I'm copying Dennis Conti, the code author of the NEBcheck tool, on this response in case he has time to add an option to include any saved detrended light curves in the NEBcheck analysis. Generally though, for SG1 reviews, I'd use the undetrended light curves
and spot any NEBs by eye. The problem with detrended light curves is that a user can detrend a light and often warp it into very non-representative shapes. Without knowledge of what processing has been done to a light curve, I'd always defer to the raw NEBcheck
plot and perform my own eyeball detrending. The only case that I think might be useful to display a detrended result is to the case of very close neighbors with strong cross contamination between the two apertures. In those cases, Width_T1+airmass detrending
can help mitigate the affects of the blending and show an NEB, or an event on target. For those cases, I'd rather see an AIJ plot which also shows the airmass and width curves to aid verification of the results by eye.
Dennis, if you have time, it might be useful to have an option to plot any saved detrended plots. The approach would be to include not just plots of rel_flux_T/Cxx data columns, but also plots of any additional same-names columns that have suffixes _dn, _d,
_dfn, _fn directly following the raw plots.
On 11/11/2019 9:45 AM, james.lowenthal [via AstroImageJ] wrote:
We're part of the TESS followup program and are using AIJ to analyze light curves of TESS exoplanet candidates. We're practicing on TESS targets that have been retired from the candidate list because a nearby eclipsing binary (NEB) was identified as responsible
for a false TESS detection.
We remove noisy stars and detrend as usual, and get light curves flat to RMS~2 ppt. We clearly see the NEB signature in the correct star's (T2 in our case) light curve. We save new columns in the data table with normalized de-trended relative fluxes, e.g.
rel_flux_T2_dn, and errors.
But when we then use the nifty AIJ "Create NEB search reports and plots..." tool, it seems to grab only the NON-detrended, noisy light curve data. The resulting plots are not flat at all, and fail to identify the NEB that we can see pretty clearly in the detrended
normalized curves in Multi-plot.
Of course we don't need the NEB tool if we can see the NEB by eye. But we certainly want the NEB tool to use the best, flattest, lowest-noise, detrended data when we do use it. But how?
I'm probably missing a simple step in there somewhere -- thanks for any help identifying it!
To start a new topic under How to accomplish specific tasks in AstroImageJ, email
To unsubscribe from AstroImageJ,