Post processing

[Jm1973]

Hi everyone. I've just recently started imaging. One of my first targets was Andromeda Galaxy M31. I watched a bunch of tutorials, took a  bunch of images, stacked them and prcoessed them in Photoshop. However they don't look right. Whereas all the images I've seen show the actual galaxy as being somewhat oval, no matter what I do with my images they are kind of round. I know may not make a lot of sense said like that, but either I am doing something totally wrong in my post processing, or the data in my shots has some kind of problem.

 

Are there any people on here who are good at post processing, and able to take my stacked tiff and see what they can do with it? It would be good to see if someone else can make something reasonable out of it. Then I would know if I need to research post processing more, or whether there is some flaw in my photography.

 

For me this stuff takes hours, but I'm aware that there are people who may have a workflow that enables them to process an image quite quickly.

Edited September 27, 2020 by Jm1973

[Chefgage]

Post up your processed M31 for us to have a look

[Aramcheck]

If you post the stacked TIFF (or a link to it, if saved to googledrive or dropbox), I'm sure some folks will have a quick play with it.

From your description it sounds to me that you've caught the central part of M31, which is much brighter than the outer edges. Have you stretched the image to convert if from linear to non-linear?

M31 is also a big target, so worth checking that it will fit in your camera/scope's field of view. I like to use the free Stellarium program to do this, when planning a night's session, but there are other tool's out there too.

Cheers
Ivor

[happy-kat]

How many lights did you take, what was the exposure length and what equipment camera/telescope-lens/mount was used please?

[Jm1973]

Hi guys. Thanks for the replies. 

I took about 100 30-second lights, at iso800, with an unmodded d3300 connected directly to a Sky watcher 200p explorer, in an eq5 synscan pro. I know it's a bit overweight for that mount, but at 30 second exposures it seems ok. I'm only doing 30 seconds because I can't get the d3300 to do longer exposures with the intervalometer software I'm using.

I used 10 darks, 10 flats and 10 bias frames.

I have stretched out with adjustment curves, and also set the black point with adjustment levels. I tried masking off the core so it didn't blow out everything else when I stretched it. Here is a compressed jpeg of the processed image:

Here is a link to the stacked tiff:

https://www.dropbox.com/s/8n1nl4d337sv52m/M31 260920.tif?dl=0

 

Thanks!

 

 

 

 

Edited September 28, 2020 by Jm1973

[Jamgood]

There's definitely more data in there to be stretched out. I did a quick 30 seconds stretch on it. I would recommend taking more Darks and Flats. Flats being the most important. I usually take around 30-50 Flats. I use a Tablet with a white jpg open full screen, on the telescope pointing up. AV Mode on the camera, click away. Always take Flats every session. Darks and Bias at least 20 of each. (You can re-use Darks and Bias for other sessions)

Looks like you have drifting issues as well. Are you Polar Aligning? Have you collimated your polar scope or is it still as it was from the factory? Chances are it is out.

The two things I found that are absolutely essential are good Polar Alignment and good Flats. Makes the world of difference. Combined with lots of exposure time and you get some nice images.

 

Edited September 28, 2020 by Jamgood

[Jm1973]

10 minutes ago, Jamgood said:

There's definitely more data in there to be stretched out. I did a quick 30 seconds stretch on it. I would recommend taking more Darks and Flats. Flats being the most important. I usually take around 30-50 Flats. I use a Tablet with a white jpg open full screen, on the telescope pointing up. AV Mode on the camera, click away. Always take Flats every session. Darks and Bias at least 20 of each. (You can re-use Darks and Bias for other sessions)

Looks like you have drifting issues as well. Are you Polar Aligning? Have you collimated your polar scope or is it still as it was from the factory? Chances are it is out.

The two things I found that are absolutely essential are good Polar Alignment and good Flats. Makes the world of difference. Combined with lots of exposure time and you get some nice images.

 

Wow! How did you get it to look like that?

I do polar align, but I think it's not perfect by any means. I haven't collimated it since I had it, so that definitely needs to be done.

I will take your advice with regards to the flats, (and darks and bias). 

I think I also need to learn how to stretch more, as you've got it to look 100 times better than me in 30 seconds, whereas it took me about 2 hours to make it look worse.

 

 

[Chefgage]

13 minutes ago, Jm1973 said:

Wow! How did you get it to look like that?

I do polar align, but I think it's not perfect by any means. I haven't collimated it since I had it, so that definitely needs to be done.

I will take your advice with regards to the flats, (and darks and bias). 

I think I also need to learn how to stretch more, as you've got it to look 100 times better than me in 30 seconds, whereas it took me about 2 hours to make it look worse.

 

 

Take a look at this YouTube video. The poster uses GIMP to process but the same will apply if you are using Photoshop.

 

Edited September 28, 2020 by Chefgage

[Jamgood]

54 minutes ago, Jm1973 said:

Wow! How did you get it to look like that?

I do polar align, but I think it's not perfect by any means. I haven't collimated it since I had it, so that definitely needs to be done.

I will take your advice with regards to the flats, (and darks and bias). 

I think I also need to learn how to stretch more, as you've got it to look 100 times better than me in 30 seconds, whereas it took me about 2 hours to make it look worse.

 

 

Post processing is difficult. I've only been doing this since July so I'm a newbie too. Great images can be achieved, it just just takes time to learn and refine the art. (Not that I've mastered anything yet by a long shot)

There's an add on for Photoshop that you want if you haven't got called Camera Raw Filter. It's a great tool for helping you to edit your images. (In the video I used the keyboard shortcut but when installed it is under Filter)

Here's a short video I've literally just made showing a rough way of stretching data, without using layers or anything special. I do a lot more than this now when editing but this shows that some simple editing can make a good enough image for a beginner. This is the Wizard Nebula I shot last night. 

EDIT* Forgot to say, definitely collimate your polar scope. It makes life so much easier. It's very simple to do, it just takes a little time so when you have an hour or so to spare, do it. There's some good tutorials on YouTube to explain how it's done. I'm only using an EQ3Pro but I can get polar alignment that works without hitch.

 

Edited September 28, 2020 by Jamgood

[Jm1973]

Thanks Jamgood. I will take a look at that. I appreciate you taking the time to post this. 

Post processing does seem like a bit of a dark art. I haven;t used PS for years either so I don't really have a solid base to build on, but I'm sure I'll get there in the end.

Justr googled that. I think my polar scope is a bit out. If I foces on something and turn the RA, it does move about. I'll get that sorted out as well.

 

@Chefgage Thanks for that. I have GIMP too so I'll take a look at that also.

[Jamgood]

This is a good tutorial for setting your polar scope. I used this. This fella is awesome. Watch all his videos and learn. He'll teach you so much. 

[happy-kat]

Also before changing to AV mode on the camera make sure the lens is set to manual focus, otherwise focus is liable to change when a shot is taken if left on automatic.

[ollypenrice]

Post processing is the interesting bit and begins with the stretch.

Here's where to start. (This is in Photoshop.) We open the linear (unstretched) image and check it in Levels. There is a big gap on the left of the histogram where the thin dataline doesn't reach the left hand side. That means you're not using the full range of brightnesses available...

...so, move the black point slider (lower left) to the right almost to meet the start of the data line. Why 'almost?' Because it is vital not to clip into that line. It contains your precious faint data. I'll respectfully suggest that this was where Jamgood slipped up. The error is known as 'black clipping' and is very common among beginners.

Now when we look at the image we see this:

...so we can very carefully bring in the black point to meet the data line like this:

...an d this gives us a nice healthy histogram:

Now we are getting close to a reasonable background sky brightness. My favourite is 23 per channel but I'll settle for 21 or 22. To find out where we are we can use the eyedropper tool to put 4 colour sampler marks on the background. I've ringed those in red and ringed the place where Ps shows you the values.

We notice that green is consistently low (which is actually unusual) so we can give it a lift either in Levels, Curves or the Colour Balance menu. I went for colour balance and didn't try to get it right up to the level of the others because green is 'the problem colour' in AP.

That'll be a bit better and we can tweak it more carefully later. At this point there are different approaches. A good number of people would stretch again in Levels, which move the histogram peak further to the right, and then bring the black point in to the right as well to cut back to the desired background level of 23 or whatever. The trouble with that is that you'll be stretching the background sky and so increasing its noise. This will mean you'll need nasty noise reducing algorithms to make your background look shiny black. No no no!!! 😁  So I've chosen to stretch further in Curves rather than levels. Below, I have a placed a fixing point at the background level and a further fixing point below it, and then I've lifted the curve above that to stretch only the galaxy and stars. Notice that the curve rises steeply and then flattens to a straight line to the finish. This stops the core of the galaxies and the stellar cores from being over-stretched to saturation point.

There is still a long way to go with this image but a careful stretch, always looking at the image and always thinking about what this particular image needs, is the way to proceed. Work in small steps and never black clip. When you've clipped it, it's gone for good.

Olly

 

 

 

 

Edited September 28, 2020 by ollypenrice
Typo

[Jm1973]

@Jamgood - Thanks for the recommendation! There are so many videos out there, it's good to have someone filter the good stuff!

@happy-kat - That's a good point. I will check that next time I shoot. Thanks!

[Jm1973]

@ollypenrice

Hi Ollie. Thank you so much for the very detailed reply. That is above and beyond mate. 

I will be following along with this as soon as I finish work, trying to see if I can do the same.

This site is so full of helpful people, it really it a pleasure to come here to learn.

 

[Jamgood]

@ollypenrice

Cheers for that. As I said, I'm a complete newbie too. Yours is a much better explanation than mine. 👍

[Aramcheck]

I had a quick play with the TIFF. Looks to me that you had movement of the mount on one image in the stack, resulting in the star trail.

Cheers
Ivor

[ollypenrice]

1 hour ago, Jamgood said:

@ollypenrice

Cheers for that. As I said, I'm a complete newbie too. Yours is a much better explanation than mine. 👍

I normally run lots of imaging tutorials with guests but we are, of course, 'Covided out' for now!

Olly

[gorann]

2 hours ago, Aramcheck said:

I had a quick play with the TIFF. Looks to me that you had movement of the mount on one image in the stack, resulting in the star trail.

Cheers
Ivor

Good point Ivor!

Jm1973: you should go over your stack and find that one (or those few) with the star trailing and remove it (them) - and stack again. You are onto a good beginning!

[Jm1973]

Hi Aramcheck. I think you are right, It is due to some bad images rather than the tracking being out that much.

I'll go through the images and see if I can get rid of the dodgy ones and stack it again.

Thanks for having a go at the tiff.  Your processed image looks great! Gives me something to aim for!

 

(sorry got confused with who posted what earlier)

Edited September 28, 2020 by Jm1973

[Jm1973]

4 hours ago, Aramcheck said:

I had a quick play with the TIFF. Looks to me that you had movement of the mount on one image in the stack, resulting in the star trail.

Cheers
Ivor

Hi. You were correct. There was one particular RAW file which had very pronounced trailing, much more so than any other.

I have taken that one out, and also a number of other ones with a smaller amount of trailing. I am re-stacking now, and will see how it turns out.

Thank you for your insight, and you're great work with the image.

[Jm1973]

@ollypenrice

II've just spent an hour or so following along with your steps. I feel I have learnt a fair bit about where tools are in PS

I start with moving the black point up the same amount as you did:

Giving this:

I can't move the black point again as much as you did now, but I move it up to 7, being careful not to clip the histogram:

Giving this:

 

Next  I zoom in to make sure there are no stars etc where I sample, and I sample 4 points in the black. But the RGB values I get are very low. Nowhere near 22-23. Not sure what I did there.

 

 

I played about with changing values with colour balance, but I could only tweak them in relation to each other, not increase them overall.

I cannot slide the black point any further to the right without clipping the histogram, so I've had to leave the RGB values where they are. Ah well, onward and upward!

So I continued on to the curves. I have tried to emulate the curve you did, and I got something like this:

 

I noticed that my histogram is further to the left than yours at this point, not sure why that is. Setting the points on the curves is quite the fiddly task!

Finally I got something resembling your curve and ended up with this!

 

 

This is a lot better than I had previously, and I feel I have learnt  some useful stuff. Lots of practise is what's needed now!

I am confused about how the RGB values were so off though....

 

[wimvb]

6 hours ago, Jm1973 said:

I am confused about how the RGB values were so off though....

Light pollution is a likely candidate. If you get the histogram so far to the right with only 30 seconds exposures on an unmodded dslr, you probably have a fair amount of light pollution. If this is the case, consider investing in a light pollution filter. To compensate for the added noise, you will need long total integration times, ie lots and lots of data.

[ollypenrice]

OK, the histogram on yours is slightly different to mine, partly because you are using fewer subs and partly because the poor subs will have introduced a little spurious faint signal into the stack. This signal would appear (I reckon) in the tiny flat line to the left of the histo peak which mine has and yours doesn't. Compare the raw histogram in yours and mine. Mine (your data but you know what I mean!) has a short flat line before the peak. Yours doesn't. While I've offered a theory as to why that might be, I can't be sure it's correct. This does change things in the stretching.

When you get to this stage...

...your background is still too low to move into the Curves phase with the pinned background sky. You'll need to give it a further stretch in Levels to get the background higher, closer to 23. Now your histogram isn't going to behave as mine did so you may not be able to get it to 23 without saturating the cores. If so, settle for less but certainly not less than 20. When you went into the Curves phase the faint nebulosity had not been adequately separated from the background so you couldn't be sure what you were pinning, background or nebulosity. (You had realized this when you said, 'I noticed that my histogram is further to the left than yours at this point, not sure why that is. Setting the points on the curves is quite the fiddly task! '  Quite so, it was fiddly because the Levels stretch had not, on the new stack, opened up the histogram sufficiently. Give it another go in Levels. Or two. Or three!

With one of my cameras I find the background sky stubbornly refuses to rise to the desired 23 so I settle for less but, when I see this coming in a stack, I go back to the beginning and give a pre-stretch to the lowest brightnesses using a curve like this:

 

I lifted the lowest values here form 6 to 9 and then straightened the curve to the finish in hopes of lifting the background a tad to start with while leaving the rest linear. Then I'd go back to the Curves stretches as usual. Your histogram, since it rises so sharply from nothing, might be more manageable after a pre-stretch like this but I can't be sure. While this technique does help, I must say that it's less effective than I expected it to be when I first thought of it! 🤣

One thing we do see here, though, is that every image is different. There is no workflow that suits all - thank goodness. That would be dull.

Olly

PS The trick for placing the background marker on the Curve is to put the cursor on a bit of background and then Alt Click and the point goes onto the Curve in the right place.

Edited September 29, 2020 by ollypenrice

[wimvb]

Before you stack the sub frames, you need to calibrate them. For the setup you used to take this image, good flats are essential. This is what the gradient in your image looks like. The circular structure is vignetting. But there is also a colour gradient with a red band at the top. Most software can handle one source of an uneven background quite well, but a combined gradient can be a challenge to remove. Especially if you also have different gradients in the three colour channels, which makes it even more of a challenge to remove.

Here's what I managed to pull out of your data, before the remains of gradients started to pop up. Nothing too fancy. Arcsinh stretch (available in PS) lifted the colour in your image. after that, curves transformation as Olly already wrote about.

(click on the images to see a larger version)