Context Navigation

Section_2

Timestamp:: Oct 15, 2013, 6:28:19 PM (11 years ago)
Author:: Christophe Delaere
Comment:: --

Legend:

: Unmodified
: Added
: Removed
: Modified

Private/RefereeComments/Section_2

-              v17
+              v18
 make such an unrealistic assumption.
 >>> Showers are not produced in Delphes as the particle momenta are simply smeared
 >>> according to the relevant calorimeter resolution. The discreteness of the calorimeter
 >>> has simply an impact on the (eta,phi) resolution of the final observables.
 >>> As it is, the pessimistic spatial resolution assumed in ECAL only affects photons,
 >>> since for electrons we have the track information and we assume infinite (eta,phi) resolution.
 >>> The "computational reasons" are that with the current state of the implementation, the combinatorics are
 >>> reduced, and this has an impact on the the jet clustering procedure, especially in the presence of pile-up.
+> Showers are not produced in Delphes as the particle momenta are simply smeared
+> according to the relevant calorimeter resolution. The discreteness of the calorimeter
+> has simply an impact on the (eta,phi) resolution of the final observables.
+> As it is, the pessimistic spatial resolution assumed in ECAL only affects photons,
+> since for electrons we have the track information and we assume infinite (eta,phi) resolution.
+> The "computational reasons" are that with the current state of the implementation, the combinatorics are
+> reduced, and this has an impact on the the jet clustering procedure, especially in the presence of pile-up.
+>>(CD) I would add that improvement with this respect are considered for future releases.
 Par 3:
 …
  * What is the physics motivation for doing a log-normal instead of a Gaussian smearing?
+>>> The lognomal distribution resembles to a gaussian when mean > 6*sigma, that is for most values at high energy, but has the advantage at low energy of being always positive. This ensures to avoid the positive bias in the effective mean and the >>> s.d induced by having a truncated gaussian.
+> The lognomal distribution resembles to a gaussian when mean > 6*sigma, that is for most values at high energy, but has the advantage at low energy of being always positive. This ensures to avoid the positive bias in the effective mean and the s.d induced by having a truncated gaussian.
+>> (CD) add: in that sense, this is a purely ad hoc choice.
  * To define a log-normal distribution, one usually gives the mean and sigma of the logarithm of the distribution, which is normal. Here, are the authors talking about the mean and variance of the log-normal distribution? I guess so, but it would be good to clarify.
+>>>We are talking about the mean and the variance of the lognormal, which have pretty complicated expression in terms of the mean and variance of the normal variable, which, as the referee correctly remarked, are the ones that are usually given. >>>However, it is clear from the text that m, and s are the mean and variance of the log-normal distribution and not the mean and variance of the normal distribution.
+>>>We are talking about the mean and the variance of the lognormal, which have pretty complicated expression in terms of the mean and variance of the normal variable, which, as the referee correctly remarked, are the ones that are usually given.
+>>>However, it is clear from the text that m, and s are the mean and variance of the log-normal distribution and not the mean and variance of the normal distribution.
 …
 The authors may want to be more explicit about their motives here.
 >>> If the detector is not highly granular and cell size are irregular, spikes can appear in angular distributions, if proper binning is not chosen. Therefore we apply uniform smearing for purely cosmetical reasons, and has indeed no impact whatsoever on the angular resolution of calorimeter towers.
+> If the detector is not highly granular and cell size are irregular, spikes can appear in angular distributions, if proper binning is not chosen. Therefore we apply uniform smearing for purely cosmetical reasons, and has indeed no impact whatsoever on the angular resolution of calorimeter towers.
+>> (CD) I know what you mean, but it is still impossible to understand without more details if you didn't face the problem. In particular, it is difficult to believe that you will still see discrete position after jet reconstruction. I think that this answer must be more explicit.
 == Section 2.3 ==
 …
 content is not related whatsoever to the information in this bullet.
 >>> the discussion about hits is not present anymore nethier in the text nor in the actual implementation of the algorithm.
+>>> the discussion about hits is not present anymore neither in the text nor in the actual implementation of the algorithm.
 PAGE 5
 …
 enough to detect it, is almost correct for most detector designs.
 >>> Although this is an un-realistic assumption, it shows to correctly reproduce the performances. However, as the referee correctly points out,
+>>> Although this is an un-realistic assumption, it shows to correctly reproduce the performance. However, as the referee correctly points out,
 >>> this assumptions restricts the variety of detector studies that can be performed with Delphes. We point out, however, the Delphes is primarily a tool used
 >>> by phenomenologists, who do not need to make advanced detector studies. We believe that these should be done inside experimental collaborations by means of
 …
 good resolution of jet pT at low pT.
+>>> The version of the energy-flow algorithm presented here, was designed specifically to address these concerns by the referee.
+>>> All the reconstructed tracks are now stored as energy-flow tracks.
+>>> Fake-neutral hadrons are now created when the calorimeter energy excesses that of the tracks.
+> The version of the energy-flow algorithm presented here, was designed specifically to address these concerns by the referee.
+> All the reconstructed tracks are now stored as energy-flow tracks.
+> Fake-neutral hadrons are now created when the calorimeter energy excesses that of the tracks.
+>> (CD) shouldn't we add few words saying that the old scheme was a good approximation in the absence of PU, but that indeed it was noticed to be insufficient when PU is introduced?
 Last par, last line : It is not true that the emulation of the PF algorithm
 …
 resolution parameters used in DELPHES to get to this performance.
+>>> The disagreement in Figure 5 has been sensibly reduced with the present energy-flow implementation.
+>>> A small discrepancy is still observed in the low energy 20-40 GeV. The Delphes authors are satisfied with the present performance,
+>>> as most of the discrepancy is observed in the 20-30 GeV bin: with increasing luminosity (trigger rates, pile-up),
+>>> most physics analyses will reject such low energy jets anyways.
+> The disagreement in Figure 5 has been sensibly reduced with the present energy-flow implementation.
+> A small discrepancy is still observed in the low energy 20-40 GeV. The Delphes authors are satisfied with the present performance,
+> as most of the discrepancy is observed in the 20-30 GeV bin: with increasing luminosity (trigger rates, pile-up),
+> most physics analyses will reject such low energy jets anyways.
+>>(CD) please don't say "the Delphes authors"... we is much better. Then, is it 20-40 or 20-30? Also instead of saying that that bin is not relevant (it is), I would say that even for that bin the resolution is reasonable (as this is the absolute resolution that matters, not the relative difference).