Help me with my color balance - M8 Lagoon.

[2020BC]

I'm experimenting with a new workflow which has the result of accenting the red data in the Lagoon nebula especially on the outer parts and which reduces the strength of the strong blue signal near the centre.

Would you please comment on the merits of the blue versus red version. Which do you like best ?

The new red version is is here (200KB) (http://www.zodiaclight.com/images/m8Lagoon3xDataCombinedWinter2006Cur vedtoRedWeb1280.jpg).

The previously posted blue version is here (300KB) (http://www.zodiaclight.com/images/m8Lagoon3xDataset100MinsWeb1280.jpg ).

Both images start with the same data.

Subject: M8 Lagoon Nebula
Main Scope: Orion 80ED with crosshairs
Guide Scope: LXD-75 SN-8
Mount: Meade LXD-75
Camera: Canon 350D unmodified
Exposure: 100 mins total (10x3minsISO800 man-guided; + 7X4Mins man-guided; and 14x3minsISO1600 auto-guided)
Process: MaxDSLR & Photoshop & Noiseware

[mickoking]

Both images are great but I prefer the blue one on aesthetics. But what one is closer to the real colour of the Lagoon?

[2020BC]

What a great question, and no one I ask can give me a definitive answer. The best I've heard is "It depends....". The reasons it's hard to get an answer are many.

Nebulae emit light strongly at specific frequencies. Hydrogen H-alpha (red), H-beta and OxygenIII (blue-green), Nitrogen (blue), Sulphur, etc.. Different nebula emit light more strongly at some of these than others. The bright stars within nebulae also contribute a broadband 'white' light component that is dispersed and scattered through any surrounding gas and dust which in turn take on their own hues. All of these are on the 'transmit' side of the equation.

On the 'receiver' part of the equation our cameras each respond to the light in different ways. A stock Canon 350D has its peak sensitivity in the green part of the spectrum and is much more sensitive to blue-green light than it is to the wavelength associated with H-alpha. A Hutech modified Canon 350D is five times more sensitive to H-alpha than a stock camera and has H-alpha response that gets closer to its response for blue-green. Astronomical CCD cameras on the other hand tend to have their peak sensitivity skewed more to H-alpha since most emission nebula emit very strongly at the H-alpha wavelength.

Combine the different strengths of emission lines on the 'transmit' side of the equation with the different responses to those emission lines on the 'receive' side of the question and its very puzzling. Some photos you'll see include an Infra-red or even ultraviolet component neither of which we can see but which combine to create "color" in an image.

So even if there was a standard and known range of colors being emitted the different types of camera would record the color balance differently.

The question of what is the 'real' color is further complicated because our eyes are poor receivers of H-alpha light especially in near-darkness. A stock Canon 350D therefore (because that's the way its been designed) responds to the different frequencies of light in an approximation to how our eye responds (but to daylight!). The low-light spectral response of the eye is different to the bright-light response.

It all gets very confusing, and I find the whole issue of choosing a color balance for imaging very vexing. The closest I've come to a plan to nail exact visual color is to locate what's called G2V main sequence stars (white like our sun [it is not yellow]) and use these as a color reference but it just adds more work, ho hum...

Combine all of the above with skyglow adding a wash of unwanted color into the image and aaaarghh.....:eyepop:

Usually I just set the black point on each color channel to balance the histogram peaks on each color, just accepting that the camera recorded what it recorded and be done with it....:lol:

[Doug]

G'day Bill, your web site seems to be down at the moment.
You might like to experiment with this proceedure in PhotoShop.
Load in the newer verion.
Select Image>>greyscale
Image>>RGB
Image>>Adjustments>>Hue/saturation
Click on the colorize check box and increase the saturation slider to about 50, leave hue at zero and also adjust lightness to about -8.
There is not alot of 'other colour in the Nebula, or in the stars to worry about. This method will kill any other colour, but for this image I believe it should be ok. But you must judge for yourself. It is the easiest way to get rid of the salmon pink in the center, yet give a natural red, overall.

Here is an LRGB image acquired with an STL 11000 and 32"RCOS. The colour is a bit sttrong , but there was no grayscaling or desaturation used ...there is just no other natural colours there.

Cheers,
Doug

[mickoking]

Apparently David Malin's images have been processed to achieve natural colour. But the million dollar question is, what are the real colours of many astronomical objects?

[Doug]

Sorry Bill, no matter how many times I proof read.............
The first to instructions should read:
Image>>Mode>>greyscale
Image>>Mode>>RGB
the rest is ok.
cheers,
Doug

[2020BC]

I think the 'natural' nebula color could only be observed with our own eyes and only if we were close enough to the object such that its light was bright enough to stimulate a daytime color response in our retinas.

The only alternative to being close enough is to use a camera to collect enough light and have the color balance in the camera match the spectral response of your eye. This is a tough one given the vagaries of collecting enough light and the effects of atmosphere, camera response, noise, skyglow, etc. etc.

Anyhoo, being limited to the spectral response of our eyes discards much of that lovely H-alpha detail we're so fond of seeing ! Horses for courses, heh, what!?

It would be nice however to know what some of these objects look like 'for real' if you could ever get close enough.

I guess what I'm looking for here is help with the question: which image do you find most pleasing ?

[Striker]

I still like the red image ......I just dont associate the lagoon being blue.

Its still nice to look at though.

[Astroman]

I think finding out the true colours of the objects can only be obtained with filters that only pick up certain wavelengths of light and subtracting the variations of wavelengths from our atmosphere. Giving us ONLY the colours of the object.

[Lester]

Hi Bill,

I am with Striker on this one. And your red version still shows a hint of blue, which is very nice. I look at other Lagoon images now and think they look too red, after seeing yours. But for me the blue one is too blue.

Your image is stunning to say the least.

[spearo]

I agree with Lester and Striker.
I also feel that the "red" one shows a better range of the different gas/clouds/emissions in the nebula.
But I doubt there is any right/wrong answers here but if we knew a particual nebula had a certain set of frquencies i the emissions we could aim for images that show a distinction between them i guess.

basically, I liked the "Red" but the "Blue" i snice too
frank.
frank

[spearo]

I need to learn the art of spellchecking...
f

[Doug]

Bill, I just re-read your opening post.......properly this time:doh:

I'd say the Red verion is better, more natural (unpopular comment to follow::evil: )

The trouble with DSLR cameras is that some photons will hit blue filtered photo sites (as they will with filtered monos too), and sometimes they are not treated correctly by the 'debayering' of the raw image. Any debayered image is an estimate based on adjacent pixels. This sometimes leaves a blue/magenta cast in the image that should not be there. This is only an issue with Astro-images, not terrestrial.

Below is the blue sub frame from the previous M8 that I posted. Blue data is there alright, however in a ballanced colour image it will be fully suppressed in areas that contain stronger Red and or Green signals. The M8 image data that was acquired by Jim Misti, has no bayer matrix to interfere with the final colour result.
Having said all that, Astro-images are largely art and so this allows for a lot of artistic expression.
Your red version as I said looks more realistic I think.
Sorry for misreading your original post.

cheers,
Doug

[2020BC]

Very interesting, Doug. It'll take me a while to digest that technical info. Very helpful. Thanks.