Strange artefact in one processed sub

Post Reply
Kevin Gurney
Posts: 15
Joined: Tue Nov 27, 2018 5:42 pm
Location: Sheffield, UK
Contact:

Strange artefact in one processed sub

Post by Kevin Gurney »

I came to do the routine processing of a reference star for an Alpy target and saw something strange in the final profile.
Thus, there were several well-defined, and equally-spaced 'notches' therein.

I finally tracked the source of this artefact to the processing of a singe sub-exposure (1 out of 5). This led to an artefact in the intermediate '@' file - see attached.
Note the 'notch' just to left of H_beta which gets repeated at regular intervals as you go towads the red end of the spectrum.
This is not present in the original data.
No other subs suffered from this.
Indeed, as a 'control experiment', I reworked the whole reduction on an alternative platform - Demetra. No problems there.

One possibility I considered was that, this was the brightest sub, and internal processing in ISIS may be sensitive to some kind of 'saturation' which is not otherwise apparent.
To test this hypothesis, I scaled the sub using the 'Multiply image by a constant' tool, but this made no change.
As a clue to what might be going on, note the weird uniform grey sector at the top left...
I also notice all files of the '@' form have negative numbers in them ....any cause for concern?

Kevin
Attachments
ISIS_ampersand_file_all.jpg
Kevin Gurney
Posts: 15
Joined: Tue Nov 27, 2018 5:42 pm
Location: Sheffield, UK
Contact:

Re: Strange artefact in one processed sub

Post by Kevin Gurney »

I think I may have cracked it...
Having tried many things, I finally came to try turning the cosmic ray filter off... This eradicated the artefact.
I can see that very bright portions of a spectrum can look like a cosmic ray strike. I think I am just going to stop using this filter from now on...

I was pointed in this direction by this posting:
http://www.spectro-aras.com/forum/viewt ... f=8&t=2004

Kevin
Robin Leadbeater
Posts: 1926
Joined: Mon Sep 26, 2011 4:41 pm
Contact:

Re: Strange artefact in one processed sub

Post by Robin Leadbeater »

Hi Kevin,

Sorry I missed your post

ISIS produces intermediate (@map) files showing the pixels that have been healed by the cosmic ray filter in each image. I particularly check them on bright targets. Problems typically show up as a number of pixels in a line down the centre of the spectrum. You can then tweak the threshold to avoid this while still healing cosmic rays

Cheers
Robin
LHIRES III #29 ATIK314 ALPY 600/200 ATIK428 Star Analyser 100/200 C11 EQ6
http://www.threehillsobservatory.co.uk
Robin Leadbeater
Posts: 1926
Joined: Mon Sep 26, 2011 4:41 pm
Contact:

Re: Strange artefact in one processed sub

Post by Robin Leadbeater »

Kevin Gurney wrote: I also notice all files of the '@' form have negative numbers in them ....any cause for concern?

Kevin
Negative numbers in the background of dark/sky subtracted images are fine. They should be there due to the random variability. Any program that clips them will bias the zero level and will not give the correct result

Cheers
Robin
LHIRES III #29 ATIK314 ALPY 600/200 ATIK428 Star Analyser 100/200 C11 EQ6
http://www.threehillsobservatory.co.uk
Kevin Gurney
Posts: 15
Joined: Tue Nov 27, 2018 5:42 pm
Location: Sheffield, UK
Contact:

Re: Strange artefact in one processed sub

Post by Kevin Gurney »

Many thanks Robin

I'm glad I know what the @map files are for now! I'll experiment a bit and see...

Kevin
Kevin Gurney
Posts: 15
Joined: Tue Nov 27, 2018 5:42 pm
Location: Sheffield, UK
Contact:

Re: Strange artefact in one processed sub

Post by Kevin Gurney »

So I have determined that, for 2x2 binned images, there is no setting of the cosmic ray threshold that allows eradication of cosmic rays without compromising the spectrum itslelf. I suspect that (for a well-focussed star with no DEC drift) the vertical intensity gradient across pixels is comparable with that for detecting cosmic ray events.

Things seem to work reasonably well, however, for non-binned images ('bin 1x1'). I am using a 460EX (pixels size 4.54mu)

Kevin
Post Reply