Blue blotchy background cannot correct out

Dieter Lubbe

I have a blue blotchy background with M81 I can't correct out. ZWO 2600MC with Antlia RGB Ultra filter. EDGE 8 at 2032mm without reducer. Image attached is after DBE before anything else, but oversaturated with curves to enhance mottling effect. Does not look like gradient. 3 nights total but same issue with individual nights only. This image from one night set 55 x 5min subs. To try and isolate the issue, I ran WBPP with and without flats, no difference in blue mottling. As expected, no change with ABE vs DBE vs Graxpert, with or without background neutralization. SPCC, PCC or CC does not remove it, as expected. Re-ran WBPP with or without local normalization, no difference. I'm thinking it may be the filter as new to my setup and have never had issue with this on same 2600MC/EDGE combo with more than a dozen targets. If filter though would flats not correct out the error? Any insight appreciated. Thanks. Dieter.

BlockHead

Yeah... flats should correct it out. Still... did you check for zeros in your calibrated data.. are you certain there isn't an over subtraction?

Otherwise, you can diagnose with Night Sky Flats (something else easy to try).

-the Blockhead

Dieter Lubbe

No zeros, lowest value is 0.0027. I also ran without any flats to see if somehow related to flats and same color issue. I was even thinking about just desaturating the background starless to get something out of the data but the galaxy is also affected by same weird blotchy color.

BlockHead

Wait... You did not convince me. Not the master.. but in calibrated frames. No zeros?

-the Blockhead

Dieter Lubbe

Nope. I have included a link to dropbox folder of masters and a calibrated light frame.

A little nervous now, I have watched most of your videos and starting to feel like a failing student!

https://www.dropbox.com/scl/fo/etphw0fyywtnv9k9fs0dn/h?rlkey=u2dnzxhx6g9sjt7r9izgcdyvp&dl=0

BlockHead

Dieter,

My overall opinion is that is a data/sensor "issue". I put issue in quotes because it is such a low level effect that I wonder if it is just the nature of things. You simply do not have enough useful signal in the image.

That being said, I looked at the image and ran it through both DBE and GraXpert. Both basically come up with the same answer (DBE might be a little better). The blue channel is simply blotchy...and the values are almost single digit differences. You are almost looking at a binarized image to see the effect. This is why I say there just isn't enough data.

There are three ideas to consider:

1. We are seeing low level effects of the blue filtered pixels and the manufacture of the substrate.

2. Drizzle effects are correlated and giving us the pattern.

3. There really is a flat issue. 

#2 would be a tidy explanation- and simple to test. You did not provide the undrizzled result. You did not mention it. I refuse to assume anything. Drizzling is a manipulation of the data and there are correlations between pixels. Please verify that undrizzled result shows the issue as well. This is output by WBPP also.

If the pattern is there... then I think it truly is a sensor effect. You would have to demonstrate it really did not exist in other data...and wasn't simply there at a low enough level you just did not notice it.

I do not understand why your sky is so bright in blue to begin with. Your flat seems saturated in Blue... is your flat field a good one? Sky Flats might give you a better result. Are you using a flat field that has a lack of red?

-the Blockhead

Dieter Lubbe

Adam, thank you for the thoughts.

1. The non-drizzled master has same effect.

2. Regarding data quality, there was almost 12 hours of integrated data from 3 nights

3. Your third point is interesting. I posted same question on Pixinsight forum and a few similar suggestions that it is my flats, which I only briefly considered. It's a fairly new panel purchased for size to cover the 8" of the EDGE 8 and some convincing data from some on the forum that it is indeed unbalanced light with blue channel blown out and actually saturated. I am going to redo flats as my optical train has not been changed once I test for a more balanced light source.

What a humbling hobby!

Thanks again for all your help with this and your great site.

Dieter

BlockHead

It isn't just the color of the flat... indeed your flat appears to be in part saturated. You need to find a light intensity that will allow you to take 1-2 second exposures with only 1/2 saturation. You will probably want to stack quite a few since your panel apparently does not emit as much red light (common).

-the Blockhead

P.S. I might ask members- if they wish for multiple opinions on a matter to first ask questions on the official forum and then come to me if the answer is not good enough or no answer. I know it is easy enough for people to post across many forums and find a consensus... but I am a lone responder (basically) on this forum and it is tough to give quality answers (or dive into mysteries from scratch) each time. 

Of course any questions on MY content should be here. lol

Dieter Lubbe

So, interesting follow-up findings:

I made new flats with a Photoshop calibrated white background on iPad Pro using initially (for 1 and 2 below) the default ASIAir recommended automatic flat exposure setting:

1. with Antlia RGB filter at ASIAir recommended flat exposure: same extremely widely split channels after delayer seen with original light panel (image 1)

2. without any filter at ASIAir recommended flat exposure: much better RGB channel grouping in mid range (image 2)

3. repeated (with Antlia filter) at half the recommended ASI auto flat exposure (to minimize blue blow-out) and got image 3 after flat field debayer...still wide separation but blue much tamer.

I then applied number 3 as a Masterflat and image processes normally.

My take-home, and feel free if this is wrong conclusion, the Antlia RGB ultra causes a very wide channel split if used with Flat light source that cannot adjust individual channel brightness which leads to blue overexposure if used with ASIAir recommended flat autoexposure.

BlockHead

I am not certain of your conclusion.

The way to think of it is that the combination of your filter and light source (your effective sensitivity) is weak in red. So your exposure time at a given light level is given by monitoring the blue pixel values. Once you know what gives you 1/2 saturation at a given light level of your panel.. you will want to take 10-15 flats in order to minimize the noise in the red channel since it is low S/N per exposure. This should result in the best flats you can take. In order to avoid some other pitfalls finding a brightness setting that yields 1-2 second exposures is best.

-the Blockhead