A bit of an experiment to use a higher res Ha from a bigger scope - in this case a 12.5 RCOS with my humble Tak Sky 90 for RGB to get up a bit closer without too much expense. As I said a bit of an experiment and star colours among other things not right. But see what you think.
Craig
RGB
15 X 7 min ISO 800
8 X 7 min darks
Hutech Modded Canon 450 with LPS V3 filter
Tak Sky 90 @ F4.5 400mm
Ha
8 X 10 mins cal
Lightbuckets 12.5 RCOS @ 2857mm F9
Last edited by Craig_L; 07-07-2009 at 06:22 PM.
Reason: Update of image
Great image Craig - there's a wealth of texture and detail in the pic.
Might be a little "overcooked" (word of the day!!) in the highlights and shadows of the nebular detail to the left of the pillars, but that's minor and addressable.
This experiment has been a success!!
Doug
Not knowing much about this type of blending myself but just on face value the image has loads of detail but just looks a little sharp and maybe a bit over cooked. I would probably look at toning the Ha for red channel down a little bit.
Great image Craig - there's a wealth of texture and detail in the pic.
Might be a little "overcooked" (word of the day!!) in the highlights and shadows of the nebular detail to the left of the pillars, but that's minor and addressable.
This experiment has been a success!!
Doug
Thanks Doug, normally like my meat medium rare but know what you mean. Struggling with a few other issues and lost sight of the tonal balance. Might try some re-processing.
Not knowing much about this type of blending myself but just on face value the image has loads of detail but just looks a little sharp and maybe a bit over cooked. I would probably look at toning the Ha for red channel down a little bit.
Nice detail so to me a nice image.
Thanks Doug. I must get my Ha Baader filter and do my own Ha like your Eta. Hard work though with a DSLR.
...I must get my Ha Baader filter and do my own Ha...Hard work though with a DSLR.
Craig
Sorry to go off topic but...
I've developed a simple little workflow with the Synscan controller for framing and focusing my Ha shots. For me, it would be a nightmare with the DSLR as you point out Craig.
1. Find and frame object to be imaged - select "User" on Synscan keypad, save as Object #1.
2. GoTo a nearby bright star, centre and save as Object #2.
3. I then insert the Ha filter into the imaging train - bright star will still allow precise focusing - Focus!
4. Go back to Object #1 using User control menu to return to the object you wish to image secure in the knowledge of decent focus and framing.
If focus goes out over the session you can return to the bright star (Obj #2) to refocus then go back to Obj #1 etc etc.
I find this helps me out as I can't see a damned thing through the DSLR with the Ha filter installed.
id back off the intensity of the Ha a tiny bit, the detail is impressive, the core just appears a little burnt out, perhaps run the stars through some sharpening to crispen them up a bit plus alittle in the curves to brighten the dimmer ones.
has a lot of potential if done just right.
Sounds like a good idea Doug. I'll try it next. Might save me a lot of time focussing.
By the way, with the Lobster, are you using Nebulousity to combine the RGB and Ha?
It's the only way i can be assured I'm in the right spot as GoTo can be hit and miss often.
Lobster - I used Nebulosity to align the separate Ha and RGB images but combined in Photoshop. I may need to get some Nebulosity lessons from you soon!
A good experiment Craig. I went down the same road two years ago producing a similar composite to what you've portrayed using a long focal length for luminance and fast wide field data for RGB. You'll find a few examples on my site such as M27 RGB, M83 and my first attempt the Cat's Paw Neb.
Its not too difficult to match up, but you guys trying to match the rich contrasty nature of Ha data can present its problems. You'll find you will obtain success or at least a better result if the approach is iterative. i.e. don't go straight for an LRGB blend, but an LLRGB. Downsizing the high resolution luminance to match the RGB to create superRGB layer, then upscaling it again to match the original luminance will assist. It gets harder to match as the differences between the two focal lengths increase. If the two data sets are at a similar focal length, the better the result.
I've since ceased experiments to progress the technique as I found I began to cross the line with what I felt was ethical. PS is an extremely powerful tool in the right hands - when you start to get out the paint brush to manipulate star colours such as diffraction spikes and other features I felt I was taking it too far. Each to their own I guess. I just wasn't happy about it. Nothing can compare to capturing RGB data on the same instrument the luminance was acquired.
I like what you've put together. Perhaps a little too saturated, but Ha blends take sometime to get right. I still struggle with them. Keep at it if this is the desired path you're keen to investigate. Also, make sure you're not downloading the precalibrated light frames from Lightbuckets - make sure you download the RAW FITS and perform calibration manually with the correct Flat based on PA.
A good experiment Craig. I went down the same road two years ago producing a similar composite to what you've portrayed using a long focal length for luminance and fast wide field data for RGB. You'll find a few examples on my site such as M27 RGB, M83 and my first attempt the Cat's Paw Neb...
Keep at it if this is the desired path you're keen to investigate. Also, make sure you're not downloading the precalibrated light frames from Lightbuckets - make sure you download the RAW FITS and perform calibration manually with the correct Flat based on PA.
Cheers
Thanks for your thoughtful advice Jase. I can see the long road ahead but am not sure the will is up to it. Your Cats Paw, M27 and M83 are certainly examples to aspire to.
I am still waiting for a flat, so used the calibrated frames in the meantime.
Might practice a little more with the data given your comments and see if I can improve the processing.
Just thought I would do a bit of a reprocess before going onto to something else. Not much fiddling but I did increase the size of the cropped RGB image to the size of the Ha before combining for a L(Ha)RGB. And less time in the oven.
Craig, nice, but as Jase says, more time in PS would be appropriate. Zooming in is a bit dissapointing, I suspect the Ha data is a lot better than presented, given the gear. Just look at a recent post Mike made. Could you post a mono Ha image?.
I did increase the size of the cropped RGB image to the size of the Ha before combining for a L(Ha)RGB. And less time in the oven.
Craig
Craig,
As per my previous post, to improve on the result you should try downsizing the luminance to match the RGB first. This is to create a so called "superRGB" layer. Blur and saturate this layer as its technically now the new RGB! You can then upscale this superRGB layer to the full resolution of the luminance again. You'll find the superRGB layer is a much better match than the conventional RGB alone. Give it a try. Don't be afraid to blur the superRGB layer hard to deal with noise. Upscaling any data requires a good quality S/N so keep this in mind. Skimping on the RGB will make it difficult.
Craig,
As per my previous post, to improve on the result you should try downsizing the luminance to match the RGB first. This is to create a so called "superRGB" layer. Blur and saturate this layer as its technically now the new RGB! You can then upscale this superRGB layer to the full resolution of the luminance again. You'll find the superRGB layer is a much better match than the conventional RGB alone. Give it a try. Don't be afraid to blur the superRGB layer hard to deal with noise. Upscaling any data requires a good quality S/N so keep this in mind. Skimping on the RGB will make it difficult.
The Ha data looks on the money IMO.
Thanks again Jase for your time. Got it and will give it a go.
Craig,
....to improve on the result you should try downsizing the luminance to match the RGB first. This is to create a so called "superRGB" layer. Blur and saturate this layer as its technically now the new RGB! You can then upscale this superRGB layer to the full resolution of the luminance again. You'll find the superRGB layer is a much better match than the conventional RGB alone. Give it a try. Don't be afraid to blur the superRGB layer hard to deal with noise. Upscaling any data requires a good quality S/N so keep this in mind. Skimping on the RGB will make it difficult.
The Ha data looks on the money IMO.
Here's a quick new version - downsized the image size of the Ha to the RGB size then blended it in Nebulousity as L(Ha) RGB. On the resulting image I applied a gaussian blur (didn't think it needed anymore saturation) then upsized it to the original size of the Ha and then used Nebulousity again to produce a Ha HaRGB image. Then levels. No colour adjustment.
My apologies for the low res file.
See what you think.
Looks like a furnace. I quite like it. 
.
