Deep Sky Stacker Alignment.

[Prolifics]

Hi guys

I am suffering to being able to understand something as a new imager.

I imaged an object that went past meridian and my mount done a automatic meridian flip but from then on the pictures were upside down. Using APT
I put them in deepsky stacker and made a reference frame for each of LRGB images and got them all to stack fine and saved them as L.tiff R.tiff G.tiff and B.tiff. Each colour are aligned fine. (now each colour are all the same way up)
However when i put them in photoshop when I make an RGB channel and paste each one in I get red and green dots in the RGB master which shows although each colour are aligned they are not aligned with each other.

So question is when I have stacked each LRGB although individiually they are stacked fine how do I get them to stack with each other. Stars are slightly missaligned in each colour to each other.

Very hard to describe but hopefully you understand.

 

Thanks in Advance David

 

 

 

 

 

[Craney]

Hi David,

I use a program called  REGISTAR which scans and aligns any image you care to put into it. 

As long as it has a few stars as a reference it will make a good job of arranging all the stacked final components of L R G B or Ha O3 S2  into a common fixed frame.

Then I use PS to do the adding and tweaking, although REGISTAR does itself have a basic combination feature.

It is quite expensive, and there may be other programs that do the same within a bigger package, but I find it works very well.

 

Sean.

Edited July 26, 2019 by Craney

[Freddie]

I think what you are describing are hot pixels on each of the colour stacks. Because there is slight drift from image to image but the hot pixels are in a fixed position on the chip you see the hot pixels show up across the combined RGB. If you look closely, I’m sure you will also see blue “dots”. What you need to do is set up to have hot pixels removed during the DSS stacking process. It has been years since I have used DSS so can’t remember where that setting is but it’s in DSS somewhere!

[Prolifics]

Hi Sean

 

Thanks for your reply. Each colour is stacked and aligned by itself. But each colour has drift like you said of a few millimeters. Isn't there a program that aligns your LRG and B data after they are stacked?

For an example say I take 30 exposures of Lrgb 1 night then decide to finish it in 2 months time. They are going to be slightly different.

 

Here are examples of my RGBL data as you can see the stars have rotated slightly. My mount has just been hyptertuned and was guiding at 0.40 average in PHD2. Polar aligned with Sharpcap to 00.08 on each or better.

L 60x 1 Min
R 30 x 1 Min
G 30 x 1 Min
B 30 x 1 Min
30 Flats of each 
60 Darks

 

 

 

Finally RGB when put in there channels without luminance. Now you can see the green and blue where we arent alligned with each other.

 

 

 

Regards David

Edited July 26, 2019 by Prolifics

[Prolifics]

Also using deepskystacker to register LRGB. darks and flats all togther as one lump works fine but not sure if I can work with that in photoshop. So if it can align is there an option to compare frames from each stacked colour?

 

Cheers David

[wimvb]

In dss you should be able to set one subframe as alignment reference. Always use the same sub for this, eg one L sub, even for R, G, and B subs. Just exclude that sub from the stack.

From the dss faq

How do I align the resulting images of 4 stacks (red, green, blue and luminance)?
To align the resulting images on the same reference frame just add the reference frame to the list even if it is not from the same stack, force DeepSkyStacker to use it as the reference frame (using the context menu) but left it unchecked. 
This tells DeepSkyStacker to use it as the reference frame but to not add it to the stack.

http://deepskystacker.free.fr/english/faq.htm

 

 

Edited July 26, 2019 by wimvb

[Prolifics]

8 minutes ago, wimvb said:

In dss you should be able to set one subframe as alignment reference. Always use the same sub for this, eg one L sub, even for R, G, and B subs. Just exclude that sub from the stack.

From the dss faq

How do I align the resulting images of 4 stacks (red, green, blue and luminance)?
To align the resulting images on the same reference frame just add the reference frame to the list even if it is not from the same stack, force DeepSkyStacker to use it as the reference frame (using the context menu) but left it unchecked. 
This tells DeepSkyStacker to use it as the reference frame but to not add it to the stack.

http://deepskystacker.free.fr/english/faq.htm

 

 

Thanks Wim I am going to try that shortly. I had already used a reference frame but one for each colour now I see the problem. I will let you know if it works for me.

 

Kindest regards David

[Prolifics]

Thanks everyone!

 

It worked and with my rubbish post processing I got this in 10 minutes with very limited data. This .jpg so very low detail. Iris Nebulae of course

[wimvb]

Very nice. What camera do you use? If you get 0.4" guiding rms, you should increase the subframe exposure time. But if your camera has low read noise (modern cmos), you should be able to stretch your final image harder to pull out more detail and colour.

For example, this is what 3.6 hours of 30 and 45 seconds exposures got me.

https://www.astrobin.com/318629/C/?nc=user

Compare the different versions.

[Prolifics]

32 minutes ago, wimvb said:

Very nice. What camera do you use? If you get 0.4" guiding rms, you should increase the subframe exposure time. But if your camera has low read noise (modern cmos), you should be able to stretch your final image harder to pull out more detail and colour.

For example, this is what 3.6 hours of 30 and 45 seconds exposures got me.

https://www.astrobin.com/318629/C/?nc=user

Compare the different versions.

Hi Win

 

All my equpment is in my signature. It's not far different from yours except the smaller scope

I only had half your data and don't think I autofocused before the session as I was just playing with my new hypertuned mount. HWFM and HDR normall around 1.3 when i focus.

I'm using unity game on my ZWO asi1600mm pro. 139/50 the 50 is so it shift the histogram to the right slightly and dosnt crop it. I was using gain 75/15 but saw a youtube video and decided to give the new gain a go.

Plus my post processing is just basics. Adjusting curves and levels. Removng noise and making an RGB layer and adding the Lum channel.

Thats about all I can do sa far. This is my 4th ever image

 

Regards David

 

 

 

[Prolifics]

45 minutes ago, wimvb said:

Very nice. What camera do you use? If you get 0.4" guiding rms, you should increase the subframe exposure time. But if your camera has low read noise (modern cmos), you should be able to stretch your final image harder to pull out more detail and colour.

For example, this is what 3.6 hours of 30 and 45 seconds exposures got me.

https://www.astrobin.com/318629/C/?nc=user

Compare the different versions.

Hi Wim,

 

I think I am a little out of focus but here is a link to my stacked LRGB files it would be nice to see what you can get from it, to see if my data is any good.

 

https://www.dropbox.com/sh/skqoxrbe5huosg4/AAAdQ_UEP0XsX0S0b6oBPiiKa?dl=0

 

Regards David

[wimvb]

I'll have a look at it and see what I can produce (with pixinsight).

[Prolifics]

Wish I could afford that program

My stars are not sharp and numerous like other pictures. I think it's more exposure time and gain settings that you were saying. Any good pointers would be appreciated but first I need to know if its my data or my post processing. I think its both

Cheers for looking at it for me! Oh by the way there are no dark frames yet only flats. I still havent took them

 

Kindest regards David

 

[wimvb]

In the preprocessing (calibration and stacking) stage, I use darks, flats, and dark flats. The latter to calibrate the flats.

1. Create master dark flat by combining the dark flat subframes (same time, gain, offset and temperature as flats)

2. Calibrate flat subframes by subtracting the dark flat master from the flat subframes.

3. Combine the dark frames into a master dark. Dark subframes must have the same temperature, gain, offset and exposure time as the light frames. This is important.

4. Calibrate the light frames with the master dark and master flat.

I don't know if dss can calibrate flats that way. But it is able to use master darks and master flats in stead of the dark and flat frames.

My ASI camera has more amp glow than the 1600, so matching darks are crucial, both for lights and flats.

[Prolifics]

Thats basically what I do when I stack. I don't use bias frames as apparently my Cmos camera dosn't need them. I have been doing 1 min exposures. I think it may be better aiming for 3 minute ones to bring in more stars. I still need to work with the gain and see whats best on certain objects.

[wimvb]

I had a quick look at your data. Some general remarks:

1. the  R, G, B frames weren't completely aligned. Espcecially the green was off: 

rotation : +0.01 deg

dx : +1.62 px

dy : +2.39 px

2. your flats are over correcting. This is either a result of flat calibration, ie, your flats need to have their own darks subtracted as I wrote previously, or the absence of darks. 

Here's what the central area of your image looks like after a rough process in PixInsight.

Here is the rgb master with only a stretch applied to show the flat over correction

 

[Prolifics]

Thank you for that. My flats are taken with APT with an ADU of 25000. Using a lit flat panel and a white T-shirt. I am not using a flattener yet. Any suggestions?

I will take some darks tomorrow remake all the LRGB files again in Deepskystacker and see how that does. Then I will know for sure.

 

Regards David

 

Edited July 26, 2019 by Prolifics

[Prolifics]

Hi Wim,

 

I managed to do the darks same temperature gain offset and exposure time as my lights. I may have found something else I was doing wrong. I was using the same master flat frame for each colour. This time I done them all seperate.

I have replaced the files and re-uploaded them. All with 30 frames of lights and 60 darks. I also saved them in 32 bit Tiff format as to not lose any data.

Hopefully you will have time to look at these and see if its the darks not being added has caused this problem or something else I need to work on.

https://www.dropbox.com/sh/skqoxrbe5huosg4/AAAdQ_UEP0XsX0S0b6oBPiiKa?dl=0

 

Kindest regards David

 

 

 

 

 

[wimvb]

I've only downloaded the data, but I must say that it looks a LOT better. Just a slight vignetting effect left, no problem for PixInsight DBE.

[Prolifics]

 

 

1 hour ago, wimvb said:

I've only downloaded the data, but I must say that it looks a LOT better. Just a slight vignetting effect left, no problem for PixInsight DBE.

I just finally spent out and bought https://www.365astronomy.com/ts-optics-photoline-2-0.79x-4-element-reducer-and-field-flattener-corrector-for-astrophotography.html

Cheaper than Pixinsight

I did speak to my friend Zoltan there and he said I may be getting Vigneting because of the backspace 2 inch going to a 1.25 cone and into back of my scope. Thats the adapters that came with my filter wheel and camera.  So not only will this fix that but is a matched flattener for my scope and 0.79 reducer. So the Flattener should stop the elongated stars at the far left and right of my shots.

Also the Reducer will lose 20% of my size because of the reduction of focal length but will gain x1.6 exposure time. Turning my F6.5 to a F5.13

I am away all of next week  so I will keep you updated somewhere on this forum with the results

Would be nice to compare the new data with the old if you get 10 minutes.

 

Have a good weekend.

 

Regards David

 

 

Edited July 27, 2019 by Prolifics

[wimvb]

4 hours ago, Prolifics said:

Would be nice to compare the new data with the old if you get 10 minutes.

A little more than 10 minutes, but here is my result. There is a lot of noise in the image, and it's pushed a bit too far, but at least it shows what's in there, and that your calibration process now works. I didn't do anything about the stars.

[Prolifics]

Thanks Wim for that. You can see its like day and night compared to the one without darks.

You have even got some of the black nebulae out from it even with only 2.2 hours of data. Not sure what to try next. I think its expreminent with gain next. Raising the from 1 Min to 3 Min exsposures will help bring a sharper and more detailed image as my tracking is good. Plus my Focul reducer/Flatterner will improve things. I also took 3 minute dark frames today while I did the 60 second ones

Not sure what the white foggy stuff is? Did you set the black levels and grey and white levels? I know its zoomed in alot so will show a lot more noise and imperfections.

Looking forward to any tips you may have.

 

Thanks David

 

[wimvb]

The new field flattener should give you round stars, once you got it tuned (correct distance to sensor). Further than that, you can experiment with longer exposure times. When you image R, G, B the sensor receives only about a third of the number of photons as compared to L, so in theory you can expose up to 3 times longer. In your Iris image, the L frames seem to have a good exposure time, but the R, G, B frames could then be exposed eg 2 minutes. If you aim at 2 x as much integration time for L than R, G, B, you'd need 4 x as many L frames as colour frames.

120 x 1 min L

30 x 2 mins RGB

Don't vary temperature and gain/offset too much, because you need darks for every combination of exposure time, gain and temperature. At low gain and longer exposure times, your camera has a larger dynamic range, and you'll need fewer subs to get to a certain level of signal to noise. If storage space and computer speed are a concern, you may need to factor this in.

The total integration time you need to get to a certain result (signal to noise ratio) depends very much on your sky conditions. According to an article by Jerry Lodriguss, you need 2.5 times more integration time for every magnitude of sky darkness/brightness you lose. Eg, if you image 1 hr from a dark site (magnitude 21), you need 2.5 hours to get an equivalent result from a magnitude 20 site, and 6.3 hours from a magnitude 19 site. It pays off to look for dark skies, if you have that possibility.

9 hours ago, Prolifics said:

Not sure what the white foggy stuff is? Did you set the black levels and grey and white levels?

I played with "curves" after initial stretching to increase local contrast. I measured the lightness in two points where I wanted to increase contrast, then pinned one of those points and increased/decreased the other.

Good luck.

[Prolifics]

29 minutes ago, wimvb said:

The new field flattener should give you round stars, once you got it tuned (correct distance to sensor). Further than that, you can experiment with longer exposure times. When you image R, G, B the sensor receives only about a third of the number of photons as compared to L, so in theory you can expose up to 3 times longer. In your Iris image, the L frames seem to have a good exposure time, but the R, G, B frames could then be exposed eg 2 minutes. If you aim at 2 x as much integration time for L than R, G, B, you'd need 4 x as many L frames as colour frames.

120 x 1 min L

30 x 2 mins RGB

Don't vary temperature and gain/offset too much, because you need darks for every combination of exposure time, gain and temperature. At low gain and longer exposure times, your camera has a larger dynamic range, and you'll need fewer subs to get to a certain level of signal to noise. If storage space and computer speed are a concern, you may need to factor this in.

The total integration time you need to get to a certain result (signal to noise ratio) depends very much on your sky conditions. According to an article by Jerry Lodriguss, you need 2.5 times more integration time for every magnitude of sky darkness/brightness you lose. Eg, if you image 1 hr from a dark site (magnitude 21), you need 2.5 hours to get an equivalent result from a magnitude 20 site, and 6.3 hours from a magnitude 19 site. It pays off to look for dark skies, if you have that possibility.

I played with "curves" after initial stretching to increase local contrast. I measured the lightness in two points where I wanted to increase contrast, then pinned one of those points and increased/decreased the other.

Good luck.

Thanks for explaining the exposure times for each filter. I was doing it half right. I always do double the Luminance amount but didn't relise about the RGB impact in light gathering. Now it all makes sense.

Regarding Backfocu I will have 56mm I need 57mm so I bought some 1mm 1.5mm 0.75mm etc Baader adjustment rings so I can play with that. What is the best way to test your Flatterner? Guessing I would say take a 3 minute exposure of Luminance? Check for elongated stars at the edges and adjust back focus from there?

 

My garden is a Bortle 5 site magnitude is around 20.27 for me.  So not terrible. I guess where you are is great because you get less daylight than anyone else  Plus the cooler conditions means less noise and I guess the northern lights or Aurora may hinder you at time?

 

I also have a friend who lives up there in the north of Sweden.

 

Kindest regards David

[wimvb]

2 hours ago, Prolifics said:

Regarding Backfocu I will have 56mm I need 57mm so I bought some 1mm 1.5mm 0.75mm etc Baader adjustment rings so I can play with that. What is the best way to test your Flatterner? Guessing I would say take a 3 minute exposure of Luminance? Check for elongated stars at the edges and adjust back focus from there?

That's how I would do it. There is a chart on this forum somewhere that shows the direction of star elongation related to flattener - sensor distance.

Here is the link:

 

2 hours ago, Prolifics said:

My garden is a Bortle 5 site magnitude is around 20.27 for me.  So not terrible. I guess where you are is great because you get less daylight than anyone else  Plus the cooler conditions means less noise and I guess the northern lights or Aurora may hinder you at time?

I have about the same darkness where I live. I also have the light bubble of Sotckholm in the Sourthwest to deal with. Not too bad, as it is low enough. Auroras can be spectacular, but if they are very faint, they're just a form of light pollution. Unfortunately we get too few of the spectacular ones. The long winter nights are great, I can start imaging from about 5 in the afternoon during the darkest season. Otoh, from late April to late August I have my gear in summer storage, as it just doesn't get dark enough to to any imaging.

The problem is often that during the winter the cold not only stiffens my limbs, but also any cables hanging from my rig. I've started to replace all power cables with silicone insulated cables. And I try to keep my data cables in such a position that they don't need to bend/move too much. As for camera temperature, I tend to keep that at -20 C all year, but sometimes I need to lower it to -30 to keep it constant. I sometimes worry what the cold will do to the grease and belt in my mount. I think that I run most of my gear well outside operating temperature range at times. As the saying goes; "there ain't no such thing as a free lunch".