NGC6188, a star forming bright nebula in the constellation of Ara with the open cluster NGC6193 providing all the action as it kicks off, along with being the source of the reflection nebula.
You also have a rare type of emission nebular, NGC6164 at bottom right.
Presented vertically as the main object sits in a dense backdrop of stars in the middle of the Milky Way along with some dark nebula and open clusters. Also has a jaunty angle which I quite liked in DSO browser.
3 hours of Ha was combined with the original 2 hrs of RGB data. Luminance added nothing, combined using LRGB combination in PI.
FSQ106n, AP Mach1, STL 6303e. The internal guide chip caused issues so switched over to a Baader Vario finder with an STi ccd which worked really well.
It is indeed a lovely composition, and the underlying data look good.
CC: The stars and the brighter bits of nebulosity are burned out. With that scope, if you use more curve and less linear gain, you should be able to produce pinpoint stars, resolve the main cluster, and get more detail in the bright bits of nebulosity.
Another possibility is to use the Ha as a luminance channel. That would also produce pinpoint stars and improve the nebulosity, excepting you'd lose a small amount of blue. If that was an issue, you can experiment with using Ha as a partially transparent luminance channel, to strike a balance.
Hi Mike
Thanks very much for the CC.
Have looked at some scripts for blending Ha in PI and it gives much better results on stars and reduces burnout, especially when you don’t over tinker with all the other Knobs and whistles
I like this one a lot. It's normally imaged in standard SII, OIII and Ha, but I think that your version with the RGB comes out much better. The reddish star makes a nice eye for the top dragon--you would probably lose that in straight narrowband.
I still had blooming artifacts in the ha data which I had missed, the original was overworked and given Mikes comments on how the FSQ should be able to produce much tighter stars I applied drizzle for the undersampled data then used a masked stretch to go non linear. Solved the star problems and blown out nebula but still getting a few halos.
Instead of using LRGB combine, I used the HARVB script in PI to blend Ha with the red then sat back and did as little as possible. Completely different outcome but I managed to keep the yellow star colour in the dragons eye!
That worked brilliantly from the point of view of tiny Takahashi stars and sharper H-alpha nebulosity, but sadly it did lose a lot of the blue nebulosity.
If you think you can stand it, you could aim for a halfway house. There are many ways to do this.
In Photoshop for example, you can set the H-alpha as a luminance layer, and then adjust the transparency of the luminance layer to produce a best compromise where you've still got plenty of blue but the stars are still small.
Not at all Mike. I value the feedback. Have spent the last hour or looking at how to do this in PI. Will need to dig around some more. Whilst I have photoshop am trying to avoid using it as much as possible.
I need to look at what I did with the RGB to get the blue in the first place then try your method but using Pi .
I REALLY like the framing on that one David! Beautifully composited. I never like looking at the Dragons horizontally - I like vertically or askew like yours.
Cheers Lewis.
DSO Browser is really good for fiddlng around with framing. What I also like is that you can travel around once you have found your target of choice and the site renders the surrounding areas as you go so you pick up stuff you might normally miss.
The repro is awesome David. I'd be stoked with a colour shot like that. Definitely your best to date and right up there in term of processing and acquisition. The way you'be blended the Ha channel and kept your stellar profiles is text book. A definite for the cool wall.
You can blend two images using PixelMath with an expression like "0.25*Image1+0.75*Image2". You can use previews and tweak the blend amounts to taste. This will work for final images, blending Ha with Lum, etc.
You might be able to re-introduce some of the blue by building a ColorMask from the old version, applying it to the new version and then tweaking the colours with CurvesTransformation (like boosting blue, reducing red and/or green, even more exotic stuff with CIE a* and b* curves.)
Adding Ha to the red channel though is old school and a very unforgiving approach.
A way more flexible approach is the Don Goldman Astrodon screen layer method. It allows you to play with different levels of narrowband data and you have ultimate flexibility. Adding to the red channel is a one way street with no later changes possible.
There is a tutorial on how to do it on the astrodon website. I often use other blending methods rather than Screen to do the blending though. Screen often is too strong.
You can also add ha as luminance as a layer on top and sometimes get away with it. It usually wrecks the colour and damages the stars sometimes but turn the layer opacity down to the point where its adding detail but not smashing the colours and stars and you can get an improvement. I don't think it will be the same for 2 images in a row though.
Adding your blue back in would be easy in layers in Photoshop. Set it to lighten blending mode and adjust opacity to suit.
