Uploaded image for project: 'Data Management'
  1. Data Management
  2. DM-8533

Investigate "bump" in PSF vs cmodel at bright magnitudes

    XMLWordPrintable

    Details

    • Type: Story
    • Status: Done
    • Resolution: Done
    • Fix Version/s: None
    • Component/s: None
    • Labels:
      None
    • Story Points:
      10
    • Epic Link:
    • Sprint:
      DRP S17-1, DRP S17-2, DRP S17-3, DRP S17-4
    • Team:
      Data Release Production

      Description

      Per DM-6817,

      the CModel vs. psf plot on DM-6818 (plot-t0-HSC-I-mag_modelfit_CModel-psfMagHist.png) reveals... a "bump" at bright magnitudes in the "stars"... [This] seems to be rooted in the psf modeling at the single frame stage.

      Figure out what's going on here, and work out how to fix it.

        Attachments

          Issue Links

          is blocked by

          Story - Created by Jira Software - do not edit or delete. Issue type for a user story. DM-9028 Process the HSC RC dataset with the latest hscPipe 4.0.5

          • Done
          relates to

          Story - Created by Jira Software - do not edit or delete. Issue type for a user story. DM-9107 Investigate isotropic footprint growing/dilating behavior: HSC vs. LSST

          • Done

          Story - Created by Jira Software - do not edit or delete. Issue type for a user story. DM-9108 Investigate distribution of footprint sizes: HSC vs. LSST

          • Done

          Story - Created by Jira Software - do not edit or delete. Issue type for a user story. DM-9109 Create ellipticity residuals quiver plots

          • Done

          Story - Created by Jira Software - do not edit or delete. Issue type for a user story. DM-9110 Investigate source of systematic difference of model psf trace radii: HSC vs. LSST

          • Done

          Story - Created by Jira Software - do not edit or delete. Issue type for a user story. DM-9252 Investigate bimodal distribution of footprint sizes for "stars"

          • Done

          Story - Created by Jira Software - do not edit or delete. Issue type for a user story. DM-9383 Investigate propagation of visit flags for certain patches in HSC RC processing

          • Done

          Story - Created by Jira Software - do not edit or delete. Issue type for a user story. DM-9535 Assess whether differences in Brighter-Fatter implementations are contributing to the trace radii differences: LSST vs. HSC

          • Done

          Story - Created by Jira Software - do not edit or delete. Issue type for a user story. DM-9554 Process HSC "RC" dataset through the LSST stack using HSC's Brighter-Fatter implementation

          • Done

          Story - Created by Jira Software - do not edit or delete. Issue type for a user story. DM-6817 Compare HSC and LSST processing of RC dataset

          • Done

          Story - Created by Jira Software - do not edit or delete. Issue type for a user story. DM-6818 Quality check LSST processing of RC dataset

          • Done

            Activity

            Ascending order - Click to sort in descending order
            Hide
            Permalink
            lauren Lauren MacArthur added a comment - - edited

            The issue revealed in DM-9109 and tracked down in DM-9110 leads to a very small difference (~0.5% systematic offset between model psf trace radii between HSC and LSST stack). I'm not sure how much its propagated effect could add up to (and I've only looked in detail at this one visit...others could be worse). In order to ensure we are comparing apples with apples, we will need to rerun one of the RC datasets. Given that the newer defaults are meant to be used, I'm not sure if it makes more sense to rerun the RC dataset with the updated HSC stack, or with the "down-dated" LSST stack. Any opinion on this Paul Price?

            Show
            lauren Lauren MacArthur added a comment - - edited The issue revealed in DM-9109 and tracked down in DM-9110 leads to a very small difference (~0.5% systematic offset between model psf trace radii between HSC and LSST stack). I'm not sure how much its propagated effect could add up to (and I've only looked in detail at this one visit...others could be worse). In order to ensure we are comparing apples with apples, we will need to rerun one of the RC datasets. Given that the newer defaults are meant to be used, I'm not sure if it makes more sense to rerun the RC dataset with the updated HSC stack, or with the "down-dated" LSST stack. Any opinion on this Paul Price ?
            Hide
            Permalink
            price Paul Price added a comment -

            It might be useful to have an RC dataset processed with hscPipe 4.0.5 (the latest), but I don't have a strong opinion.

            Show
            price Paul Price added a comment - It might be useful to have an RC dataset processed with hscPipe 4.0.5 (the latest), but I don't have a strong opinion.
            Hide
            Permalink
            lauren Lauren MacArthur added a comment -

            After discussing with John Swinbank, we have indeed deemed this be a useful exercise (see DM-9028).

            Show
            lauren Lauren MacArthur added a comment - After discussing with John Swinbank , we have indeed deemed this be a useful exercise (see DM-9028 ).
            Hide
            Permalink
            lauren Lauren MacArthur added a comment - - edited

            The "bump" in the PSF vs CModel mags at bright magnitudes seen here:

            is now gone with the processing of DM-9554 having adopted the original HSC method for the Brighter-Fatter correction as described in DM-9535, as shown here:

            Summary: we had adopted a slightly different implementation in the application of the Brighter-Fatter correction in the LSST stack (vs. the original one used in the HSC stack). DM-9535 revealed this to be a regression and the RC dataset was rerun with the LSST stack on a branch reverting to the original HSC version. This turned out to be the root cause of the "bump" of this ticket, thus this issue is now resolved.

            Show
            lauren Lauren MacArthur added a comment - - edited The "bump" in the PSF vs CModel mags at bright magnitudes seen here: is now gone with the processing of DM-9554 having adopted the original HSC method for the Brighter-Fatter correction as described in DM-9535 , as shown here: Summary: we had adopted a slightly different implementation in the application of the Brighter-Fatter correction in the LSST stack (vs. the original one used in the HSC stack). DM-9535 revealed this to be a regression and the RC dataset was rerun with the LSST stack on a branch reverting to the original HSC version. This turned out to be the root cause of the "bump" of this ticket, thus this issue is now resolved.
            Hide
            Permalink
            lauren Lauren MacArthur added a comment -

            John Swinbank, is this satisfactory for the conclusion of this investigation?

            Show
            lauren Lauren MacArthur added a comment - John Swinbank , is this satisfactory for the conclusion of this investigation?
            Hide
            Permalink
            swinbank John Swinbank added a comment -

            It certainly looks like it to me.

            Show
            swinbank John Swinbank added a comment - It certainly looks like it to me.

              People

              Assignee:
              lauren Lauren MacArthur
              Reporter:
              swinbank John Swinbank
              Reviewers:
              John Swinbank
              Watchers:
              John Swinbank, Lauren MacArthur, Paul Price
              Votes:
              0 Vote for this issue
              Watchers:
              3 Start watching this issue

                Dates

                Created:
                Updated:
                Resolved:

                  Jenkins

                  No builds found.