > Subject: Re: [G12] GFlux
> > > > > > > > >
> > > > > > > >
> > > > > > >
> > > > > >
> > > > >
> > > >
> > >
> >
>
> > > > > > > > > > The output of gflux includes a file that has the
> > > > > > > > > > energies
> > > > > > > > > > of
> > > > > > > > > > the
> > > > > > > > > > individual energy paddles and the counts. These
> > > > > > > > > > files
> > > > > > > > > > are
> > > > > > > > > > then
> > > > > > > > > > used
> > > > > > > > > > to fill a histogram in the g12-gflux script where
> > > > > > > > > > the
> > > > > > > > > > range
> > > > > > > > > > and
> > > > > > > > > > bins
> > > > > > > > > > are specified on the command line. The energies are
> > > > > > > > > > slightly
> > > > > > > > > > "corrected" to the ones found in the following
> > > > > > > > > > file:
> > > > > > > > >
> > > > > > > >
> > > > > > >
> > > > > >
> > > > >
> > > >
> > >
> >
>
> > > > > > > > > > /group/clas/parms/pcor/g12/tagger_energies.dat
> > > > > > > > >
> > > > > > > >
> > > > > > >
> > > > > >
> > > > >
> > > >
> > >
> >
>
> > > > > > > > > > But, the correction is very small and averaged over
> > > > > > > > > > the
> > > > > > > > > > whole
> > > > > > > > > > run
> > > > > > > > > > period (i.e. not one a run-by-run basis).
> > > > > > > > >
> > > > > > > >
> > > > > > >
> > > > > >
> > > > >
> > > >
> > >
> >
>
> > > > > > > > > > And before you ask, the the summing of the flux is
> > > > > > > > > > valid
> > > > > > > > > > to
> > > > > > > > > > at
> > > > > > > > > > least
> > > > > > > > > > 4 figures - probably 5.
> > > > > > > > >
> > > > > > > >
> > > > > > >
> > > > > >
> > > > >
> > > >
> > >
> >
>
> > > > > > > > > > -- Johann
> > > > > > > > >
> > > > > > > >
> > > > > > >
> > > > > >
> > > > >
> > > >
> > >
> >
>
> > > > > > > > > > On Tue, Mar 25, 2014 at 4:28 PM, Michael C. Kunkel
> > > > > > > > > > <
> > > > > > > > > > mkunkel at jlab.org
> > > > > > > > > > > wrote:
> > > > > > > > >
> > > > > > > >
> > > > > > >
> > > > > >
> > > > >
> > > >
> > >
> >
>
> > > > > > > > > > > Greetings,
> > > > > > > > > >
> > > > > > > > >
> > > > > > > >
> > > > > > >
> > > > > >
> > > > >
> > > >
> > >
> >
>
> > > > > > > > > > > If the data shows a hole at beam energies, how
> > > > > > > > > > > does
> > > > > > > > > > > the
> > > > > > > > > > > gflux
> > > > > > > > > > > for
> > > > > > > > > > > this
> > > > > > > > > >
> > > > > > > > >
> > > > > > > >
> > > > > > >
> > > > > >
> > > > >
> > > >
> > >
> >
>
> > > > > > > > > > > energy have events?
> > > > > > > > > >
> > > > > > > > >
> > > > > > > >
> > > > > > >
> > > > > >
> > > > >
> > > >
> > >
> >
>
> > > > > > > > > > > Below is a table of Gflux for 15MeV binning,
> > > > > > > > > > > which
> > > > > > > > > > > is
> > > > > > > > > > > to
> > > > > > > > > > > fine
> > > > > > > > > > > for
> > > > > > > > > >
> > > > > > > > >
> > > > > > > >
> > > > > > >
> > > > > >
> > > > >
> > > >
> > >
> >
>
> > > > > > > > > > > analysis but great for illustrating this
> > > > > > > > > > > question.
> > > > > > > > > >
> > > > > > > > >
> > > > > > > >
> > > > > > >
> > > > > >
> > > > >
> > > >
> > >
> >
>
> > > > > > > > > > > Also, notice the 0 entries at 3.5 GeV, and low
> > > > > > > > > > > events
> > > > > > > > > > > in
> > > > > > > > > > > neighbors.
> > > > > > > > > >
> > > > > > > > >
> > > > > > > >
> > > > > > >
> > > > > >
> > > > >
> > > >
> > >
> >
>
> > > > > > > > > > > FORMAT: Emin:Emax:Flux
> > > > > > > > > >
> > > > > > > > >
> > > > > > > >
> > > > > > >
> > > > > >
> > > > >
> > > >
> > >
> >
>
> > > > > > > > > > > _______________________________________________
> > > > > > > > > >
> > > > > > > > >
> > > > > > > >
> > > > > > >
> > > > > >
> > > > >
> > > >
> > >
> >
>
> > > > > > > > > > > G12 mailing list
> > > > > > > > > >
> > > > > > > > >
> > > > > > > >
> > > > > > >
> > > > > >
> > > > >
> > > >
> > >
> >
>
> > > > > > > > > > > G12 at jlab.org
> > > > > > > > > >
> > > > > > > > >
> > > > > > > >
> > > > > > >
> > > > > >
> > > > >
> > > >
> > >
> >
>
> > > > > > > > > > > https://mailman.jlab.org/mailman/listinfo/g12
> > > > > > > > > >
> > > > > > > > >
> > > > > > > >
> > > > > > >
> > > > > >
> > > > >
> > > >
> > >
> >
>
> > > > > > > > > > _______________________________________________
> > > > > > > > >
> > > > > > > >
> > > > > > >
> > > > > >
> > > > >
> > > >
> > >
> >
>
> > > > > > > > > > G12 mailing list
> > > > > > > > >
> > > > > > > >
> > > > > > >
> > > > > >
> > > > >
> > > >
> > >
> >
>
> > > > > > > > > > G12 at jlab.org
> > > > > > > > >
> > > > > > > >
> > > > > > >
> > > > > >
> > > > >
> > > >
> > >
> >
>
> > > > > > > > > > https://mailman.jlab.org/mailman/listinfo/g12
> > > > > > > > >
> > > > > > > >
> > > > > > >
> > > > > >
> > > > >
> > > >
> > >
> >
>
> > > > > > > _______________________________________________
> > > > > >
> > > > >
> > > >
> > >
> >
>
> > > > > > > G12 mailing list G12 at jlab.org
> > > > > > > https://mailman.jlab.org/mailman/listinfo/g12
> > > > > >
> > > > >
> > > >
> > >
> >
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: https://mailman.jlab.org/pipermail/g12/attachments/20140329/2a6d380f/attachment-0001.html
More information about the G12
mailing list