Hubble Palette - The Emperor's New Colours?

[Stub Mandrel]

I've been 'doing' narrowband imaging for nearly a year now and something is really bugging me.

I understand the Hubble Palette in principle - it takes Sii,Ha and Oiii and substitutes theme for red, green and blue.

This give images that are predominantly green, due to the abundance of Ha in most objects, so i can understand the approach of rebalancing the colour to get an image that highlights the distribution of all the colour elements.

To me a scientifically 'ideal' mapping would be balanced so Ha-only areas were green, Sii only (very rare) red, and areas with both Sii and Ha yellow, for example.

But it is almost universal to apply heavy processing to remove all green from the image so the Ha becomes not just yellow, but brown. Areas which also have Sii just become a barely distinguishable brown.

Another impact of this 'green removal' is that there is little to distinguish areas strong in both Ha and Oii - you would expect these to be blue-green, some sort of teal/turquoise, but instead most nebulas end up divided into two blocks of brown and pale blue with very little subtle differentiation in the colouration.

Essentially, even from the most respected of imagers we get bi-colour images, blue and brown, much information on the makeup of the nebula is lost and the image loses subtle colour variations that should make it more aesthetically pleasing.

To me, such images are aesthetically most pleasing with the Ha in yellow, and at least this does allow Sii rich regions to at least have a somewhat duller brown colour.

Why is there such effort made to remove green? It may not be present in 'true colour' DSO images, but these are false colour so isn't this a 'false argument'?

An alternative is to use a mapping of HSO instead of SHO.

This gives hydrogen and oxygen colours closer to their 'natural' ones and sulphur, always a minor element of any image and combined with some Ha signal, appears as yellow highlights. Far less drastic processing is needed to achieve an aesthetically pleasing result and more information on the distribution of elements is retained.

A criticism might be that the colours are too bright and cheerful with SHO mapping, and some may find the magenta hue of Ha +Oii regions as unpleasant as I find mud brown.

What do others think?

[TerryMcK]

I’ve experimented with SHO, HSO, OHS and OSH and of course they all give radically different colours dependant upon single strengths and targets. However I tend to alter the greens to get more pleasing combinations (at least to my eyes). I seem to end up with golden hues or pinkish tinges. I haven’t yet got a standard I consistently use and have a look at other peoples work to get inspiration.

I also do Bi-colours especially when using the OSC. My cameras are both the same sensor, one being a colour the other is a mono, and I sometimes blend both if I have subs of the same target to get more “realistic” looking images.

[Dmack1]

Thats a really interesting question Neil - and as a OSC/HOO newbie I may not be qualified to input much but here my thoughts anyway.

I think you have part of it with your "true colour" "false argument" idea. I'm new enough at this that I am still impacted by the cries of "there's no green in space (except great Orion)" and "Use HLVG" or "Use SCNR". You become programmed that if you have a green cast in your RGB image you are showing yourself up as a neophyte (or maybe a heretic ?) so you become allergic to green in all its shades - and perhaps that stays with you even into your NB False Colour Career. 

But I think another part of it is our desire to gain approval for our images. And the more the respected imagers (your phrase) go in that direction the more those of us just learning the ropes will follow. It is extremely difficult to be putting out images which you personally find aesthetically appealing but are not in keeping with current trends.

And that's a pity. Because, at the end of the day any false colour image (including OSC/HOO) is, in reality, a construct diverged from "reality" and should be considered as that - a series of personal artistic choices resulting in a final presented image. But, in truth, thats the case in any form of AP. Otherwise all we'd be looking at is the dark brown canvas of a single unstretched sub-frame.

And thats where we run into the "AP as art/science" conundrum.

For me personally, my goal in AP is to produce pleasing images which I, and hopefully others, can appreciate as just that -pleasing, thought provoking, maybe even awe-inspiring images of something we simply cant see with our own eyes. But there are others who eschew the artistic principle and see AP as a purely scientific endeavor. I was recently told (on another forum) that "AP is a mathematical concept" - (the quote is real and accurate ?) - followed by the implication that if I could not fully understand the maths behind Pixinsight then I wasnt a real AP'er and would never scale those lofty heights. 

So I think it really comes down to what we want out of AP - to produce nice pictures that you and some others like, to gain the approbation of other more experienced imagers or present scientific data for analysis. (FWIW I am firmly the first but with an unfortunate dollop of the second ? )

 

 

 

 

[GazAstro]

Interesting read ... I'm personally hooked in narrow band as I'm in bortle 7 to 8 so when I tried Ha it was a proper Eureka moment.

I realised straight away that I'd moved from taking pictures to taking pictures and making art with them.

I'm ok with that as I tweak away til I like the look of it ?

I probably aim for something Hubble like as that's what people expect when they see my efforts.

Dylan O'Donnel has a useful tool that shows all the outcomes possible if you upload a small jpg of your Ha Oii and Sii.

[Carastro]

I am a regular Hubble palette imager living in Bortle 8.  But without the Green colour toned down, it tends to overwhelm the other filters, which only tend to show through if the Ha is no longer green.

In the past I used to hate narrowband with yellow but subtle tones of orange, and beigey/pink colours look quite nice.  I also saw a deep pink and blue once, which was a combination of Filters of Sii/Ha in the red channel, and Ha and Oiii in the blue channel, (with varying opacities). 

I Have also done HOO which I quite like with certain targets, and also HSO and HSS even where there was no Oiii (California nebula). 

I think it all comes down to taste, but I agree there has come about some sort of recognised uniformity to the sort of colours people expect.  

I am not keen on the muddy browns myself. 

Carole 

  

[Gina]

A very interesting discussion  ? 

[AstronomyUkraine]

I'm strictly a narrowband imager. I process mainly in SHO combinations, with slight variations, depending on the colours I want to achieve. HOO is another favourite combination of mine, although these targets are somewhat limited. I don't believe there is a right or wrong colour combination, after stretching the image, it's all down to preferences.

[Gina]

Posted by: @AstronomyUkraine

I'm strictly a narrowband imager. I process mainly in SHO combinations, with slight variations, depending on the colours I want to achieve. HOO is another favourite combination of mine, although these targets are somewhat limited. I don't believe there is a right or wrong colour combination, after stretching the image, it's all down to preferences.

Same here.  I agree, false colours are purely a matter of taste.  I don't find SHO works very well in general.  Plus removing the green produced by Ha must be counter-productive!

 

[Hallingskies]

I care not for the fashionable greens in emission nebulae and for the Hubble pallet in general. My skies restrict me to narrowband most of the time, and then I tend to stick to HOO. On the rare occasions I bother with SII, I blend it with the Ha in the green channel, which gives some pleasant gold highlights in some nebulae. It’s all down to personal taste of course, but I just think emission nebulae should be red, not lurid blue or green.

[AstronomyUkraine]

Posted by: @Gina

I don't find SHO works very well in general.  Plus removing the green produced by Ha must be counter-productive!

 

What I have read on the Pixinsight forum regarding SCNR to remove the green caste, because of the dominant Ha data, is only to use it to align the green channel with the other two channels. Removing too much green can result in a washed out image, or a magenta background. I prefer to keep some green in a nebula image, to help with the gold colours associated with the hubble palette.

[Gina]

Posted by: @AstronomyUkraine

Posted by: @Gina

I don't find SHO works very well in general.  Plus removing the green produced by Ha must be counter-productive!

 

What I have read on the Pixinsight forum regarding SCNR to remove the green caste, because of the dominant Ha data, is only to use it to align the green channel with the other two channels. Removing too much green can result in a washed out image, or a magenta background. I prefer to keep some green in a nebula image, to help with the gold colours associated with the hubble palette.

Yes, I agree.

 

[Astrobdlbug]

Just got around to reading this  - bit late to the party ...again, Anyway very interesting to read people's thoughts on colour combinations of narrowband channels.

I just started imaging with narrowband filters this year, and enjoying the journey so far - interestingly it has got me using Pixinsight significantly more than when I imaged LRGB.  I think for a few reasons - not straying too off topic, but Ha lends itself to a lot more de-convolution and stretching - due I assume that S/N is much better than the noise fest I endured with Lum. Plus the Starnet++ application that very effectively separates nebula structure and stars to allow different processes to be applied to each. Finally its the PixelMath function I have been using to combine Ha, Oii and Sii and there seems to be a whole lot of equations out there to 'play' with.
I have been combining stars as HOO to avoid the pink SHO style, I also read Jon Rista's web blog on HOO combinations and also Coldest Nights who uses some quite interesting PixelMath functions that give different results.

I am still learning about how the narrowband channels combine and contribute to a final image so that I don't hold many strong views right now apart from avoiding garish over saturated rainbow mixes...but I do like a bit of green merging to yellow in the transitions within gas clouds.

I am interested in the way the equations mix contributions from each filter in odd ways, not necessarily just 20% of this and so 80% of that into the final RGB render.

I look for colours that highlight structures and the boundaries between different emission layers in nebula gas clouds. SHO is the goto for me as that is where I can compare what I am managing to create from my setup and processing skills.

I collected a few hours on Sadr region before the cloudfest that has been last millennium - well ok several weeks but feels that long.... and out of boredom and some learning I have been pushing and pulling that Sadr data and created umpteen different versions with good 'ol Crescent Nebulafr    floating up in top right like a brain in space. I've produced a sort of SHO, Red 'Autumnal' shades, Yellows and yes browns as well....

Finally I do find myself back in Photoshop and using selective colour tool that pulls all the Pixelmath work to pieces - but it can be fun when trying to just get that bit more colour where it probably shouldn't be....

Finally, and this is a bit 'arty' but I generally think the data pulls the end result into certain directions, we cant always , if ever gather the perfect billion hours of subs - what we can pull down from our back garden or balconies is what we get and then the image takes you down certain ways due to how much particular channel can be stretched and relative noise in channels etc....possibly a bit too much Zen and the art of astroimaging there for some :)

Hopefully some clear skies soon......
Bryan

[Jkulin]

Yep @stub-mandrel an interesting thought provoking subject.

Bottom line is each individuals choice and whichever floats your boat.

I have done SHO, HOO and even HOS, whichever I think suits the image.

I remember Sara Wager doing B33/Horsehead etc in SHO and it is still my favourite images of it: -

I feel she broke the mould on this one and succeeded.

Technically there is supposed to be no green in space and I can partially accept that, but until we can visit these areas in space, we'll never know and to my mind, they may never represent any colours we process at, but so what, enjoy being creative.

[Dmack1]

@Jkulin Well if I had just read "Horsehead in SHO - behind HH is green/gold" I'd have thought ? . But that is just a stunning image. I'd not seen it before but it is fantastic and I'd probably agree that it has to be one of the top picks for that target. But then she is a virtuoso imager.

But I'd wager (and that pun was absolutely NOT intended!!) she got a few tuts and shaken heads from the purists about it.

Thanks for sharing that because it really does make a point.

[Stub Mandrel]

Posted by: @Jkulin

Yep @stub-mandrel an interesting thought provoking subject.

Bottom line is each individuals choice and whichever floats your boat.

I have done SHO, HOO and even HOS, whichever I think suits the image.

I remember Sara Wager doing B33/Horsehead etc in SHO and it is still my favourite images of it: -

I feel she broke the mould on this one and succeeded.

Technically there is supposed to be no green in space and I can partially accept that, but until we can visit these areas in space, we'll never know and to my mind, they may never represent any colours we process at, but so what, enjoy being creative.

I have to agree that's an outstanding image by Sara.

To my mind it's a good combination of aesthetics and says something about the distribution of gases.

Interesting how the pretty little blue refection nebulas near some of the larger stars have disappeared, being a narrowband image.

 

[Stub Mandrel]

Lots of interesting responses here, I think the consensus (not surprising given the origins of this group) is that what matters most is getting a result that you find compelling yourself :-)

[Gina]

Sara's images are really superb!

[Stub Mandrel]

Well I think it's only fair if I have a go and put up two images to compare:

HSO

SHO

Both adjusted with selective colour

<edit - and tweaked so they have similar contrast>

[GazAstro]

Had a crack at that last year ... Nowhere near Sara's but I enjoyed it .