PDA

View Full Version : M-27 H-alpha - RGB



RickJ
2007-Sep-16, 05:34 PM
When the moon was too bright to image much a couple cycles back I tried an H-alpha shot of M27. Even then the moon did some damage but it came out quite well except I had remounted the camera with a set screw loose and it twisted a bit. Only a column down the center was in focus but that's where M27 was! I tried merging it with last years RGB data but the result was not good. I tried again this weekend and must have held my tongue right as this time it went together better. 90 minutes of H-alpha combined with only 15 in each color of RGB all binned 2x2. I used my standard 14" LX200R at F/10 on a Paramount ME and the STL-11000XM

Rick

paul f. campbell
2007-Sep-16, 05:39 PM
Hi Rick.
This is one great photo. The H-alpha filter sure does the j.o.b.

andyschlei
2007-Sep-16, 05:59 PM
Rick,

This is a great image. The detail in the nebula is amazing!

:clap: :clap: :clap:

--Andy

tvdavis
2007-Sep-16, 08:24 PM
Nicely done Rick.

Tom

Bokmakierie
2007-Sep-17, 05:07 AM
Great image, Rick.

Phil

parejkoj
2007-Sep-17, 05:11 AM
That's pretty cool. Lots of wisps around the outside.

It looks like you might be applying a bit too much unsharp masking, though: the stars have dark halos around them, I think. Or am I missing something?

RickJ
2007-Sep-17, 05:54 AM
The halos are due to H-alpha being the Luminosity image. It makes for very tiny stars compared to a true luminosity image. Also blue stars don't register at all. The RGB stars are three times the size of the H-alpha ones yet I only had 15 minutes of that data. Those are what made the halos around the stars and some holes where there are no stars in the H-alpha image (attached). I did no unsharp masking after combining and very little on the main image, mostly to the star field at a level of 50, the nebula was done at 20. There was no deconvolution at all. The H-alpha filter at 6nm seems to create a very sharp high contrast image so I don't need much if any unsharp masking and suffers for it unless done very lightly. I did some contrast and curves after combining but no unsharp masking. Those holes came when I combined the images. That cost me some detail as well. I needed a lot stronger color data to compete with the H-alpha but it was all I have.

If anyone knows a way around them besides stronger color data let me know. I'm still very basic in my processing skills.

Rick

seeker372011
2007-Sep-17, 08:38 AM
oooooh lots of outer nebulosity...well done

Yes it would be nice to find a way to match H alpha star sizes with those in RGB images..wonder if anyone knows of a technique?

maybe one could use Noel Carboni's "make stars smaller" action but when I tried it the results were not all that great

parejkoj
2007-Sep-17, 02:43 PM
Huh... I stand corrected. It makes sense that the stars don't show up in h-alpha, though it certainly looks odd.

One way might be to apply a gaussian smoothing filter to the H-alpha image, but that will reduce the resolution in the nebula.

RickJ
2007-Sep-17, 04:59 PM
oooooh lots of outer nebulosity...well done

Yes it would be nice to find a way to match H alpha star sizes with those in RGB images..wonder if anyone knows of a technique?

maybe one could use Noel Carboni's "make stars smaller" action but when I tried it the results were not all that great

I tried that routine too, I haven't found it particularly useful for anything as it kills all around the star as well as the star's glare. If it would determine the glare from a star against dark sky then subtract that it would leave the nebulosity. I'm not good enough yet to try that as yet.

His technique for removing halos from over used unsharp masking does work well however. Still, if I see halos developing I back off the Unsharp masking as I'm likely making false detail anyway. But when I tried it on these holes it did nothing at all. It neither helped nor hurt, a big nothing.

Rick

JAICOA
2007-Sep-18, 01:55 AM
A Beautiful shot it is Rick!, Well done. On the processing in my case with my type camera in which it will bloom easilly i sacrifice the star size a little and combine my small amount of luminance 15-20min and combine with my Ha to bring out the stars that showed too small or not at all. and eliminate the small rings on the stars with a lighten background on it. Clear Skies

RickJ
2007-Sep-18, 04:13 AM
A Beautiful shot it is Rick!, Well done. On the processing in my case with my type camera in which it will bloom easilly i sacrifice the star size a little and combine my small amount of luminance 15-20min and combine with my Ha to bring out the stars that showed too small or not at all. and eliminate the small rings on the stars with a lighten background on it. Clear Skies

I tried that but the ratio of color data to H-alpha was way to low. So to get any color at all the stars were horrid in size in the RGB image. Including them to any extent in the L image just hid H-alpha detail to a surprising extent. I decided to go with the holes and the detail.

I knew the color was weak, but with constant clouds of late and a problem with the camera (now fixed) I wasn't getting any more data so went with what I had. When I get a more balanced color series I'll try again using the blend you suggest.

Rick

parejkoj
2007-Sep-18, 04:30 AM
Hmmm... I don't know if this method is known to amateur observers, but it probably should be: Lupton et al. (2004) (http://adsabs.harvard.edu/abs/2004PASP..116..133L) describe the arcsinh color stacking method that is used in SDSS to produce more "realistic" colors, with both good contrast at the dim end, and non-saturated colors at the bright end.

Might be worth a try, if there is any common software that supports it.

RickJ
2007-Sep-18, 06:12 AM
Hmmm... I don't know if this method is known to amateur observers, but it probably should be: Lupton et al. (2004) (http://adsabs.harvard.edu/abs/2004PASP..116..133L) describe the arcsinh color stacking method that is used in SDSS to produce more "realistic" colors, with both good contrast at the dim end, and non-saturated colors at the bright end.

Might be worth a try, if there is any common software that supports it.

Arcsinh is one of the stretch methods available in FITS Liberator. A free NASA program many, including me, use to move FITS images into Photoshop. It has many other curves. Sometimes I find one, including arcsinh useful. But most of the time I prefer to design my own stretching routine using the Curves fuction in Photoshop. No two objects have the same light range nor is the detail you wish to show in the same part of the light curve every time or even most of the time. So I usually move the image into Photoshop without using one of the predefined stretch methods. I then design my own for that object using Curves in Photoshop. It usually takes many iterations. Doing it slowly, one step at a time works far better than trying to do it in one or two steps I find. At least at my stage of learning this art.

Predefined curves like arcsinh were handy when I first started out as I had no idea how to read a raw image and come up with a stretch that was any good. But with experience you learn and soon can design your own curves for a particular object that is better suited to the predefined ones. With any home brew curve in photoshop it is quite easy to be sure you don't saturate a star too much to totally wash out color. Still, in a textbook you'll find Betelgeuse is a red star but in reality is is really a orange star so you can carry this way too far if not careful.

It's very tempting (I yield too much of the time) to push the star colors past reality. In a scope stars rarely show much color unless two very different ones are side by side as in the case of Alberio, Iota Cancer etc. Carbon stars alone are very red in a few cases. V Aquila is my favorite. I often incorrectly describe it as "the stoplight of the heavens." Still I'd not want it brilliant red in a photo. Designing your own curves to fit the exposure and object seems to me better than any predefined curve, including arcsinh, at least the version in FITS Liberator.

Rick