M31: what went wrong here?

[siovene]

Hello, I have imaged this M31 earlier this week, but the result is way below my expectations.

I managed 46x5min frames, 12 flats, 12 darks for the flats, 25 offsets.

I didn't take any darks because I thought I'd take them the next day, but then the temperature changed dramatically (over 10 degrees warmer). I thought that having 46 frames in -16°C should eliminate the need for darks.

Anyway, after some minimal processing, I realized that this is an insult to astrophotography:

What a mess! Besides that dust fragment in the top left, I can see lots of noise, gradients and overall a lousy image.

Can anybody help me figure out what went wrong and how to do better next time?

Telescope: SW ED80 on HEQ5

Camera: Canon 450D unmodded

Also, if somebody feels like giving it a shot, by clicking here you can download the Autosafe.tif that Deep Sky Stacker spit out.

Thanks a lot!

[jgs001]

Doesn't appear to be anything too wrong... it looks a bit red...

If you're taking Flat Darks, you don't need Bias/Offset frames and vice versa... so I'd start with losing one of them. Drop the Bias and restack as a starting point.

[siovene]

Nothing too wrong? What about all that noise? And what about that rain effect? (lack of dithering?)

What about the lack of signal? My site is indeed light polluted, but I've seen much better shots with similar equipment even from the cities.

[brantuk]

I wouldn't be too dissapointed with that. Your trackings spot on, the cores a little bright, but you've got dust lanes, and bought out the other galaxy nicely too. I'm no good at processing but I'm sure someone will have ideas on how to bring out the best in your data. Good one I say

[philj]

I agree theres not much wrong with that image that a bit of colour balancing and gradient removal wouldnt sort.

If you havent tried to clean the dust from the camera try doing a few post event flats then try a process with those. Not the best way I admit but its worked fo me in the past when Ive either forgotten or not been able to do to do flats at the time of image acquisition.

Phil

[jgs001]

I can't open the autosave.tif... GIMP can't handle the bit depth.

Restack without the bias frames, the flats are affected by both those and the flat darks, and are probably introducing more noise...

In fact, try stacking just the lights and see how it comes out.

Actually, thinking about it, if you took flats, and are still showing dust, then does that not imply a problem with the flats ??

[lukebl]

I'm no expert at these things, but that looks pretty good to me.

Here's my first attempt at adjusting it in PS (from the jpeg, not the TIF, so the quality isn't good):

[siovene]

Like, pretty please, explain to me how you did it!

[peter shah]

all looks pretty good to me ...just some processing and more frames and it all should be ok....stick with it.....its a very large learning curve

[lukebl]

Like, pretty please, explain to me how you did it!

This was my quick and dirty way of doing it but, believe me, there are people on this forum who are far better at it than me, and will do it much, much more effectively!

STEP 1:

STEP 2:

Click [OK] then STEP 3:

[siovene]

Cheers Luke!

[peter shah]

This was my quick and dirty way of doing it but, believe me, there are people on this forum who are far better at it than me, and will do it much, much more effectively!

STEP 1:

STEP 2:

Click [OK] then STEP 3:

that is one way but you can clip the black point very easily doing it this way.....one thing to take note of is the shape of your histogram, it has two peaks this tells you that there is a colour bias issue(NOT THE SAME AS COLOUR BALANCE) you can select each colour channel in the histogram drop down box, as you scroll through each channel you will notice the histogram will have one peak and jumps forward....all you need to do is to adjust the black point slider in each channel until they are the same. then you can go into the colour balance and adjust the balance.

[lukebl]

that is one way but you can clip the black point very easily doing it this way.....one thing to take note of is the shape of your histogram, it has two peaks this tells you that there is a colour bias issue(NOT THE SAME AS COLOUR BALANCE) you can select each colour channel in the histogram drop down box, as you scroll through each channel you will notice the histogram will have one peak and jumps forward....all you need to do is to adjust the black point slider in each channel until they are the same. then you can go into the colour balance and adjust the balance.

I did say there were others who know far more than me! A dark art indeed. How do you get the individual RGB histograms to match, other than by trial and error as you can't see them at the same time?

[peter shah]

I did say there were others who know far more than me! A dark art indeed. How do you get the individual RGB histograms to match, other than by trial and error as you can't see them at the same time?

a dark art indeed.....your method does work but can give variable results

you can show the histogram in a window and split them to show the different colour channels......so when you do any adjustments it gives you a live view of the histogram...you can then be sure not to clip anything

[Shibby]

siovene, as others have said, processing your image is a real art form and takes a lot of patience and experimentation. I'm also on the learning curve and it's a steep one! However, with 45 subs you're probably justified in presuming there would be a bit less noise. So I'd try restacking with only light frames added. Darks and flats will add noise - the fewer of them there are, the more noise you add.

Either way, I wouldn't be disappointed. The outer areas of M31 are very faint and you've got a nice image there, with perfect framing I might add!

[lukebl]

Is anyone else going to have a go? I'd like to see how it should look, as the data seems to be in there.

[siovene]

Thanks Shibby, that's encouraging!

To Luke and the others: how come the image looks quite monochromatic? I can't see any contrasting colors in the stars, despite there being some stars which should be red and some which should be blue.

[narrowbandpaul]

i agree siovene, with 45x5mins or whatever, you should have an image better than that. You are right to be suspicious that something has gone wrong.

try stacking just the lights...if you have poor quality flats or darks they can add noise to the final image.

since you live in a high LP environment maybe an LPR filter would be of good use.

paul

[siovene]

Thanks Paul. Which LPR filter would you recommend for an unmodded camera?

I'm considering getting a B&W CCD, but that's about a year from now.

[blinky]

The noise in the image I am pretty sure is casued by the flats. I have had similer problems before if the flats are slightly underexposed (I think, it's under exposed)

[siovene]

blinky: how can I tell if a flat is correctly exposed? Today I got a flat foil so I should be able to take them consistently from now on.

[Tim]

Dont worry, all the shots look like that when they first come through

Keep at it, you'll soon learn

[jgs001]

As you're using an SLR, set the camera to Av mode and let it work out the exposure. I normally grab one shot in Av mode, then set to manual and the same shutter speed and capture the rest. The camera puts the peak of the histogram between the 50 and 75% mark, and seems to work very nicely. This is based on the info in the DSS help.

[siovene]

Yeah John, that's exactly what I did...

[MartinB]

Thanks Paul. Which LPR filter would you recommend for an unmodded camera?

I'm considering getting a B&W CCD, but that's about a year from now.

An astronomic CLS CCD clip filter