HMSTG195 Work In Progress. + A sidonio.. (Already)

[AlexN]

Just for something a little different

This nowhere near finished, however I think (With my "I've been up for 29 hours" eyes) that its worth posting the current progress.. I think if I were to double the exposure time I might get a bit more detail and be able to pull a little more out.. especially with regards to colour..

Details:
Captured 25km from Brisbane CBD

Field centered around :
RA: 9hrs 37min
DEC: -49° 45'

Center of this FOV lies 9° 7" SE of Gum17 in Vela.

Image Scale : 2.32 Arc Seconds per Pixel
Field Of View : 96.7 x 128.5 Arc Minutes

Acquisition details.
LRGB (120:15:15:15)
All 5 minute subs, RGB binned 2x2..
15x5m darks (B1 and B2)
20 flats (B1 and B2)

Equipment:
TMB/LOMO 80mm F/6
ST8300M @ -10C
FLI CFW 2-7
Astrodon I series LRGB filters
HEQ5 Pro
QHY5 + OAG Guiding.

50% Resolution - 800kb
Small image - 195kb

Reprocess using Greg Bradley's information regarding Deconvolution.

50% resolution
Small image

A thank you goes out to Tom Davis, for both the inspiration to chase dust, and his helpful guide to processing widefield images.

Cheers for taking a look..
Alex.

[strongmanmike (Michael)]

Well we have people doing Sidonio's (easy peasey things to do) and now we have you doing a Davis... where wil it end?...well done, some big dusty shoes to fill there though mate That area is full of all the most amazing stuff in the sky I recon.

29hrs is a long exposure for your eyes

Mike

[AlexN]

Big shoes indeed Mike... However there is no point drooling over images of dust and wondering if you could do it yourself... Its better to get out under the stars, go against the odds and give it a go.. Considering battling clouds and light pollution, I tend to think this image came up Ok.... It seems to have lost some star colour when I converted it to JPG will have to have a fiddle..

Vela has about 10 targets I want to image this year... some I can do from home, some I'll have to do when I'm out at dark skies... Looking forward to the challenge though!

30 hours now, I think I need a shutter and an RBI flushing system...

[multiweb (Marc)]

Excellent shot Tom, err.... Alex!

[TrevorW]

Hey Alex theres are dark spot on the image that could do with PS bandaid touchup

Oh it's meant to be there you say

Sorry then, silly me, well done then!!!!

[gregbradley]

Good one there Alex. You are accumulating some nice gear there.

I doubt the RBI flush is that important despite the remarks of some.

I haven't had any problem or bad artifact from my FLI 8300 at all. Its super clean. Then again you need Maxim DL to run the RBI flush.

What you need here is some deconvolution work.

I'd recommend this:
1. On the luminance open in CCDStack and click on deconvolution.
2. Set iterations to 40 and use positive constraint (not sure about max entropy when its used, I am sure there is a criteria about when its best).
3. Click on a few stars near the centre of the image until you find one that is lowest FWHM. Now I could have this wrong perhaps you need one with the worst FWHM as deconvolution means deblurring. Check your theory on this point but that is what I have been doing. A few stars will be much lower FWHM than the others. Usually just off centre.

4. Do the deconvolution. It is a slow process but at least the 8300 files aren't as big as the 16803 files.
5. Save it.
6. Now do it again and repeat until you get maximum benefit. That may be 3 times though. If you get oversharpened stars go back to one of the earlier versions - hence the step of saving each time through.

Do the same on each of RGB.

Now do your final alignment and colour combine. You'll have way sharper stars and less bloat and more detail on bright objects.

I think this sort of routine is done by guys like Johannes Schedler whose images always have tiny pinprick stars. No scope does stars like that that I have seen. So its tight stars in the imaging from excellent careful focus, good seeing of course and then deconvolution. Add to that well engineered filters that don't bloat and you should achieve some nice tight stars.
I would expect the stars to be tighter with a Lomo 80/600mm triplet though as it is very widefield and the stars should look like sand as in an FSQ106.

I had a Lomo 80/480 triplet WO and apart from focuser not being square causing bad seagulls at first until shimmed the lens itself was pretty awesome.

Did you get tight focus with an electronic focuser?

Greg.

[gregbradley]

Also is there some reason you chose 5 minute subs? I don't think this is optimum for your chip.

There is an exposure calculator where you can calculate optimum exposure times for different cameras. For example an ST2000XM was around 10 minutes.

Here is one caclulator, there are others, I think perhaps on the SBIG site:

http://starizona.com/acb/ccd/calc_ideal.aspx

I use 10 minutes and that works very well except for bright objects.
15 minutes may be too long for this chip as it has small wells and I think that translates into bright stars spilling over when too long an exposure and makes them look bloated.

Unless you are having tracking problems I'd try 10 minutes.

Colour should be there already and not a matter of more time as well.

How are you processing the stars? There are ways to bring out the stars separately from the main object. Involved and I tend not to do it but there are techniques to create a layer with the stars only so you can process them separately. Or simply isolate the stars using the colour range tool and then increase saturation and use curves etc.

Or even easier use the Carboni action a few times (works a treat and nice and easy for the lazy processer). His reduce star size works fairly well to but it is based on the minimum filter which is a very limited tool and deconvolution is way superior.

Greg.

[Hagar (Doug)]

Nice clean image Alex but to fill the vacumn bag you'll need lots more.
Should be nice when completed.

[David Fitz-Henr]

Looks good to me - great image Alex ! Love those dark nebulae, and nice subject with some good detail in the nebula itself and contrasting against a rich star field.

[DavidU (Dave)]

Top effort Lex ! You sure put the work in.

[Omaroo (Chris Malikoff)]

Yep - top effort indeed there Alex Every time you put another up I feel a pang of jealousy because I believe that mine's still on its way from the USA.

Thanks also to Greg for the great advice. Looks like CCDStack is in the cards. Looks like I've got to retain my PC for that though. Yuck...

[AlexN]

Thanks everyone...

Greg - I agree... RBI flush is not needed with the 8300 sensor... But my eyes needed it this morning after I'd been awake for 30 hours!! Vision was going a little weird.. haha

Thanks for the tips on deconvolution... I didn't run any deconvolution on this image.. I'd always figured deconvolution was more suited to oversampled and long focal length images... I'll give that a go and see where it takes me...

As for focus, the average FWHM was 1.9, (the lum being slightly tighter on average at about 1.1) however I was shooting between clouds and there may have been some high clouds through some of the frames that could have bloated stars a little I suppose.. Agreed, the 80/480 lens is capable of producing tiny tiny little stars, however I've only managed to keep them tiny during processing a few times... when it happens the stars are pretty amazing..

Reason for 5 minute subs = I live 25km north of Brisbane CBD, and I was imaging to the south, into the worst sky glow visible from my location... The setup is capable of 30 minute subs from dark skies, however at home I've set myself a bit of a limit at 5min.. I will experiment though.. with the LPS filter it may be possible to go out to 10 minutes. Another limiting factor would be clouds.. they were in and out during the night... I would not have accumulated much data at all with 10min subs last night.

Im not doing too much in the way of processing the stars separately, I reduce them a little in the L R G B frames before combine and I give them a saturation boost right before a final s curve... appart from that, nothing special...

Thanks for your thoughts here Greg, you've been quite helpful.. I'll give Deconv a try today and see where it gets me...


Doug - Cheers, Yeah agreed.. another 2hrs lum and another 45 mins per colour channel should see this one through to the finish line... Hopefully I'll get it in tonight or tomorrow night.. Just gotta wait for that pesky moon to go away... (I still say every astro imager on the planet should put in $100 and see if we can't pool enough money to get nasa to fly up and cover the moon in adhesive black velvet..... its just such a pain)

David -Thanks, I too have a bit of a love for both dark and reflection nebulae.. Not only are they pretty awesome when imaged well, but take a look around online... There are literally hundreds if not thousands of relatively large targets (to fill my massive FOV) that are so rarely imaged... I did a google search on HMSTG195 and came up with nothing.. Not a single solitary image of this dark neb to be found.... This is my new playground

DavidU - Cheers Buddy.. I do spend a bit of time at my scope.... Quite literally ever clear night of the year providing I dont have to work the next day.. Although sometimes I break that rule and go to work on 2hrs sleep.... It hurts... but its worth it..

[AlexN]

Chris - Yours will be here soon enough Chris.. Prepare to be dazzled!!

I've just downloaded the CCDStack demo and I'm running deconvolution over the image as I type this.. I'll post up a comparison shot once its all done.. MaximDL's Deconvolution is a horrible experience..

[telecasterguru (Frank)]

Alex,

Glad to see that you have gone over to the dark side.

Fabulous image.

Frank

[jjjnettie (Jeanette)]

Lots of nice subtle detail in the dust cloud Alex. Beaut work.

[AlexN]

Thanks Frank... The dark side is very very appealing!

JJJ - Thanks.. Glad to have got something out of last nights effort

Theres a repro in the oven... Greg Bradley ought to get a medal for his post regarding deconvolution.. The stars are --TINY!--
Still a little way to go yet... Got to work on the channel alignment a bit more, but the sharpness of the image after running deconvolution is amazing...

[AlexN]

Repro is cooked... Thanks to Greg, the stars are looking a lot more like they should.. More data will see this one out though... Theres quite a bit of selective noise reduction used in the dust as I really pushed the data to bring it out, and unfortunately, the data simply wasn't adequate... Monday night I'm at this one again chasing more Lum and a few more colour frames.

Large Repro (50% total image resolution)
Small Repro

[gregbradley]

Wow, that did make quite a difference.

Another little trick is to rub the sponge tool (set to saturate and 5%)
over the brighter coloured stars. This brings them up.

Cheers,

Greg.

[Paul Haese]

Nice image Alex. Heaps of detail in the dark nebula. Looking the goods.

[AlexN]

Thanks Again Greg. I'm going to start keeping a word document of all the processing tips I get.... Every one is so useful, and sometimes it can be hard to remember them all... I got the trial version of CCDStack to run deconv, and I re-registered/stacked/aligned and combined all the data.. It seems to have made a great deal of difference.. Despite its cost, It definitely seems worth it purely based on these results...

Paul - Thanks mate... For a work in progress its well on its way...