Sword of Orion - Progress report

[Uranium235]

I had to take full advantage of Wednesdays clear sky to grab more of the sword as it was properly clear for the first time since posting the 50% done version. Im still missing 2 panes from the Ha, but at the next opportunity I will complete those and simultaneously grab another 4 1/2 hours on the 1000d. That would be just one more good night needed - 8.15pm (when it clears the tree), until 12.30pm (when it disappears behind a house). But the forecast is depressingly grim, I hope it clears up soon. In the meantime I might grab some darks in the fridge as havent taken enough to keep up with the lights (only 34 darks).

Processing this has taken most of the evening, so im not going to go the whole hog because I dont have all of the data to hand. I havent also gone as far as put in a proper core for both versions (need sleep!), but it should illustrate what approach im going to take with the processing - I will be treating the DSLR data as if it were LRGB by working on seperate luminance and RGB masters. That way I can get a better grip on the noise in the outer areas, Ive also experimented with selective contrast enhancement (via layer masks).

I'd say its about 70% done, so still the marathon continues..... by the time this is finished, the combined total for all data collected should roughly equal 24 hours... and it will probably take another 24 hours to process!

Sword of Orion:

1000d: 46x600 (L) Darks+Flats

Atik 314L+: 6x600 + 5x900 (Ha) 4 panes

ED80 (0.85) x2, NEQ6

Thanks for looking

First the luminance layer Ive created from the DSLR data, for an uncooled, noisy camera - thats not too shabby. I will work on the colour layer in the final run.

Next is what I have from the CCD so far... obvious as to what is missing

Once its complete I will selectively add it to the DSLR data so not to destroy the detail in some areas. If you study the two, you can see the differences in structure.

[Tom OD]

Thats coming along nicely.

[Mik3]

Wow thats nice.

Is that DSLR luminance just RGB coverted to greyscale with no fancy filters used?

[Gina]

Wonderful stuff there - a real labour of love I salute you

[zakky2k]

Great update! hope you get the clear skies soon to complete the missing data

Cant wait to see the end result!

[Steve 1962]

Thanks for the update - great levels of detail in there. I'll be very interested to see how this pans out - I have a modded 1000d getting dusty on the shelf and have been thinking about a dual scope CCD / DSLR setup

[Uranium235]

Wow thats nice.

Is that DSLR luminance just RGB coverted to greyscale with no fancy filters used?

Yes, all I did was discard the colour information in photoshop (change to 16bit greyscale), then treat it as I would if it were from a CCD. The only filter used on the 1000d was the CCD CLS clip filter. Doing it this way I can work on the contrast without worrying about the colour, the final luminance master will be a blend of DSLR and masked narrowband data. I'd like it to pan out so the CCD data takes over strongest where the DSLR starts to give up the ghost in terms of noise and detail.

Steve, if you get the chance.... use your old camera! This is data I wouldnt otherwise have if I had just stuck to using the CCD alone, a very welcome brucey bonus

[Mik3]

Uranium if you have the time could you post some more details about how you processed it?

This is the level of detail I would also like to achieve. I hope these clouds clear soon so you can finish it.

[Uranium235]

Uranium if you have the time could you post some more details about how you processed it?

This is the level of detail I would also like to achieve. I hope these clouds clear soon so you can finish it.

Give us a few hours and I will do a quick write up (need to go shopping first!). One or two of the methods im using are still experimental, like selective NR and constrast enhancement. I had the idea that layer masks can not only be used for improving the dynamic range, but can be used to blend/apply (or mask certain areas from) contrast and sharpness routines. For instance, I applied contrast enhancement to only bring out the dark stuff in M42 while leaving the rest of the image untouched. It might be old news to some, but to me its the brainwave of the year so far

[Mik3]

I had the idea that layer masks can not only be used for improving the dynamic range, but can be used to blend/apply (or mask certain areas from) contrast and sharpness routines. For instance, I applied contrast enhancement to only bring out the dark stuff in M42 while leaving the rest of the image untouched. It might be old news to some, but to me its the brainwave of the year so far

Thats how I do mine, I heavy sharpen the core to reveal lots of detail, but sharpen the outer areas less.

I have learned a lot about processing the last few months due to these clouds, but haven't had the chance to get much data to really test it yet.

How did you reduce the bright ring around the bright stars? I was thinking about using shorter exposures and blend them in with layer mask and gaussian blur.

[Uranium235]

How did you reduce the bright ring around the bright stars?

That came as a result of the composting between medium, long and very long stretch versions. In the version above it was 5 layers, to get right down to the trapezium with the DSLR it might need 6 - but ive yet to collect any really short subs (the shortest I have is 60s).

[Uranium235]

Right, here we go... To produce the greyscale luminance version I did the following:

1) Load the calibrated and stacked (10min subs) Autosave file into PS (needs to be 16bit)

2) Change to greyscale (discard colour data)

3) Create three different versions, one with all the outer details revealed, another with less, and a third with little adjustment in curves (very long, long, medium)

4) Compost them all into a single image (VL+L, flatten, then add the M, flatten)

5) Do the same again with the shorter subs (ensure they are registered with the 600s master), compost them in until you get down to the core.

6) Flatten the lot, then add a very light touch of NR, strength 4 (preserve details=80)

7) Use noels select brighter stars to create a selection, then expand that by 1 pixel. Then CTRL+J

8) Reselect the background layer, then create a duplicate - set blend mode to overlay.

9) Select the overlay layer, run a high pass filter - setting 4.0

10) Adjust opacity, merge/flatten

11) Repeat step 7, then reselect the background layer then select brighter stars again, expand by 1, then inverse the selection

12) Run smart sharpen at about the setting of 50.

13) Apply a slight curve to the star layer to compensate for the dimming created by sharpening everything except the stars.

14) Merge/flatten

15) Create a copy of your current state and run noels "enhance local contrast" on it

16) Copy + paste this back into your original version - blend mode darken, adjust opaicty to suit

17) To make step 16 more selective, use a layer mask as you would when composting.

18) When done, flatten then one final dose of NR to tidy up the noisy bits (but not M42 or the running man, leave those out of it if poss). But the more data you have, the less you need to tidy up.

All of the above will need to be repeated for the Ha master, and the colour data - though the colour will not need any sharpening (just loads of NR) as its only the colour information im after.

[Uranium235]

Will post the above pic again as I got an error message when trying to post (it says I dont have permission!). This is a merge of the DSLR and NB luminance using a layer mask.

[Mik3]

Fantastic image, looking forward to see the final colour version.

Thanks for the help, now just need some clear skies!

There's some good processing techniques on this website you might be interested in, the screen mask invert is good for increasing faint nebulosity.

[ultranova]

I wish the damn weather would sort itself out so i can see it finshed in all it glory.

Looking vey good.

Paul

[Uranium235]

Fantastic image, looking forward to see the final colour version.

Thanks for the help, now just need some clear skies!

There's some good processing techniques on this website you might be interested in, the screen mask invert is good for increasing faint nebulosity.

Thanks, thats a very interesting link to LLRGB processing, though it mentions deconvolution which for me has never really worked (always makes it look worse!). Everything else is do-able, ive never really tried noels deep space noise reduction before, it might be a cure for any chromatic noise that creeps in from the DSLR.

[Gina]

Wonderful stuff

[Sp@ce_d]

....... ive never really tried noels deep space noise reduction before, it might be a cure for any chromatic noise that creeps in from the DSLR.

I've had some success with that & then using the remove colour blotch action.