The Fun of Processing a Stellar Spectrum—the Hard Way (Abstract)

Volume 47 number 1 (2019)

Stanley A. Gorodenski
9440 E. Newtown Avenue, Dewey, AZ 86327; stanlep@commspeed.net

Abstract

(Abstract only) Freeware exists for processing spectra: vspec, iris, isis, audela, bass, and midas, to name some. One feature in common with all is that they can, to some extent, be viewed as black boxes, and they are limited to doing certain very specific functions. Not only may some users feel it would be too daunting to attempt to write computer programs to do some of the same things these packages do, many of the packages are limited to the extent a spectrum can be explored in a statistical and graphics sense. There are a huge number of gifted computer people in astronomy. This talk is directed toward those, like the author, who is not so gifted but found it very interesting and exciting to be able to write programs to explore a spectrum. Although there are many things one can do, such as computing equivalent width, computing radial velocity, and estimating a continuum, this talk will only focus on two things. It will show the fun of developing a Gaussian curve and using it to identify large deviations from an ideal Gaussian distribution at the pixel column level of a spectrum. It discusses some of the difficulties of doing a dark sky subtraction and it describes an exploratory method for doing one that, although it may or may not be better than the method these other packages use and actually is very computer intensive, demonstrates what can be done if you can computer program and have a good statistical package with good programming capabilities.