deletions | additions       diff --git a/Stellar atmosphere model.tex b/Stellar atmosphere model.tex  index 1497c90..619fcfa 100644  --- a/Stellar atmosphere model.tex  +++ b/Stellar atmosphere model.tex 
...
To accomplish this, we use the FDBinary tool \citep{ili04} on the spectral window 5402--6750 \AA. Following the approach in \citet{bec14}, we break the window into 25 pieces that each span about 10 \AA. FDBinary does not require a template spectrum, and instead uses the orbital parameters of a binary system to separate spectra in Fourier space. We tested FDBinary's capabilities by creating a set of fake double-lined spectra from a weighted sum of two identical spectra of Arcturus. When the orbital solution and flux ratio is correctly specified, the program accurately returns a pair of single-lined spectra that are indistinguishable from the original.  FDBinary requires six parameters that define the shape of the radial velocity curve:  orbital period, time of periastron passage (zeropoint), eccentricity, longitude of periastron, and amplitudes of each star's radial velocity curve. We set these to 171.277 days, 321.189576 days\footnote{Units of BJD--2454833}, 0.36, 18 deg, 33.7 km s$^{-1}$, and 33.1 km s$^{-1}$, respectively. While FDBinary does include a rudimentary optimization algorithm for any subset of these parameters, it consistently ``preferred'' solutions that are wildly incorrect when compared to the observed radial velocities.  These orbital parameters parameters, which  come from a preliminary set of photodynamical models, which were ultimately held fixed.  %They  are later refined with stellar atmosphere models (see Section \ref{model}).TO ADD: WHY WE DIDN'T USE FDBINARY'S FITTING ABILITY... IT SUCKS  FDBinary also requires the flux ratio of the two stars. Because KIC 9246715 contains two very similar red giants---the oscillating giant is slightly larger and cooler than its companion---we set the flux ratio to 1. To learn how this assumption affects the result, we further tested FDBinary by specifying an incorrect flux ratio for the set of fake double-lined spectra based on Arcturus. All of the correct absorption features were recovered, but one of the two output spectra had features that were systematically too deep while the other's features were systematically too shallow. Since this appears to affect all absorption lines in the same way, we are confident that choosing an incorrect flux ratio for two stars of similar brightness will not affect atmospheric modeling.