Back to forums

Attacking weird background noise

Processing Deep Sky M81 NGC3031 Galaxy
12 replies•614 views
Dimitris Kavallieratos avatar
Dimitris Kavallieratos avatar
Hello everyone,

Since galaxy season is on, i started imaging M81/82 with my ASI533MC camera. After integrating my frames(PI or DSS),a weird red noise have been swamping my background and even though I tried everything (multiple DBEs/SCNR/RangeSelection to apply masks/ColorMask), no effiecient way to do this without weirding out the image has been found or hurting my galaxies.

Everytime I go broadband from my backyard this red noise is profound and only L-Extreme has been able to manage this, also this noise is only found when I use my new astro camera only.I did not use my L-pro filter because this also can not contain the noise and I lose some color on my broadband captures,so I prefer to go without filter for these.

So your advice is very valuable,since it seems that I have to live with it, maybe something can be done is pre-proccesing ? No idea...

Here are some .fits exports of mine, my total integration time is about 8 hrs (30secs exposures all calibration frames where used):

Initial integration after Linear fitting only
DBE Clone
Starless

Thanks in advance!

Edit: Also I dithered every 5 frames.
Helpful Engaging
andrea tasselli avatar
andrea tasselli avatar
HI Dimitris,

You should always stick to the L-PRO as this is your only way to contain the LP, I cannot stress this strongly enough. I processed the original image and removed the LP gradients (mostly, the far right corner still shows some of it), see below. Just a judicious use of DBE would do the trick. I can repost the image to a Google drive if is of any help.
Supportive
Dimitris Kavallieratos avatar
Dimitris Kavallieratos avatar
andrea tasselli:
HI Dimitris,

You should always stick to the L-PRO as this is your only way to contain the LP, I cannot stress this strongly enough. I processed the original image and removed the LP gradients (mostly, the far right corner still shows some of it), see below. Just a judicious use of DBE would do the trick. I can repost the image to a Google drive if is of any help.


Thank you Andrea, so DBE repetion/correct use is the way, thank you for the feedback. Is the trick to play with the tolerance or just placing the samples in the right places? I mainly go for tolerance 2/shadows relaxation 6 and then divison&subtract after selecting the samples. I did it more than 3 times but the red background remained...

As for the L-pro ,even with it this red noise was still present, but for sure you know way better than me so I will follow your advice!
andrea tasselli avatar
andrea tasselli avatar
Hi Dimitris,

The image above was made by using the default DBE setting with 55 samples per row, 0.15 sample weight, tolerance 1.5. I removed all samples nearby bright stars and ALL the galaxies including halos. Then 3/4 iterations of DBE again, with 0.25/0.3 sample weight, 55 samples x row standard tolerance until you get rid of most of the blue gradient in the right corner. Repeat until pleased with the background flatness. All corrections are to be made by subtraction.

Please note that you have a problem with the stars, the black-eyed bean effect.
Helpful
Dimitris Kavallieratos avatar
Dimitris Kavallieratos avatar
andrea tasselli:
Please note that you have a problem with the stars, the black-eyed bean effect.


Tried the axial method described by Adam Block with fixed values for R/G/B to flatten the field and did not work or it did work but trying subtraction drew the red again on top, will definitely try your method and let you know! Thanks a lot!

By black-eyed bean effect, what do you mean?
andrea tasselli avatar
andrea tasselli avatar
The upper right side of the brighter stars has a dark spot in it, see below:



This is usally due to something you've done when calibrating the raw light frames.
Rafał Szwejkowski avatar
Rafał Szwejkowski avatar
Despite dithering you have overwhelming walking noise, my guess is that the dithering was insufficient in distance, you probably want at least 20 pixels per move.  Every 5 30s frames is frequent enough.
Dimitris Kavallieratos avatar
Dimitris Kavallieratos avatar
Despite dithering you have overwhelming walking noise, my guess is that the dithering was insufficient in distance, you probably want at least 20 pixels per move.  Every 5 30s frames is frequent enough.

I did dithered with the default NINA setting of 5 pixels RA only, as a have a star tracker...Pff every time something different, aint it?

Thank you!
Dimitris Kavallieratos avatar
Dimitris Kavallieratos avatar
@andrea tasselli Andrea your info was very valuable ,I did implement your method and I managed to force denoise/extract the thing and salvage somehow my data.Though that I threw out 2 days of clear skies (again)...

I posted the image to my AB page ,take a look if you can and give me your feedback if it is not bothersome.

Thanks again for your time
Rafał Szwejkowski avatar
Rafał Szwejkowski avatar
Despite dithering you have overwhelming walking noise, my guess is that the dithering was insufficient in distance, you probably want at least 20 pixels per move.  Every 5 30s frames is frequent enough.

I did dithered with the default NINA setting of 5 pixels RA only, as a have a star tracker...Pff every time something different, aint it?

Thank you!

Using star tracker you only dither in RA most likely, what you need to do is to manually dither in DEC once in a while, just use the up and down knob between every 10-15 mins, it should help eliminate the issue.  One thing that exacerbates the walking noise is the large number of exposures at 30s per.
Concise
andrea tasselli avatar
andrea tasselli avatar
I posted the image to my AB page ,take a look if you can and give me your feedback if it is not bothersome.


Hi Dimitris,

Had a look. There are a number of things you might want to look at:

1. The background is both too bright and way too green. Use BackgroundNeutralization to achieve a more neutral colour.
2. You haven't the data depth to push contrast too high, so is probably best you don't.
3. The galaxies are way too reddish. Their "natural" colour is off-white with more than a hint of green. Are you using PhotometricColourCalibration to colour calibrate you stacked up image? I'd use ArcsinhStretch to bring out the colours while keeping the background at bay and then selectively use HistogramTransformation to bring in a bit more depth in the mid-tones. Finally I's use CurvesTransformation to change the green cast to a more blueish one
Helpful Insightful Respectful
Andres Salaverria avatar
Andres Salaverria avatar
Hi Dimitris,

I image from a Bortle 9 sky, and have tried everything I could think of to tackle light pollution noise. I have used an Idas LP filter, and just started using the L-Pro to image galaxies. It appears to do a better job (at least for me) compared with the Idas. you can see some of my images in my AB page. In fact, I imaged M81/M82 with a SpaceCat 51 (the gray version of the RedCat), and I was not impressed at all! In the last couple of weeks I imaged M81 and M82 separately using a Celestron 8 Edge at F/7. My integrations are short, but I am not terribly displeased with the results.
Helpful
Dimitris Kavallieratos avatar
Dimitris Kavallieratos avatar
andrea tasselli:
. The background is both too bright and way too green. Use BackgroundNeutralization to achieve a more neutral colour.
2. You haven't the data depth to push contrast too high, so is probably best you don't.
3. The galaxies are way too reddish. Their "natural" colour is off-white with more than a hint of green. Are you using PhotometricColourCalibration to colour calibrate you stacked up image? I'd use ArcsinhStretch to bring out the colours while keeping the background at bay and then selectively use HistogramTransformation to bring in a bit more depth in the mid-tones. Finally I's use CurvesTransformation to change the green cast to a more blueish one

I did some SCNR for green and "upped" the blue color with curves and masking the background, to take care some of the issues. I had already done some BE and PCC so this color is the most neutral I can get with these data unfortunately...
Andres Salaverria:
My integrations are short, but I am not terribly displeased with the results.

I would be terribly pleased with your results! When I upgrade to a proper mount and OTA for galaxies, I am hoping to have your results...
Using star tracker you only dither in RA most likely, what you need to do is to manually dither in DEC once in a while

I will have to live with RA only for now unfortunately. Will try to reduce the number of frames with my LP filter on, at least!