So this is the first thing that I see. A plot of Flux vs. Wavelength for a particular order.

Step one is to fit a "continuum" — this essentially means drawing a line that follows the general trend of the background shape (ignoring "lines").

The next step is to fit complicated functions to the "lines" that we see. They are often fairly obvious that "something's there" (notice the 3 blue spikes).

Now that we have a continuum and lines fit for the whole data, we can divide the data by the continuum, to make everything "normalized" to 1.

Conceptually, nothing really that tricky has gone on yet. What you think you see happening, is probably what's happening.

I'm going to put quite a bit more on here, but for now, here's a movie that looks kinda neat.

## How the Convolution Looks

The following image shows how the convolution looks as the iodine lines are shifted.

You are seeing the convolution that was calculated at each shift.

The green Keck lines do not move (of course), while the convolved data can only move vertically because the convolution is calculated at x-values of the Keck data. If you watch the red peak just to the right of 5320.5, you can see how the convolution evolves as an iodine line moves from one side of the x-value where the convolution is calculated, to the other.

The best Chi-Square fit is at a shift of 0.06025.