Data Analysis, good , bad, indifferent ?

[Newforestgimp]

HI All,

 

I wonder if anybody would care to review a set of data I've captured, post processing for me is a massive learning curve and ive been concentrating on trying to improve the DATA IN.
If anybody could look at the following and maybe do a quick process and advise what the main issues are to set me going on an improved path id be most grateful.

The following were captured via NINA, with an ASI071mc Pro via an RC8" scope with apprx 0.71 reduction giving apparent FL of 1166mm, 2 min guided subs at 90 Gain 20 Offset, a CLS CCD filter is in the image train

https://www.dropbox.com/sh/ye0fs1ynm67a8bz/AAC3Br8t0mlMmS4ufXvwVxNMa?dl=0

Any pointers as my attempts to process are not good, I dont know if its crap data IN Crap data OUT or Decent Data IN, crap bloke in middle causing crap data OUT 🙂 

Either way im sure it will cause the gurus some laughter, i wont be offended.

Regards & Clear Skies

Andy

[Clarkey]

I would not claim to be an expert imager - far from it, but I am in the process of downloading. Is there any issues in particular that you are having?

[Newforestgimp]

I just can’t seem to get a nice image without stretching the nuts off it, I need a workflow that proves to me if the data is good so I can have faith in the data capture technique

[Clarkey]

Ok. I have had a VERY QUICK go at this one. I have not done any star reduction just a bit of stretch, background correction and a bit of colour correction plus a bit of denoise. I'm not sure what you are expecting but the data does not seem too bad. I used 90% of the subs about as there were some with bad star shapes. I would say the following:

1. The tracking is not too good on some subs- but a EQM35 with an RC8 is asking a lot. O use my RC8 on the AZ-EQ6 as the HEQ5 is a push.
2. I would think about less filtering as you will be loosing some light and a difficult to remove green tint.
3. There does seem to be a bit a problem with the flats as the outer edges of the integrated result is very variable in brightness.
4. I think the collimation could be improved slightly

Having said all this, the actual data does not seem too bad to me. A bit more integration time will help a lot. There is also some banding - do you dither?

[Newforestgimp]

Thanks for this, it would seem I’m on the right track. It’s GEM45 mount, still getting to grips with the balance and guiding.

flats were my concern but didn’t want to lead you. Used an illuminated art panel and just let Nina sort it for me, I’ve noticed that one side cast to most images.

collimation is much better than it was not perfect though.

dithered every 4, changed to 5 for tonight, binned 2x2 also.

yes, filter, I can’t decide ? I’m in bortle 5 not sure if I need, maybe I should get just get a IR UV ?

I just didn’t want spend hours capturing if it’s unworkable and I think you’ve proven it can work but needs a bit of refinement. 👍👍👍

[ONIKKINEN]

Had a quick look as well. Flats dont seem to be working all that well and the end result has inverse vignetting for me. I dont know anything about the ASI1071MC so probably not that helpful for you but shouldn't bias frames be as short as they can be? Your bias frames are 1/50th second and the flats are 1/7th second exposure. Probably necessary to take dark-flats then to calibrate the flats properly. Looks like there is some dark current in the dark frames aswell so it could result in bad flats calibration. I dont really have dark current in my camera so cant comment on that on more than a gut feeling stance.

Also the flats have a very different vignetting look to them per color channel.

[Clarkey]

8 minutes ago, Newforestgimp said:

I’m in bortle 5 not sure if I need, maybe I should get just get a IR UV

I am in similar bortle 5 go 6 depending where the scope is pointing. Personally I don't use and light pollution filter and remove the gradients on processing. I did not like the colours with the filters.

I think you are definitely on the right track.

Edited January 14, 2022 by Clarkey

[Newforestgimp]

10 minutes ago, ONIKKINEN said:

Had a quick look as well. Flats dont seem to be working all that well and the end result has inverse vignetting for me. I dont know anything about the ASI1071MC so probably not that helpful for you but shouldn't bias frames be as short as they can be? Your bias frames are 1/50th second and the flats are 1/7th second exposure. Probably necessary to take dark-flats then to calibrate the flats properly. Looks like there is some dark current in the dark frames aswell so it could result in bad flats calibration. I dont really have dark current in my camera so cant comment on that on more than a gut feeling stance.

Also the flats have a very different vignetting look to them per color channel.

Dark current,  not sure what this is ? I will Google.

I did think the images were very noisy  considering it’s cooled (-10) maybe I need to rethink my settings and build better calibration library.

[Newforestgimp]

I do have another 100 or so subs from night 2, but I ran out of Dropbox space 🤪

[ONIKKINEN]

6 minutes ago, Newforestgimp said:

Dark current,  not sure what this is ? I will Google.

I did think the images were very noisy  considering it’s cooled (-10) maybe I need to rethink my settings and build better calibration library.

Its the thing dark frames are for. When you expose for a dark frame in complete darkness you still get signal in the image from thermal noise. When dark frames for flats are not available (darkframe with same exposure length as flat frame) a bias/offset frame is subtracted from the flat. But in your case the bias frame is 1/50s and the flat frame is 1/7s, so its possible there is unaccounted for signal in the flats = calibration doesn't work as intended. Reason i said its a gut feeling is i dont have nearly as much dark current in my camera, so bias and darkflat are similar enough that i dont care about the difference.

[vlaiv]

1 minute ago, ONIKKINEN said:

Its the thing dark frames are for. When you expose for a dark frame in complete darkness you still get signal in the image from thermal noise. When dark frames for flats are not available (darkframe with same exposure length as flat frame) a bias/offset frame is subtracted from the flat. But in your case the bias frame is 1/50s and the flat frame is 1/7s, so its possible there is unaccounted for signal in the flats = calibration doesn't work as intended. Reason i said its a gut feeling is i dont have nearly as much dark current in my camera, so bias and darkflat are similar enough that i dont care about the difference.

Difference between 1/50s and 1/7s is very small at -10°C according to this:

Even if doubling temperature is ~5°C - that would mean 0.00128K/px/s - or really negligible value at 1/7s

I suspect that it is bias instability that is the issue here (higher bias value than flat dark). I'm downloading subs and will inspect bias vs darks to see if I can find something useful.

 

[Newforestgimp]

Does the diagram suggest I should set temp to say -13oC ?

[ONIKKINEN]

7 minutes ago, Newforestgimp said:

Does the diagram suggest I should set temp to say -13oC ?

Judging from the diagram it seems that this camera also has very low dark current. Below -5 its arbitrarily low and probably doesn't cause any issues. At your temperature of -10 there is 0,0768 electrons of noise in a 120s dark frame = so pretty much irrelevant.

Actually judging from this you could run the camera much warmer and probably still not see a difference.

[vlaiv]

11 minutes ago, Newforestgimp said:

Does the diagram suggest I should set temp to say -13oC ?

It suggests that there is not much difference between say -15°C and -10°C and you should use what is convenient.

In warmer weather it is harder to reach set point - and that depends on your deltaT. If your deltaT is say 30°C - and it is 20°C outside - you'll struggle to get it down to -15°C.

[vlaiv]

@Newforestgimp

What software did you use for capture and did you adjust any sort of  "white balance" controls in that software?

[Newforestgimp]

33 minutes ago, vlaiv said:

@Newforestgimp

What software did you use for capture and did you adjust any sort of  "white balance" controls in that software?

I used NINA, only exposure, binning & dither were used.

for the flats I let Nina work it out with a flat light panel and cloth over scope.

[vlaiv]

Well, something is wrong with your flats. I can't really understand what exactly - but if I take any two flats - they don't correct each other properly.

When dividing one flat exposure with another - result should be just more or less uniform noise - this is what I get:

if I use two flats from second batch and this:

If I use two flats from first batch.

You say that you've been using some sort of cloth over objective? Did you by any chance change focus position between flats and lights?

What is difference between flats marked as 9.40 and those marked as 10.50?

[Clarkey]

The top flat explains the banding on the final image.

For my RC8 I use an A4 LCD panel for the flats. It is a fraction small but works fine.

Edited January 14, 2022 by Clarkey
Auto correct

[Newforestgimp]

No change in focus, the 9.4 & 10 is the temperature -9.4 -10, just how I configured the naming convention so I could easily tell which set of dark frames I need to use.

I actually used an A3 light pad but I think the issue could be the weave pattern of the white cloth betweeen scope and light panel. I will try without cloth.

[vlaiv]

8 minutes ago, Newforestgimp said:

I actually used an A3 light pad but I think the issue could be the weave pattern of the white cloth betweeen scope and light panel. I will try without cloth.

With scope focused at infinity - there should be no way for any pattern located at aperture to be anywhere near focus - it should be all blurred.

Even if it was pattern - two successive flat exposures should have been the same and should calibrate one another properly.

[Newforestgimp]

Maybe ‘pattern’ is a poor choice of word, the fabric is muslin, it has a weave that is like a grid, I put three layers crossing to avoid the ‘weave” affecting the light, see the diffuse panel below

[vlaiv]

3 hours ago, Newforestgimp said:

Maybe ‘pattern’ is a poor choice of word, the fabric is muslin, it has a weave that is like a grid, I put three layers crossing to avoid the ‘weave” affecting the light, see the diffuse panel below

It does look alike that thing in the image. I still think it is unrelated - but I could be wrong.

[Clarkey]

19 hours ago, Newforestgimp said:

will try without cloth

You should be fine without it. The other option is a few sheets of white A3 paper if you need to dim slightly.

[Newforestgimp]

Well I’ve been trying some flat captures today and that pattern is WEIRD !!! It’s not the cloth (removed), it’s not the filter (removed), it’s not the reducer (removed).

I’ve cleaned the glass in front of the camera sensor with a cleaning kit and pattern is still there , however, and this is where it gets really weird, it’s not always there ? I did a batch absolutely clear of pattern then another batch and pattern returned. I’m at a loss now what it might be.