Help with processing M31 (is it me, my setup, my workflow or my data that's poor?)

[MikeWilson]

Tonight I thought that I'd have another go at M31 and see if I could improve on my last attempt taken last month (http://stargazerslounge.com/imaging-deep-sky/114179-my-second-attempt-dso-imaging-m13-m31-m39-m57-ic5146.html)

I'm not sure if I'm doing something wrong with the curves tool in Photoshop or if there's something inherently wrong with my setup (or me), but the results I'm getting aren't that much better than my original, despite having what I think is a better set of data to work with.

I took two flats of twilight, fourteen darks and used my library of 20 bias images.

Would it be possible to ask if someone with a few free minutes could check my .tiff file to see if I've gotten either poor data or if I'm just not very good at image processing? (I know I'm not that great, but I followed the same steps with this M31 as I did with my last one and the results aren't as good ...).

Thanks in advance for any advice.

"Screenshot" of the .tiff file (download the file here: http://dl.dropbox.com/u/3717210/M31%20Andromeda%20%2860x90sec%29%20ISO%20800%20-%2011-10-2010.tiff)

My best attempts at image processing this file thus far:

[Uplooker]

Mike,

Unfortunately I am unable to help you with this but i would be very interested to find out what you have done to achieve the results you have.

I have figured out the taking of astro images bit (although needs refining) but for the life of me I can not get the grasp of image processing. After stacking in DSS the .Tiff image I have is very similar to the 1st pic you posted but can get nowhere near your processed pic.

I admit I only have 10 x 60sec and 5 x 90sec @ ISO800 on a Canon 450D, the images nearly appear better on the cameras LCD screen than they do in the .Tiff

I am trying to process in Lightroom 3.

I would be very interested to find out what you have done.

I do apologise if i am hijacking your post.

regards,

Ian

[MikeWilson]

Mike,

Unfortunately I am unable to help you with this but i would be very interested to find out what you have done to achieve the results you have.

I have figured out the taking of astro images bit (although needs refining) but for the life of me I can not get the grasp of image processing. After stacking in DSS the .Tiff image I have is very similar to the 1st pic you posted but can get nowhere near your processed pic.

I admit I only have 10 x 60sec and 5 x 90sec @ ISO800 on a Canon 450D, the images nearly appear better on the cameras LCD screen than they do in the .Tiff

I am trying to process in Lightroom 3.

I would be very interested to find out what you have done.

I do apologise if i am hijacking your post.

regards,

Ian

Hi Ian,

No worries - and I do hope someone can check my data later - I'm sure there's more in there that I could extract without burning it out like that.

Anyway, that's exactly the point. You will need more data and then Google for, "Stretching the Histogram". I'm not familiar with Lightroom, but if it has "Levels" and "Curves" you need to use them in order to bring up the data in your image. I took about an hour and a half worth of data here, which is why I thought I'd get a better result than I did.

Also make sure that you've taken bias images, dark images and flat field images too in order to keep the noise down.

That's the theory anyway, but as you can see - I'm no master at this!

[blinky]

It's not a bad attempt at all. What length and how many subs did you get? If using a DSLR are you stacking RAW of JPG Images? You really dont want to use JPG for anything other than the final web output.

As for suggestions, well I think you have mabe pushed the stretch a little too far and you have some gradient (light pollution possibly?) which could be dealt with by using Noels Actions (google them, great value, little photoshop scripts for astro images) but other than that you have the making of a fine image there!

[MikeWilson]

Hi Blinky,

RAW images. And it's funny that you mention Noels Actions as I'm already Googling for them

Thanks for the feedback, too.

Next time around should I try for more subs of the same or longer subs?

[planetdnb]

I'm kinda at the same level, my m31 looks very similar although not as wide field.

check out this guys page for some Photoshop tricks that gave me some ideas

[removed word] Locke's DSLR Astronomy Workflow

[Beyond_Vision]

You will need to take much more than 2 flats as this will result in you adding more noise to the image. I would suggest stacking the image again without the flats to see how it turns out. It's possible that you may have some of the light frames with clouds in or either bad flats.

Regards

Kevin

[MikeWilson]

Hi Kevin,

Thanks for that. So what would the optimum number of flats be for any number of given light images? (Given that twilight doesn't last all that long). I could only manage two before stars started to appear in my flats. Would two flats really harm the image overall? They seemed to look ok (they show vignetting in the usual areas).

As for the light frames with clouds in, you might be right.

@Planetdnb: Thanks for the link. I'll go check that out.

Also I've just invested in Noels actions / tools for Photoshop. They're pretty neat, especially for adding fake diffraction spikes to stars and making the stars smaller.

[Moonshane]

I know nothing about the subject but find it astonishing that 1) you can go from no experience in two months ago to getting this sort of result with relatively 'basic' kit and 2) get from the first image to the second image!! I see what you mean (and am a perfectionist myself in other forms of photography) but think you should be slapping yourself on the back mate.

[Tim]

Familiar steps Mike

I usually go for 20 x flats, bias and darks to smooth them out.

Data taken one night will vary vastly even from data taken another night, there are so many variables including the moisture content of the air, thin/light cloud cover you cant see, moonlight, temperature and more.

Dont be tempted to push the data further than it looks like it naturally wants to go. The usual advice is, if you can see the difference your changes have made in one go, then you've done it too much

The most useful tools in Noels are Local contrast enhancement, increase star colour, and if you dont have a deconvulve tool, reduce star size.

I'll download your tiff and have a play anyway

Cheers

Tim

[Tim]

With a dslr you can take great flats automatically at twilight, but you do need a white t shirt or similar diffuser over the telescope.

I cant remember the setting for a Canon DSLR just now, but you turn the setting to that, and just take 20 pics, simple as that Someone will know what to turn the dial to.

[Tim]

Just looked at it in Pixinsight, and using the auto stretch function the result was very similar to yours.

How does it look if you leave off the flats?

[MikeWilson]

Thanks TJ and Moonshane.

Reprocessed without flats. Not sure it's any better (in other words - I'm not sure why I bothered taking flats) although I can see noticeable vignetting in the bottom right hand corner.

Thanks for all the tips. I've bought Noels Actions and have been having a good play with that. I'm not sure the data can be stretched out that much more but if I can at least rule out my inexperience, I'll know it has something more to do with my workflow, seeing and equipment