LAMOST Continuum Normalization

[annayqho]

Experiment:

• Pseudo-continuum normalize
• Identify continuum pixels using pseudo-continuum normalized spectra
• Apply continuum normalization to the original spectra, not the pseudo-continuum normalized spectra

Note from Melissa: In the tests I did the pixels did not converge, I am not sure if I was doing something wrong but if you find the same thing let me know - the tests I did were to determine the new pixels that are continuum from the spectrum just normalised using the pixels for the preceding found pixels and polynomial/trig fit. I found that a wave got built in and so when I divided it the spectra looked modulated by a sine wave.

[davidwhogg]

rinse and repeat -- the system should converge if repeated. Or if it doesn't, let's look at why.

[annayqho]

Maybe continuum normalization isn't the right goal: the goal should be consistent normalization that preserves our assumptions, not the identification of the true continuum.