M31: what went wrong here?

[RobH]

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:

You should never touch the right hand (white point) slider!

What you have done here is to lose data, and have clipped the whitepoiint.

On the graph, there is a thin black line that goea all the way to the right. This is information, and in bringing the slider down to the base of the big data mound, you have clipped all of that data. Whan doing levels adjustments, just adjust the blackpoint but leave the whitepoint well alone. Use 'CURVES' to brighten the image....this way you can brighten the dark areas without clipping the bright ones.

Cheers

Rob

[lukebl]

You should never touch the right hand (white point) slider!

What you have done here is to lose data, and have clipped the whitepoiint.

Well I did say that there were others who know better, and I don't doubt that you are quite right. That was just my 'quick and dirty' way of doing it!

But I did see it on an astro-imaging tutorial (can't find it now, but will try and find it) where the instructions were definitely to slide the right-hand arrow till it met the start of the histogram.

[RobH]

Well I did say that there were others who know better, and I don't doubt that you are quite right. That was just my 'quick and dirty' way of doing it!

But I did see it on an astro-imaging tutorial (can't find it now, but will try and find it) where the instructions were definitely to slide the right-hand arrow till it met the start of the histogram.

It is quick and dirty Luke!!

If you look at the core, you'll see that after using the whitepoint slider like that, you now have a much larger clipped area.

The astro-imaging tutorial that you saw that in is wrong, plain and simple!

Cheers

Rob

[lukebl]

The astro-imaging tutorial that you saw that in is wrong, plain and simple!

Cheers

Rob

Thanks Rob. I have been helped! My apologies for the duff advice.

[RobH]

It's exactly how I used to stretch images Luke, and I always ended up with no star colour as I'd inadvertantly clipped them in all 3 colour channels, meaning that they end up white.

What you have to be aware of when looking at the histogram, is that even a thin black line is data. If you move either the black, or white point into this line, you are losing data.

I've just finished a big section about using levels and curves which I'll be adding to my website over the next few days, which will hopefully explain things a bit better.

Cheers

Rob

[yesyes]

The autosave.tif doesn't work. It loads as a completely transparent image.

Do you have a saved TIFF file in 8 or 16 bit that we could play with?

[siovene]

yesyes, I could upload it later. Or you can open the 32bit tif with DSS and convert it to 16bit tiff.

[yesyes]

I'm at work at the moment. Don't have any of this fancy stuff on my work PC. ;-)

Will have a go when I get home...

[Luke]

It's exactly how I used to stretch images Luke, and I always ended up with no star colour as I'd inadvertantly clipped them in all 3 colour channels, meaning that they end up white.

Ah nice tip, I was thinking that doing this a bit (dragging the right point to the left a bit) would be okay if you didn't mind some areas overexposed, like a galaxy core. Didn't think about loss of colour!

[yesyes]

I've had a go at your data. Downloaded the tiff, loaded it into DSS and saved it as 16bit TIFF. Then played with levels and curves in Photoshop. I also made sure I did not move the white side slider (thanks Rob, I didn't know that either). Then a bit of a colour balance tweak.

I'm no expert at this either. So far I have only taken 2 DSO images myself. So others might still be able to get more out of your data.

But I would say it is not bad for a first image (assuming this is your first?). There is some weird halo around M31 though. I'm guessing it's from your flats. Are your flats much brighter in the middle than at the edges?

I've attached a jpeg version of my result. The TIFF version of the same image can be had here:

http://yesyesserver.co.uk/astro/siovene_m31_2.tif (70MB)

[Tim]

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.

That red sky background is all to familiar

I find it always looks (and photgraphs) worse on nights with poor transparency. The skies here in the Midlands lately have been appalling, even when it has been clear there has been loads of muck in the air. All that **** does a good job of reflecting streetlights. A decent light pollution filter will help you a bit, but the same length subs on a night with better transparency (and less moon?) would likely be better again.

The red streaks like rain are common with Canons, and it seems especially to come to the fore on M31 images for some reason. If you find an answer to it, do make a big post about it. I didn't find that dithering worked for me.

Cheers

Tim

[MartinB]

I've had a look at the image. The LP has certainly made it's mark but there are some odd areas of discolouration which I think might have been introduced by your flats. There was also vignetting. The orange bias was moderately easy to eliminate and I was able to sharpen up some detail in the dust lanes. If the background had been cleaner it would have been possible to enhance the colour a little more.

As Tim says, it's very difficult to combat this degree of LP.

[mike_777]

I think there is definatley something wrong with the data. I have had a little play bu cant seem to get alot out. With nearly 4 hours of data there should be more in there!

I am using virtually the same equipment and I'm no expert, but this is the result of approx 3 hours of the same length exposures. I used the 2" sky watched LPF.

mike

[Tim]

With all due respect Mike, to compare pics in a meaningful way the data would need to be aquired on the same night at the same time at the same location. Just too many variables which affect the SNR.

To the OP: Could you restack the image without any calibration frames by any chance?

Cheers

Cheers.

[siovene]

Thanks to everyone that helped!

I restacked without the flat darks, and got something reasonable.

Mike: your image is definitely better, but perhaps you had better sky that night?

Here's my final result:

All sizes | M31 | Flickr - Photo Sharing!

I think it leans on the green side, but I'll fix that later.

[yesyes]

Now that is a lot better... So there must be something wrong with your flat frames I think.

The green hue is probably cause by the green channel being offset in the histogram. You might be able to tweak only the green channel to get rid of that.

[mike_777]

Point taken guys, i'm relativley new to the game and dont fully understand the diferent variables when imaging.

regards

mike