Ticket #8146 (closed: fixed)
[Indirect Diffraction] Error bars defect
| Reported by: | Samuel Jackson | Owned by: | Samuel Jackson |
|---|---|---|---|
| Priority: | major | Milestone: | Release 3.0 |
| Component: | Indirect Inelastic | Keywords: | |
| Cc: | Blocked By: | ||
| Blocking: | Tester: | Gesner Passos |
Description
Reported by Franz: When running with OSIRIS in diffonly mode the d-spacing workspace appears to produce error bars that are way out of proportion. This does not appear to be the case in diffspec mode.
The .raw files that caused the defect are attached to the ticket. Runs 542/543/544 are the van files.
Attachments
Change History
comment:1 Changed 7 years ago by Samuel Jackson
- Status changed from new to inprogress
Refs #8146 changing method of scaling output workspace.
Changeset: bfbb77de28aa45319de8789a6d0c6fae60f58b27
comment:2 Changed 7 years ago by Samuel Jackson
Refs #8146 Updated reference result for OSIRIS diffonly.
Changeset: 5e0b07716d42f003631e34567b4782e25d3a540c
comment:3 Changed 7 years ago by Samuel Jackson
- Status changed from inprogress to verify
- Resolution set to fixed
Checked this with both Spencer and Franz and they agree this is seems to be the correct solution.
To Tester
Check that the running OSIRISDiffractionReduction works with the files provided:
- osiris00102605.raw is the sample
- The other raw files are vanadium.
- The calibration file can be found in the system tests data folder.
I've provided the testing script I used to make things a little easier.
Be sure to check that plotting the spectrum with errors bars produces a result where the error bars huge. I would suggest attempting to reproduce the error on a different branch first so you can see the difference.
This branch will require merging in the main repo and in the system tests
comment:5 Changed 7 years ago by Gesner Passos
- Status changed from verify to verifying
- Tester set to Gesner Passos
comment:6 Changed 7 years ago by Gesner Passos
- Status changed from verifying to closed
Merge remote-tracking branch 'origin/bugfix/8146_diffrac_error_bars'
Full changeset: a6cffa3f0e7f4ce32d2950033726314ea860bf2d
comment:7 Changed 7 years ago by Gesner Passos
The error now is reasonable. I would just put one comment on why (correctly!) it was necessary to skip the division algorithm. Besides, probably the Stitch1D could be used instead of MergeRuns, but unfortunatelly, it requires to pass on the RebinParams... Why Stitch1D can not calculate the Rebin as MergeRuns? Would it be possible?
Well, the important thing is that it was sorted out!
comment:8 Changed 7 years ago by Martyn Gigg
Merge remote-tracking branch 'origin/bugfix/8146_diffrac_error_bars'
Full changeset: 851ff85fd830a8f4474a80aad14ea1ed704894bc
