This time I have really bitten off more than I can chew.
12 hours of data taken over 3 nights this week and this object is still far too noisy. I've had to push the processing and the gradient removal extremely hard just to get anything half reasonable.
As usual, modded Canon 350D on Tak Epsilon 180ED with IDA LPS-P2 filter. 144 x 5 minute subs at ISO 800.
I don't know whether to shoot more data or just abandon it :(
(http://www.markshelley.co.uk/Astronomy/2014/cepheus_ghost_21082014_small.jpg)
Bigger nightmare here:
http://www.markshelley.co.uk/Astronomy/2014/cepheus_ghost_21082014.jpg
Nightmare - yea right !
What do you expect ? It's a difficult target and it looks to me like you've done an excellent job of pulling it out of the background. If I was to criticise anything then it would be that the brighter stars seem to me a little too saturated.
Excellent image if you ask me, especially the larger version !
Robert
Yes there is a lot of info you have dragged out from the background Mark,. i can imagine how difficult it was to process that
I think it's a very good image, and to get all that dust with a DSLR is particularly good. I did this image at Olly's last year with nice dark skies and a CCD camera.
Can't wait to see what you can achieve with a mono CCD camera which you mention is on the cards at some point.
Carole
Carole, I'd forgotten you had imaged it:
http://forum.orpington-astronomy.org.uk/index.php?topic=9305
My results are pretty similar to yours though my noise is more multicoloured and seems a bit too green.
Maybe I shouldn't be so hard on myself!
I've also got an image of the Iris somewhere so maybe I can join them together.
Mark
QuoteI've also got an image of the Iris somewhere so maybe I can join them together.
Good idea Damian ;)
Although I do like the Ghost
QuoteI've also got an image of the Iris somewhere so maybe I can join them together.
It will be sods law there will be a gap between the two but hope it works for you.
Carole
These guys have just posted on SGL the best version I have ever seen:
http://www.jeffreyjongmans.nl/photo.php?id=55
Is it their setup at Olly's?
In any case it is 34 hours of data in 12 sessions. Stunning.
Mark
QuoteIs it their setup at Olly's?
Yes it looks like it from the information and link on his website.
when he refers to "we" I wonder who the others are in his "group".
I knew of a chap called Yves who has a remote set up there and also Per, I didn't know there was a 3rd remote set up unless he is sharing with Per.
A stunning image.
Carole
Yes Ollies
That's one of the remote stations there on the bit at the top. The big 12" DK has now been replaced with a dual setup as well, the old camera there has also been changed.
Mark - Just checked this out on my 27 inch monitor & have to say it looks very good for such a difficult object. You've captured loads of detail. If anything I would say overall brightness looks a 'tad' too high for me (personal taste) but I understand that this helps with showing off all the feint brown background stuff. Also, still don't fully understand why this is know as 'ghost' Neb..
Very nice though - John
Quote from: RobertM
If I was to criticise anything then it would be that the brighter stars seem to me a little too saturated.
I disagree. Here is the star colour straight out of the stacked image linearly stretched with a histogram. Look at the 3 brightest stars.
(http://www.markshelley.co.uk/webdisk/starcolour.jpg)
I think the problem is that imagers are not used to seeing proper star colour. Instead, they are used to the star colour resulting from non-linear luminance stretching operations such as PhotoShop curves or Pixinsight screen transfer function autostretch. Both these operations wash out star colour, which has been discussed on previous occasions.
The best way to see real star colour is to do an histogram transformation (which is a linear operation), interactively moving the top slider up and down. It causes the star centre to saturate but allows you to see the true colour around its edges.
In my processing operations I strive to maintain the original star colour as best I can i.e. the ratio R:G:B in the unsaturated region of the star. It is this ratio that truly defines the star colour. PS curve and PI STF stretches wash out this R:G:B ratio making it become more white.
Mark
So here's the re-process. Processed from scatch with less green tint and better control of chrominance noise.
(http://www.markshelley.co.uk/Astronomy/2014/ghost21082014v2small.jpg)
Bigger version here:
http://www.markshelley.co.uk/Astronomy/2014/ghost21082014v2.jpg
I'm actually reasonably happy with this now.
Mark
Thanks for the explanation but it was only the position of those sliders and not the method.
Robert
Quote from: RobertM
Thanks for the explanation but it was only the position of those sliders and not the method.
I'm not sure what you're asking. Once you have set the black level then when you play with the Photoshop "levels" slider or the Pixinsight histogram slider you'll notice that the star colour remains the same, no matter what position of the slider you select. On the hand, as soon as you perform PS curves or Pixinsight stretching, the colour becomes diluted. This is an indication that something dodgy is being done to the data.
The method of maintaining true star colour? Arcsinh transform - which in opinion, is the only proper way to perform stretching. Yes you can simulate this in Photoshop by many iterations of stretching and increasing saturation but all this demonstrates is that the Photoshop method is fundamentally flawed.
Mark
Well, the Ghost does in fact overlap the Iris image I took 4 years ago:
(http://www.markshelley.co.uk/Astronomy/2014/iris_ghost_small.jpg)
But the Iris was only 4 hours of data whereas the Ghost was 12 hours. So the Iris had to be stretched quite a lot.
Bigger version:
http://www.markshelley.co.uk/Astronomy/2014/iris_ghost.jpg
It seems to indicate that I may have inadvertantly neutralised an area of dust in the Ghost image. Collateral damge whilst removing the background gradients :(
Well done Mark.
Carole