Help in bring out faint detail in galaxies, nebula etc.

[StargeezerTim]

Hi all,

I am a relative newcomer who has just got into guiding and am able to produce stacked images through DSS mainly. 

I feel I need a strategy to work with under exposed or faint objects rather than my usual hit and miss approach using various free packages (GIMP, Paint.net, Registax etc) and Elements. Does anyone have any advice or know of any good tutorials that could point me in the direction of a strategy using the software I have?

I stacked 9 picture files (Raw) and 5 darks of Pinwheel Galaxy, taken two nights ago. (5 min, 1600 iso)

I wanted to attach the dss output but the file is too large, so I attach a JPEG of the autosave file. I also attach a crop of the file after I have tried to process it  ( used mainly levels, denoise, brightness/contrast, sharpening).

Before:-

And this is after my efforts:-

So any advice or help would be much appreciated.

Tim. 

One other thing, I get a circular area in all my photos. in the first one here it is mainly illustrated by the darkened edges. Does anyone know what this is or how to prevent it? I did use a light pollution filter in for this image that has reduced it a bit, but still annoying!

[michael.h.f.wilkinson]

The darkened edges is just vignetting which is normal. Flats can take care of that.

[kalasinman]

You simply need more data more than anything else. I don't know your camera, but most DSLR like ISO 800, then increase exposure times till histogram is peaked at 25-35% of scale. I believe for objects like this total integration times of hours are the norm.

[D4N]

Take care what program you perform your post processing in. If it is unable to handle 16bit per channel you will lose data when you stretch the image.

Gimp 2.8 for example is only 8 bit and your camera likely captures in 12 or 14 bit.

Gimp 2.9 is 16 bit per channel.

Paint.net is also only 8 bit per channel.

Sent from my iPad using Tapatalk

[Freddie]

You make no mention of curves. That is what you need to use to bring out the faint detail whilst keeping the bright and dark parts under control.

[Scott]

I'm the last one to advice on processing technics, but as mentioned earlier, more data is a big issue. I've also noticed that the last couple of images I've seen of yours (this and the iris neb) are both quite tricky objects to image. Can I repectfully suggest that you turn your attention to cygnus where there are some great nebulae to be imaged. The north america neb or the pelican neb are both great targets oozing with nebulosity and will show more detail per sub (of equal length/iso) than galaxies and the fainter nebs. sorry if I sound condescending...just trying to help .

Good luck and soldier on...if it was easy, even them visual guys would be doing it   

[StargeezerTim]

You make no mention of curves. That is what you need to use to bring out the faint detail whilst keeping the bright and dark parts under control.

I'm not good on curves! I've seen good tutorials of curves on photoshop, using layer masks to be very selective but I don't have curves on elements. I'll have a longer play with them on GIMP.

[StargeezerTim]

I'm the last one to advice on processing technics, but as mentioned earlier, more data is a big issue. I've also noticed that the last couple of images I've seen of yours (this and the iris neb) are both quite tricky objects to image. Can I repectfully suggest that you turn your attention to cygnus where there are some great nebulae to be imaged. The north america neb or the pelican neb are both great targets oozing with nebulosity and will show more detail per sub (of equal length/iso) than galaxies and the fainter nebs. sorry if I sound condescending...just trying to help .

Good luck and soldier on...if it was easy, even them visual guys would be doing it   

Thanks..One problem I have is from my garden, I can only get views of north and south, and even then, not low down in the sky. At least I can polar align, but I suppose it does limit what I can image at any one time... 

[Scott]

Thanks..One problem I have is from my garden, I can only get views of north and south, and even then, not low down in the sky. At least I can polar align, but I suppose it does limit what I can image at any one time... 

Ah I see, that makes things a bit harder, especially at this time of year.Perhaps the globular cluster M13 in the south or possibly M51 in the north.

[Freddie]

I'm not good on curves! I've seen good tutorials of curves on photoshop, using layer masks to be very selective but I don't have curves on elements. I'll have a longer play with them on GIMP.

In elements, curves are under enhance, adjust colour, adjust colour curves. Not a good title as it works on mono as well !!

Even with more data, you are not going to pull out the faint stuff unless you get to grips with curves so well worth a further try. Just keep it simple to begin with and don't bother with layers, just get to grips with how to control the image with a combination or levels and curves.

[kalasinman]

M81-82 are due North

[lensman57]

You really need a lot more data. There is just enough data as you have it to reveal the core only and as it is there is not enough data to stretch out and if you attempt it the result will just be noise. Try 20+  of 600s exposure subs and that is the starting point. Last year I did 15 subs of the same target @600s each with an Atik 428 EXC and an 8" Newt @ F3.6 and I just dumped the end stack as it couldn't stand to being stretched.

A.G

[StargeezerTim]

You really need a lot more data. There is just enough data as you have it to reveal the core only and as it is there is not enough data to stretch out and if you attempt it the result will just be noise. Try 20+  of 600s exposure subs and that is the starting point. Last year I did 15 subs of the same target @600s each with an Atik 428 EXC and an 8" Newt @ F3.6 and I just dumped the end stack as it couldn't stand to being stretched.

A.G

Thanks... Sorry to be really thick, but you talked about 'stretching' data. What does this mean?

[D4N]

In an unprocessed astro image the vast majority of the data is very dark (background and nebulosity) with a very small portion very bright (the stars). When you process you are stretching the data so instead of it all being very dark it covers a much larger tonal range but whilst you do this you don't want to make the stars too bright or make the background either too bright or too dark.

Curves are great for this as they allow you to increase specific areas without overly affecting others.

Levels tend to effect the whole image but are also very useful when combine with curves.

Sent from my iPad using Tapatalk

[lensman57]

Thanks... Sorry to be really thick, but you talked about 'stretching' data. What does this mean?

The data as outputted after stacking has a tremendous dynamic range that is well beyond the capability of even the best monitors to display, it is as you might say heavily compressed and like D4N has said if you look at the histogram the whole data is bunched up to the left. By stretching either using levels and curves in PS or other methods in the more AP specific software ( Maxim, Pixinsight, Startools etc, )  you are actually pulling the data apart so that it can be seen on the monitor. That is why it is called stretching. As you stretch the data you  also stretch the noise that is in the data but unless you have a lot of good data compared to  noise or in another word a high S/N ratio, Signal to Noise, what you end up with is a very noisy image. The only way to have a high S/N ratio is to image with as long a sub exposure length that is possible ( this is set up and local condition dependant ) and as many  of them as you can gather.

A.G

[StargeezerTim]

The data as outputted after stacking has a tremendous dynamic range that is well beyond the capability of even the best monitors to display, it is as you might say heavily compressed and like D4N has said if you look at the histogram the whole data is bunched up to the left. By stretching either using levels and curves in PS or other methods in the more AP specific software ( Maxim, Pixinsight, Startools etc, )  you are actually pulling the data apart so that it can be seen on the monitor. That is why it is called stretching. As you stretch the data you  also stretch the noise that is in the data but unless you have a lot of good data compared to  noise or in another word a high S/N ratio, Signal to Noise, what you end up with is a very noisy image. The only way to have a high S/N ratio is to image with as long a sub exposure length that is possible ( this is set up and local condition dependant ) and as many  of them as you can gather.

A.G

Thank you...

[DaveS]

I've also found a CLS filter better than UHC for cutting the LP in broadband objects like galaxies as it lets through more useful light while still blocking the various sodium wavelengths.

The last time I imaged M101 was a couple of weeks ago when I gathered 2 hours of data in astronomical darkness using 5 min subs. Just something to build on for later, rather than a finished picture.