search

davidwhogg / HoneyComb

asteroseismology on the cheap
MIT License
2 stars 0 forks source link

rebin short-cadence poster-child data and re-run #22

Open davidwhogg opened 9 years ago

davidwhogg commented 9 years ago

Can we take the short-cadence data on the poster child, and:

(meaning: average adjacent pairs of data points)

and then repeat -- ie, rebin again and do it again, to see how the signals degrade as we rebin?

RuthAngus commented 9 years ago

Sure - it takes a while to run but I'll update you when I have plots.

RuthAngus commented 9 years ago

Here are some rebinning plots for superpgram (top), fft (middle) and Lomb-Scargle (bottom). I am a little confused about why the superpgram and fft versions look so similar, but the LS version looks different. I would have expected LS and spg to look more same-y. Does this make sense to you?

spg: binning_comparison_spg fft: binning_comparison_fft LS: binning_comparison_ls

And I have another question... in the above spg plot I'm not taking the binning into account when I compute the footprint. I leave the delta t in the footprint at the short cadence interval of ~1 min. When I do take the changing time interval into account (i.e. doubling it for 2 points-per-bin, etc) I get the following plot. I am confused though because I would have expected this to look better than the above example. Am I misunderstanding something? Does this make sense to you, or is it more likely that I've got a bug somewhere?

binning_comparison

davidwhogg commented 9 years ago

No, I don't understand this. Let's try a simpler thing out of the gate: Let's make Fake data that exactly conforms to the SPG model... New issue coming while this sits on hold