Forum der Vereinigung der Sternfreunde
https://forum.vdsastro.de/

A robust method to measure centroids of spectral lines
https://forum.vdsastro.de/viewtopic.php?t=5420
Seite 1 von 1

Autor:  Christian Netzel [ 28. September 2018, 23:08:04 PM ]
Betreff des Beitrags:  A robust method to measure centroids of spectral lines

Hallo,
über Centroide von Spektrallinien wurde hier bereits diskutiert. Daher möchte auf folgende Arbeit hinweisen:
Teague Foreman-Mackey,
A robust method to measure centroids of spectral lines
https://arxiv.org/pdf/1809.10295.pdf
Viele Grüße
Christian

Autor:  Robin Leadbeater [ 30. September 2018, 14:47:53 PM ]
Betreff des Beitrags:  Re: A robust method to measure centroids of spectral lines

Hello Christian,

I find this very counter-intuitive. They are suggesting that they can get a better estimate of the centroid by ignoring 80% of the data (~12/15 data points in their example line profile), independent of the line shape or SNR. I find this difficult to believe. For example in an oversampled line with a broad top, a small amount of noise could change the location of the maximum pixel giving a large error

Cheers
Robin

Autor:  Christian Netzel [ 01. Oktober 2018, 12:01:46 PM ]
Betreff des Beitrags:  Re: A robust method to measure centroids of spectral lines

Hello Robin,
my intuition says that you are right.
Cheers
Christian

Autor:  Thomas Eversberg [ 01. Oktober 2018, 15:57:31 PM ]
Betreff des Beitrags:  Re: A robust method to measure centroids of spectral lines

Robin is right! This paper is clearly not addressed to velocity measurements in stellar environments (especially lines from expanding winds). They rather apply it to radio observations with sharp lines (their model example uses Gaussians).
Cheers, Thomas

Seite 1 von 1 Alle Zeiten sind UTC+02:00
Powered by phpBB® Forum Software © phpBB Limited
https://www.phpbb.com/