search

davidwhogg / HoneyComb

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

Good news and bad news #20

Open RuthAngus opened 9 years ago

RuthAngus commented 9 years ago

The bad news is that Dan Huber didn't have any new insight to offer re finding the modes in the poster child. "They should just pop right out of the Fourier transform". I am still working on this...

The good news is that I have implemented something like what he uses to measure delta nu, described in this paper: http://arxiv.org/abs/0910.2764

You compute ACFs for lots of sections of the Fourier transform and add them up (horrendous, but gets the job done!)

Top panel shows the FFT of KIC 6442183. Middle panel shows ACFs calculated for different sections of the FFT. Darker = more correlated. Bottom panel shows the middle panel collapsed along the horizontal axis.

6442183_dnu

davidwhogg commented 9 years ago

Okay fine, I am surprised; he couldn't explain that apokasc figure? It is just magic that we did the SAME THING WITH THE SAME DATA and got totally different results? I am pretty upset about that.

With this poster child now -- let's focus on this one -- can you show me the comparison of the FFT, the L-S, and the superpgram?

RuthAngus commented 9 years ago

Here's the fft, lombscargle and superpgram, with a zoomed-in version. They look super-similar, but then these signals are pretty giant! 6442183fft_ls_spg test

davidwhogg commented 9 years ago

Okay this looks great! I am happy now.

Let's now play some games where we average down the data to longer and longer exposure times and see what happens. Also, both constant exposure time (same for every data point) and also random exposure times.

RuthAngus commented 9 years ago

Ok, I'm on it!