Star color using ES CLS filter [Pixinsight]

[jjosefsen]

Hello all,

I am trying to process data captured on friday during almost full moon, and I'm having a hell of a time of it..

Due to the full moon doing it's best sun impression my target was an open cluster (NGC 225 in Cassiopea), thinking it would be doable.

I was using my 2" Explore Scientific CLS filter, my unmodded Nikon D7100 and captured 92 usable 45 second subs at ISO 400.

Calibrated with bias and flat frames and stacked with linear fit clipping as the pixel rejection algorith.

The mission for the night was not getting particularly good data, but getting decent tracking and round stars, something I have had lots of trouble with!

 

Now that I am trying to process it, I find im dealing with some very nasty gradients / casts and I can't seem to get any good colors.

An initial strech reveals a strong blue color cast/gradient that I can deal with in a number of ways.

• Linear fit
• DBE
• ABE

followed by the usual:

• SCNR (green)
• Color Calibration

No matter what I do I end up with the whole image being either really green or really red and not much else for color in any of the stars.

I also think i burned out all the major stars by over saturation..

I know the CLS filter is going to wreak havoc on colors because it completely blocks parts of the spectrum, so maybe the CLS filter is just the wrong tool for the job?

If anyone would like to show me what they can get from this data, and how (so preferably with PI) I would be really grateful. At the moment im thinking maybe the data is just really bad and not useable?

Linked is the final integrated xisf (only xopped slightly):

https://www.dropbox.com/s/nmxds87toc7vz51/integration.xisf?dl=0

[steppenwolf]

The downloaded image is showing as all over blue on my system, not green or red? However, an AutomaticBackgroundExtraction (ABE).

There is hardly any colour in the stars as you say but importantly, the image is not in focus which isn't helping you with the post processing

Can you upload the best of your subs (to DropBox) - just a single one without any processing having been carried out?

[jjosefsen]

Yeah the red and green comes out when I try to process it, the main problem being that it applies to pretty much the whole image.

But as you say there may very well not be much colour data there to begin with..

Link to best of the subs without anything done to it at all:

https://www.dropbox.com/s/zutqq8ht47jyadf/NGC225_LIGHT_45s_400iso_2_CLS_20171201-22h14m28s491ms.NEF?dl=0

Thanks for taking a peek at it Steve, if the data is just bad, dont waste any time on it, then  atleast I know this needs to improve significantly.

 

Is it badly out of focus? This seemed to be the best I could get on that night using the focus tools in Backyard Nikon.

It was a pain, because the moon washed out the bright stars, so couldn't focus with live view, had to do it on short subs.

[steppenwolf]

I have de-Bayered your image and apart from the fact that it is out of focus, there is no strange colour cast. Are you sure that you used the correct de-Bayer matrix at the start of the process?

[jjosefsen]

Hmm 95% sure.. I will try to replicate your result.

Thanks

[steppenwolf]

OK, I am getting the same result as you when using PI but a satisfactory result when using MaxIm DL so I wonder if this is a PI issue?

I'm going to try another experiment .........

[steppenwolf]

Hmmm, PI isn't currently my field of expertise but I wonder if you have simply stretched the image too far as PI IS correctly de-Bayering the image. Auto ScreenTransferFunction is also stretching the image too far perhaps?

[jjosefsen]

This is pretty much what I get everytime with PI, I honestly thought that was "normal" as I can usually get something decent out of it after dbe or linear fit.

[jjosefsen]

15 minutes ago, steppenwolf said:

Hmmm, PI isn't currently my field of expertise but I wonder if you have simply stretched the image too far as PI IS correctly de-Bayering the image. Auto ScreenTransferFunction is also stretching the image too far perhaps?

That might be it, I'm going to go back and have a look at some data I shot a few weeks ago without the CLS filter and the full moon. Do some comparison on the colors there..

Thank you for looking at it.

[steppenwolf]

Based on my analysis in MaxIm DL the data is thin but there and the only issue I can see is that your image is over-stretched.

Have a go at processing it again but this time, start from scratch and don't use Auto STF, just work gently on the data without STF enhancement and see where that gets you?

I'd also recommend taking the CLS filter out of the equation while you conduct further testing to obviate introducing a mechanical colour cast..

[steppenwolf]

We doubled!

[jjosefsen]

Honestly I'm not even sure I need the CLS filter, was just looking for an edge to get a bit more out of the data.

I wish I had a few years of experience at this, often I have no idea if what I'm capturing is ok quality wise, until I start processing it, and even then I might screw the pooch there too.. ?

Thanks for taking the time.

[Filroden]

Try the new PhotometricColourCorrection process. Apply a default ABE then apply the PCC. You’ll need your focal length and pixel size for it to plate solve but it will try and balance star colours to thei4 catalogue BV.

[steppenwolf]

Yes, this process is quite effective providing it can match sufficient catalogue stars.