M31 / First light with 2600MC

[AbsolutelyN]

5 minutes ago, gorann said:

Well, that banding is odd - have you shown the flat to ZWO?

Yes I sent ZWO screenshots of lights, darks, bias + flat. All I got was a one line response saying please calibrate with flats. And indeed calibrating with flats comes out with a pretty much spot on image. I don't see why I should need to use flats though. Very strange, I'm keen to know if anyone else has this banding or it's just my camera. 

Edited November 10, 2020 by AbsolutelyN

[gorann]

Well I have not seen them in my lights. Could it be a cable issue, with the USB cable being exposed to DC current from another cable? Hopefully someone here on SGL has an idea.

[AbsolutelyN]

2 minutes ago, gorann said:

Well I have not seen them in my lights. Could it be a cable issue, with the USB cable being exposed to DC current from another cable? Hopefully someone here on SGL has an idea.

That's an interesting point thanks. I guess I can test that by removing the camera and taking new flats with a short USB cable and laptop to eliminate anything that might be interfering in my current setup as that is a possibility.   

[gorann]

32 minutes ago, AbsolutelyN said:

That's an interesting point thanks. I guess I can test that by removing the camera and taking new flats with a short USB cable and laptop to eliminate anything that might be interfering in my current setup as that is a possibility.   

Hopefully it is a cable issue. Keep us informed what you find out.

[AbsolutelyN]

1 minute ago, gorann said:

Hopefully it is a cable issue. Keep us informed what you find out.

Will do, thanks for all the help. 

[AbsolutelyN]

12 hours ago, gorann said:

Hopefully it is a cable issue. Keep us informed what you find out.

Well I took new flats on my Esprit 80. Different computer, shorter better shielded USB 3 cable and different power supply. The banding is still there and clearly visible with nothing more than a STF stretch.

 

[gorann]

Odd. I have only once taken flats with my ASI2600 and it was 1 second sky exposures at dusk (gain 0. cooled to -10°C). Here is what one such debayered sub looked like - no banding whatsoever.

[AbsolutelyN]

Just now, gorann said:

Odd. I have only once taken flats with my ASI2600 and it was 1 second sky exposures at dusk (gain 0. cooled to -10°C). Here is what one such debayered sub looked like - no banding whatsoever.

Now that's what the flat should be like. What ADU are you exposing at? Although you can calibrate the banding out I do now suspect my camera is a dud. I've asked FLO about possibility of replacing it so will hopefully get it sorted. 

[gorann]

2 minutes ago, AbsolutelyN said:

Now that's what the flat should be like. What ADU are you exposing at? Although you can calibrate the banding out I do now suspect my camera is a dud. I've asked FLO about possibility of replacing it so will hopefully get it sorted. 

Not sure about the ADU, I just took a series of 1 second exposures with the RASA 8 pointing at the sky at dusk, selecting an exposure time that kept the histograms for the colour channels away from being clipped. Here is what the histogram looks like (in PS)

[gorann]

PS. I also notice that you have much darker corners in your RASA8 flats with the ASI2600 than I get. That is also quite odd.

[kirkster501]

Splendid picture.  If I may say, to be constructive, I think you have gone slightly too far with the sharpening and HDR.   Here is my version with same sensor, which needs more processing:

 

 

 

[AbsolutelyN]

12 minutes ago, gorann said:

Not sure about the ADU, I just took a series of 1 second exposures with the RASA 8 pointing at the sky at dusk, selecting an exposure time that kept the histograms for the colour channels away from being clipped. Here is what the histogram looks like (in PS)

My flats look similar in photoshop - screenshot has a STF stretch in PixInsight so is exaggerated. My darks are slightly darker I think. 

[AbsolutelyN]

7 minutes ago, kirkster501 said:

Splendid picture.  If I may say, to be constructive, I think you have gone slightly too far with the sharpening and HDR.   Here is my version with same sensor, which needs more processing:

 

 

 

Thank you and yes completely agree - I've been working on a second version which has less aggressive sharpening.  Always learning. great image. 

[kirkster501]

You need to be very delicate with HDR multiscale transform, if that's what you are using.  The default is much too aggressive.  I use a range mask and set the scale to 8 or 9 with lightness mask ticked

[AbsolutelyN]

4 hours ago, kirkster501 said:

You need to be very delicate with HDR multiscale transform, if that's what you are using.  The default is much too aggressive.  I use a range mask and set the scale to 8 or 9 with lightness mask ticked

I think it was mainly Topaz Denoise in Photoshop that I must have used too aggressively - probably amongst other things.

Here is version two which I hope addresses some of the original issues. Sure there is still much that could be improved.  

Edited November 11, 2020 by AbsolutelyN

[gorann]

16 hours ago, AbsolutelyN said:

I saw the Artesky adapter and was tempted but not planning on using filters. Currently using the Baader 7.5 mm T adapter.  It would be good if Celestron just redesigned the front plate for better / repeatable image quality. As it is every time the camera moves it feels like pot luck on your star quality. It's done better than I thought though. 

Two questions though if I may? 

1. Have you removed the optical window from the RASA? Feels like just an extra piece of glass in the optical train but I'm worried if I remove it it might make focal position different. 

2. If you run a DBE on your flats do you get this banding pattern?

   

There is something in your image train that seem to totally cut out the corners, making them all black. I do not get that. The useful image circle of the RASA 8 should be 32 mm, so it should do better on the 28 mm diagonal of your camera, with some rather slight vignetting.

[AbsolutelyN]

2 minutes ago, gorann said:

There is something in your image train that seem to totally cut out the corners, making them all black. I do not get that. The useful image circle of the RASA 8 should be 32 mm, so it should do better on the 28 mm diagonal of your camera, with some rather slight vignetting.

Pretty sure it's just the STF auto stretch exaggeratig the light fall off - if you look at raw de-bayered flat in PS it's similar to yours. I suspect if you do a STF autostrech in pixinsight you'll get similar. 

[gorann]

No, I did that and also an Equalize in PS which is extreme stretching and I do not get those dark corners. I do see a defocused star in the sky (showing that collimation is fine). Here is the most extreme autostetch option in PI

[gorann]

Here is the raw sub if you want to do a direct comparison with yours in PI - I am more a PS person 😀

2020-08-15-1945_3-CapObj_0003.FIT

[AbsolutelyN]

34 minutes ago, gorann said:

Here is the raw sub if you want to do a direct comparison with yours in PI - I am more a PS person 😀

2020-08-15-1945_3-CapObj_0003.FIT 49.77 MB · 1 download

Brill thanks. Mine was a master flat but your single frame is very similar. If I apply a screen transfer function you exaggerate the fall off at the edges. 

Even more interestingly if I run a DBE and then STF you can actually just about see the same banding though not as strong. Probably because my flats are darker and it was a master flat. Maybe it is integral to the sensor after all if it's on more than one copy of the camera. 

 

Edited November 11, 2020 by AbsolutelyN

[AbsolutelyN]

28 minutes ago, gorann said:

Here is the raw sub if you want to do a direct comparison with yours in PI - I am more a PS person 😀

2020-08-15-1945_3-CapObj_0003.FIT 49.77 MB · 2 downloads

He is my master flat for comparison

FLAT_CALIBRATED_BIAS_DARKFLAT_NEW.xisf

[gorann]

If I also have banding after an extreme stretch then it did at least not bother me during the processing - I never noticed it. So maybe it is something in your processing that brings it out to an extent that it becomes a problem for you. I only use PI for a few specific things (primarily stacking and a bit of DBE), and then do the rest including all stretching in PS so I have hard to come with any suggestions. By the way, at what temperature did you run the sensor?

[AbsolutelyN]

Just now, gorann said:

If I also have banding after an extreme stretch then it did at least not bother me during the processing - I never noticed it. So maybe it is something in your processing that brings it out to an extent that it becomes a problem for you. I only use PI for a few specific things (primarily stacking and a bit of DBE), and then do the rest including all stretching in PS so I have hard to come with any suggestions. By the way, at what temperature did you run the sensor?

Yes agree, maybe I'm digging too deep into the files and you are bound to find artefacts. I do stacking, DBE and star reduction in PixInsight, most of the rest in Photoshop. However I do challenge anyone to process the attached file (M31 stack of 4 hours with no flats) and not find the banding in the image - it's pretty prominent once you run a DBE and start to work on the image. 

Temp was -10 but I got the same at -20 a few days earlier so don't think it's temp dependant. I'm using 2 min exposures form Bortle 5. 

integration4.xisf

[gorann]

9 minutes ago, AbsolutelyN said:

Yes agree, maybe I'm digging too deep into the files and you are bound to find artefacts. I do stacking, DBE and star reduction in PixInsight, most of the rest in Photoshop. However I do challenge anyone to process the attached file (M31 stack of 4 hours with no flats) and not find the banding in the image - it's pretty prominent once you run a DBE and start to work on the image. 

Temp was -10 but I got the same at -20 a few days earlier so don't think it's temp dependant. I'm using 2 min exposures form Bortle 5. 

integration4.xisf 298.95 MB · 0 downloads

I wonder if DBE could create the problem. Like Gradient Exterminator in PS DBE can cause odd effects. If you have a good master flat then in principle you should not need to do DBE. Maybe @wimvb who is a PI expert can comment on this.

Edited November 11, 2020 by gorann

[gorann]

Here is my flat run through the Equalize command in PS, which is used by many as a diagnostic tool to find banding and gradient problems and I see nothing there.