Eta Carina - Reprocessed

[David Fitz-Henr]

Well, I reprocessed my Eta Carina image as I wasn't happy with my initial effort. Also, I have read Ron Wodaski's book (New Astro Zone System for Astro Imaging) which has helped a lot.

Small Image:
http://www.pbase.com/image/124324296

Large Image:
http://www.pbase.com/david_fitz_henry/image/124324296/original

[renormalised]

Nice piccie David, but you're going to have to cut down on its size so people can see it without having to scroll all over the place!!:)

[multiweb]

Hey David, this is truly outstanding data. I feel there is a lot more you can extract out of the raws by having a look around the field. What software are you stacking with?

[David Fitz-Henr]

Ummm...yeah, you're right. What would be the recommended size - maybe 1000 x 800 or thereabouts?

>>> Update: I have downsized the large image.


I'm using CCDStack for initial calibration, Registar for alignment, CCDStack for statistical combine and Photoshop CS4 for curves / sharpening / etc. I'm sure you're right about extracting more though; I'm still learning the basics at this stage!

[renormalised]

The resized image is good. Maybe just a tad smaller....it's OK to scroll down a bit, but not across the screen.

[multiweb]

Cool - give this a go and see if it makes any difference.

1_ Find the sub in the Lum that has the smallest FWHM. CCDIS can do that. Make a note of its name. Let's call it sub X.
2_ Register everything to sub X with CCDIS plugin [high precision checked]. Use Nearest Neighbourg as your rego algorithm.
3_ Select sub X again. Push DDP to see the difference between light and dark areas. Normalise chose both: prompt select background, pick dark neb, 2nd prompt select eta and around keyhole as highlight.
4_ Select sub X again. Stack -> Procedures -> data rejection -> Reject Hot/cold pixel [strength10, upper limit(adu) 10000 - tick clear before apply -> Apply to All.
5_ STD sigma reject - tick top image 2% - untick clear before apply -> Apply to All.
6_ Stack -> Combine -> Mean
7_ File -> Remove all images excep this
8_ Process Deconvolve -> Auto-Select Stars -> Accept Selection -> [Positive constraint, number of Iterations 35]
9_ DDP to taste then save as scaled TIFF. Before you save as a scaled TIFF lower the background so you're not clipped in PS.

That should give you a real good Lum to start from. See how you go. :thumbsup:

[Hagar]

I like this one David. As repro go this is a ripper. Must be nice working with good data like this. Colour etc is lovely. Well done.

[David Fitz-Henr]

Thanks Marc - some good pointers in there. I'm away for a couple of weeks with work, but I'll include them in my next repro when I'm back. I did actually do a DDP stretch beforehand with the luminance though, but found that I had to "unstretch" it in PS to get the colour to come through. It seems to me that bright luminance data prevents the colour layer (using colour blend) from giving good saturation? Any suggestion there - is there an upper limit to brightness for the luminance layer (apart from stars)? The colour actually looks better with the luminance turned off, and with the luminance I can't achieve the same vibrance tweaking colour balance, rgb curves, etc?



Thanks Doug - yes I was far more satisfied with this one. I had pumped up the luminance on the Ha in the first attempt, and then found I had to pump up it's red colour to prevent washout so the whole image took on a very red hue. Also I handled the stars differently this time. I had big cyan halos (due to Ha and poor seeing for the blue subs) in the first which were hard to get rid of. This time I shrunk the stars first in the blue data, and replaced the Ha stars with luminance layer to get them all the same. Then I applied the Minimum filter in PS to the stars only. For my next repro I may try Deconv (as per Marc's suggestion).

[richardo]

Hi David, man for your second attempt at RGB imaging, wow, certainly nice work!

I think back many years ago to my first attempts, makes me cringe in comparison to your efforts.
Again, well done.

Rich

[David Fitz-Henr]

Thanks very much Rich, but don't give me too much credit! I did read Ron Wodaski's book first and got a few tips from Marcus (Marc4DarkSkies) as well. You can spend hours at the PC playing around with this stuff - I'm starting to wonder where the astronomy ends and the PC takes over!?!

[richardo]

In many cases with good data, it comes down to- 'Less is best';)
When it looks like a painting, and say a bit plastic, then you've gone too far!

Rich