meeting 2025 04 02 n57 - JacobPilawa/TriaxSchwarzschild_wiki_5 GitHub Wiki

Context

Just wrapping up some final testing on the kinematics before moving to the minimizations. The fits below all use the same exact settings and differ only in the spectral mask:
- The "base" fit = mask directly from BOX
- The "base" fit with mask edits = mask from BOX + two small features near 8700 and 8750A masked
- The "base" fit with mask edits + chip gap = mask from BOX + two small features near 8700 and 8750A masked + masking the "chip gap" region from ~8750 to 8770 (end of the fit region)
Takeaways:
- The additional mask + additional chip gap mask both HARDLY change the kinematics and seem to only improve the pPXF S/N and the RMS of the fit.
- Furthermore, the chip gap region being masked does very little to the overall goodness-of-fit.

First, here's a quick comparison of the prelim kinematics and the MC kinematics for the three cases above plotted against one another.

	Base Fit vs. Base w/ Addtional Masks	Base Fit vs. Base w/ Addtional Masks + Chip Gap	Base w/ Addtional Masks vs. Base w/ Addtional Masks + Chip Gap
prelim	[images/250402/PRELIM_base_vs_base_with_additional_masks.png]]](/JacobPilawa/TriaxSchwarzschild_wiki_5/wiki/[[images/250402/PRELIM_base_vs_base_with_additional_masks_and_chip_gap.png)	images/250402/PRELIM_base_with_additional_masks_vs_base_with_additional_masks_and_chip_gap.png
MC	[images/250402/base_vs_base_with_additional_masks.png]]](/JacobPilawa/TriaxSchwarzschild_wiki_5/wiki/[[images/250402/base_vs_base_with_additional_masks_and_chip_gap.png)	images/250402/base_with_additional_masks_vs_base_with_additional_masks_and_chip_gap.png

And here's a quick comparison of the "banana" SN and RMS plot, as well and histograms of the two quantities which are a bit easier to make sense of. The distributions of the base w/ additioanl masks and base w/ additional masks + chip gap are virtually indentical, and are all better fit than the case where we only use the BOX mask:

Histograms	SN vs. RMS
[images/250402/histograms_of_rms.png]]](/JacobPilawa/TriaxSchwarzschild_wiki_5/wiki/[[images/250402/banana_plot.png)

Since the three sets above are in great agreement, I think we're fine to move forward with the kinematics associated with the BOX mask + additional masks + chip gap case. Here are what the inputs would look like for this case.
Here's a quick reminder of what these kinematics are actually showing:
- GMOS data = base fit w/ additional maskss + chip gap from above
  - These are the fully corrected kinematics using: Barth stars + bias = 0.2 + starting guess = [0, 200] km/s + new code, etc
- Mitchell data = data from Melanie/Irina/from BOX
  - Ran packinputs.py on the full data, removed the bins associated with the inner 3 fibers since they overlap with GMOS
  - Run symmetrize.sh after removing the inner data points
- Both datasets:
  - Have dummies added for h9-h12 and h7-h12, with the central values of 0 and the errors equal to the highest order even/odd moments from the same dataset
And here are the resulting kinematic diagnostics for these data:

Bin Centers
images/250402/bin_centers_mod_base_CaseZ_mask_edits_and_chip_gap_mod.png

	Symmetrized	Raw Data
Radial Moments	[images/250402/radial_moments_sym_modbase_CaseZ_mask_edits_and_chip_gap_mod.png]]](/JacobPilawa/TriaxSchwarzschild_wiki_5/wiki/[[images/250402/radial_moments_unsym_modbase_CaseZ_mask_edits_and_chip_gap_mod.png)
2d Maps	[images/250402/kinematic_maps_sym_modbase_CaseZ_mask_edits_and_chip_gap_mod.png]]](/JacobPilawa/TriaxSchwarzschild_wiki_5/wiki/[[images/250402/kinematic_maps_unsym_modbase_CaseZ_mask_edits_and_chip_gap_mod.png)
V map w GMOS and Mitchell	[images/250402/n57_sym_mod.png]]](/JacobPilawa/TriaxSchwarzschild_wiki_5/wiki/[[images/250402/n57_unsym_mod.png)