Does anyone make flats for a transmission grating, e.g, SA100?
I was lead to believe it couldn’t be done (at least well).
I’m interested in the method if it is done. I would think there would be unrepeatable rainbow effects??
I’ve heard it is not recommended to apply a flat to a slitless spectrum. Reason is that if you illuminate the SA100 with a uniform light source, then each pixel in the 2D frame is dispersed into a spectrum. Do this for all pixels in the image and you have a superposition of all these spectra such that any given pixel will have frequency contributions from multiple other pixels. I don’t see how you would separate the frequency and spatial contributions needed to create the flat.
On the other hand, and I’m just thinking out loud on this, if you remove the SA100 from the image train and take a flat as normal, then perhaps this flat can be used to remove the vignetting and dust donuts when applied to a subsequent frame taken withthe SA100 in place. Thoughts?
This question was asked a few days ago in another forum. Robin Leadbeater’s answer can be read here: Re: Easy zero point stacking software for the SA100?
/Johannes
Thanks for the input and links. You confirmed what I thought I knew.
I had thought about removing the filter and doing flats but I have two problems: I’m adding a camera rotator to the mix so I would be doing all kinds of flats, and I don’t like the idea of my fumbling fingers handling the filter all the time.
The R. Leadbeader notes were of interest in that the second paragraph mentions a “professional” way and this sounds like the technique I have been trying. It does involve breaking the image into small frames, finding peaks to mask, creating an estimated flat for each frame and integrating them together. (If anyone is interested, my source is the “photutils” package for python; 2DBackground().) I’d be interested if anyone has worked with this and can give some tuning advice.
Paul
If you are interested, these references describe how flat fielding is done (using a “flat cube”) when reducing spectra from the HST slitless spectrographs.
https://iopscience.iop.org/article/10.1086/596715/meta
It was after reading these that I decided this was likely to be too complex for an amateur setup and made the recommendations for Star Analyser users in the article referenced earlier in the thread.
http://www.threehillsobservatory.co.uk/astro/SA_flat_field_notes.pdf
Cheers
Robin
Thanks again. Sorry I missed typed your name Robin. I have been going through the documents; the one on the aXe software reduction process was interesting. I understand the data cube but I’m not there yet.
So, from everybody’s comments, I should know my camera well (good cal frames, with a flat for just the sensor(?), do a test flat with filter in place. It looks like anything else, e.g., dust motes, optic issues, etc., will be dumped in with the background correction. I am assuming that rotating the camera/filter assembly would randomize the noise.(?) I think I will keep trying out the 2D method.
Everyone is probably doing this already but here’s a follow up which probably be another topic?..
This is bringing to the next part, the annulus: for the traditional method of background correction. I use RSpec a lot. It appears that it combines all of the annuluses together and computes a mean or median (your choice) for the entire segment. If so, this would add noise and error to the results if the background is not flat. I was trying this out but wouldn’t it be better to create a profile for the annulus as well as the SED and use that to correct?
Thanks again,
Paul