What to do about these weird gradients?

[jiawen]

I took a night worth of shots of M8 and M20 a couple weeks ago. After a lot of processing in StarTools and elsewhere, I still have some pretty nasty gradients across the image. There are reddish bars toward top and bottom, then wider bluish bars, then an area that seems not to have any gradients toward the middle. I assume that the gradients are caused by ambient temperature, but there isn't really much I can do to change that. Also, I assumed that the darks would've eliminated that kind of thing, but no such luck. Is it that I'm overprocessing these images? Is there a function in ST that would eliminate these gradients?

I'm attaching two images: a copy of the finished result, and a version of that despeckled to eliminate the stars and highlight the gradients.

Acquisition details:

Camera: Olympus micro 4/3 with 180mm Nikkor legacy prime lens
Mount: Star Adventurer
Lights: 323x20sec, winnowed to about 280 subs
Darks: 50x20sec
Bias: 220 frames
No flats (no practical way to shoot them, much of the time)
Stacking in DSS; the usual processing in ST (crop, bin, develop, wipe, HDR, sharp, denoise), post-processing in RawTherapee and GIMP -- I don't have a lot of money to throw at this, so hopefully I can solve this problem within ST.

[almcl]

By no means an expert, but if you can't take flats (and they would really help if you could) then you may need to crop a bit more, how much are you cropping at the moment?

[jiawen]

Yep, thank you for correcting my mistake, I meant flats. And yes, I know how useful they are, but they just aren't practical for me so far.

Native camera resolution is 4080x3040; I've been binning it at 71%, then cropping out the stacking artifacts, and then cropping again later on. The full-res image I made those low-res shots from is 2590x2059. Not sure if cropping more would help.

[almcl]

If you are doing all the cropping before the wipe, then, no, it probably won't make a difference. But if there any artefacts (not just stacking ones) left in when you wipe, it can introduce weird colours.

[jiawen]

"if there any artefacts (not just stacking ones) left in when you wipe"

What other kinds should I watch out for? I'm pretty sure the cropping (which, yes, I do before wiping) would've gotten rid of all the stacking artifacts.

[admin]

Hi,

We could maybe have a look at the data if you could share it?

These are not typical gradients in that they undulate very fast in distinctly different color channels as well.

It its most likely an issue with a filter or lens. You should definitely be acquiring flats - they are really not optional unfortunately. They are, after dithering, the cheapest (e.g. no-cost) way of vastly improving your efforts. At the very least they should ameliorate your problem (if the gradients seem to happen in the same configuration/positions across multiple datasets).

[jiawen]

Thanks for responding!

Hi,

We could maybe have a look at the data if you could share it?

I don't have a good way to share stuff, unfortunately. The DSS TIFF is 148 MB.(I don't suppose I can upload it here?)

These are not typical gradients in that they undulate very fast in distinctly different color channels as well.

Sorry to hear that.

I did try inverting the despeckled layer, with a star/nebula mask, and then applying it as a "color" layer in GIMP. That seemed to reduce a lot of the weird colors, but had the side effect of reducing overall saturation (of course).

It its most likely an issue with a filter or lens.

My camera and lenses have no filters, that I know of. I think I've gotten a little of the same gradient problems regardless of lens, too, if I'm remembering right.

My current best guess is that it's related to ambient temperature. I live in a pretty hot place; when I'm imaging, it's recently been above 80°F here while I've been imaging, sometimes well above that. And I think the small sensor size exacerbates temperature problems.

You should definitely be acquiring flats - they are really not optional unfortunately. They are, after dithering, the cheapest (e.g. no-cost) way of vastly improving your efforts. At the very least they should ameliorate your problem (if the gradients seem to happen in the same configuration/positions across multiple datasets).

I did manage to get flats one night, about a month ago, when I was actually able to stay up until dawn. But the rest of the time, I don't have a practical way to get flats. Cheap once you've got the equipment, yes, but first I'd need to get something to produce flat white light. (It's like when people say a barn door mount is cheap. Sure is, if you already have access to machine tools! Otherwise, it's pretty expensive.)

[Starry Eyes]

For your flat problem, Have you tried sky flats? Easy to do, Cheap, just use some blue sky with the OTA of your choice during the morning after dawn. (Helps with the temperature issues & make sure that the rig is out of direct sunlight). Make a histogram about in the middle, and shoot ~ 50 of them. Stack and you'll have a Master Flat. That's all I use with my DLSR.

PJM
http://www.astrobin.com/users/Starry-Eyes/