Are we over-removing Green in our SHO-Narrowband-Images?

[graemlourens]

Hi everybody.

Due to a discussion with another SGL-Member i started critically thinking about if I (and probably others as well) am maybe over-removing Green from my SHO images 'just because all others look like that too'.

I am thinking about this more in a scientific way - obviously everybody is free to make a pretty picture with whatever color range & balance they prefer. In this case i'd like to figure out a few more scientific facts that will explain why so many SHO-Images don't show any visible green or green hue.

I got very aware of not leaving any green due to the fact that most targets do not have true greens in them (when imaging RGB). With SHO obviously HA is very dominant and has to be equalised relative to SII & OIII to maintain a balance so by definition we're trying to tone down the greens by a far amount, what leads to the brown/yellowish SH-Regions seen in most SHO images.

As in most cases the SII Regions match up with the HA regions pretty well, until now this was a no-brainer for me to run a SCNR or HLVG over the image to get rid of that green cast. But aren't we removing a crucial part of the information of the image? Essentially we're creating a more Bi-Color image with Teal & Brown with various shades of it.

I have not come across a target where SII & HA Regions differ so much, that one could produce distinct green & distinct red areas, but maybe these exist and I have not got to them yet? (I'd be happy for examples if anyone has targets like those)

This is a little of a brain twister for me, and I can not settle to continue doing this approach before understanding more why and i can not get enough arguments by myself.

I'm happy to hear any ideas, thoughts and comments on this issue!

Edit: Per request an example. Image number one is my image from this year of the elephants trunk, and the second version is a test i just did today without removing most of the greens. Obviously i would have stretched the SII a little harder to get more of the yellow/brown in the right hand corner, but you should get the essence of my question.

Number 1 (for referencing later on)

Number 2: (for referencing later on)

 

Kind regards, Graem

[Adam J]

Perhaps you could post two examples one balanced as you normally would and the other with the additional green to show what you mean?

 

[johnrt]

I don't think there is any scientific basis for removing green or leaving some / all of it in. It is purely a case of aesthetics. A Hubble palette image done with skill should not be reduced to just 2 colours, there should be red/orange into gold and blue tones all clearly visible.

Let's not forget that narrowband is a false colour image, so any colour combination goes but we more often settle on the most appealing to the eye - the Hubble Palette.

 

[graemlourens]

1 hour ago, Adam J said:

Perhaps you could post two examples one balanced as you normally would and the other with the additional green to show what you mean?

 

Hi Adam. Absolutely, i have attached an example to the original post.

44 minutes ago, johnrt said:

I don't think there is any scientific basis for removing green or leaving some / all of it in. It is purely a case of aesthetics. A Hubble palette image done with skill should not be reduced to just 2 colours, there should be red/orange into gold and blue tones all clearly visible.

Let's not forget that narrowband is a false colour image, so any colour combination goes but we more often settle on the most appealing to the eye - the Hubble Palette.

 

Thank you for your response. In this case - if you look at my example images above - wouldn't the second (green) one be more accurate in terms of representing the different atoms of the nebula? As its correct HSO, why does one see many many more Versions like the first image around, and not more with the green hue?

Kind regards, Graem

[johnrt]

12 minutes ago, graemlourens said:

Thank you for your response. In this case - if you look at my example images above - wouldn't the second (green) one be more accurate in terms of representing the different atoms of the nebula? As its correct HSO, why does one see many many more Versions like the first image around, and not more with the green hue?

Yes perhaps number 2 is a more accurate representation, but we more often see or strive for number 1 because it is much more pleasing to the eye.

 

[graemlourens]

1 minute ago, johnrt said:

Yes perhaps number 1 is a more accurate representation, but we more often see or strive for number 2 because it is much more pleasing to the eye.

 

I suspect you meant it the opposite way (nr 1 is more pleasing to the eye) ? (I've numbered the images in the original post to avoid confusion later on)

[johnrt]

3 minutes ago, graemlourens said:

I suspect you meant it the opposite way (nr 1 is more pleasing to the eye) ? (I've numbered the images in the original post to avoid confusion later on)

Yes, wrong way round, I edited my post.  

[Filroden]

On the aesthetics, I suspect it is because the combination gives a very acidic range of yellows and greens which are not as appealing as the more earthy and natural browns/golds and blues. The second image certainly gives more information about what's going on, as the transition between the colours is clearer. Interestingly, if you put image 2 through PS's auto colour, it's not far off image 1.

[graemlourens]

29 minutes ago, Filroden said:

On the aesthetics, I suspect it is because the combination gives a very acidic range of yellows and greens which are not as appealing as the more earthy and natural browns/golds and blues. The second image certainly gives more information about what's going on, as the transition between the colours is clearer. Interestingly, if you put image 2 through PS's auto colour, it's not far off image 1.

Thank you for your thoughts! I agree that version 1 is much more pleasing, but i can not for certain say if its not a cognitive bias as i'm used to looking at the pictures similar to it.

A further thought: Is it correct to say, that maybe because SII is most of the times occupying the same image area as HA, that a greenish touch can be avoided and only using the shades of yellow over to reddish brown is also adequately representing the different atoms and in the end actually 'correct'?

A reason not to leave the green overcast would be that in any case we're not obeying the relative mass of the atom clouds, as there is far more HA there than OIII & SII combined, so we're 'equalizing' those 3 and when you do so HA & SII you end up with a lots of Gold/Yellow with the few spikes here and there into red or green depending on the target and the true challenge would be to stretch HA & SII (and OIII) so that certain extremes are visible and one does not end with a average soup that is always the same color and will nearly give you a bi-color (blue/gold) image?

Looking at the hubble pillars of creation (the first version) there is a ton of green, what did surprise me a lot. In the 2014 version they made a totally different balance more towards 'our' normal approach.

Kind regards, Graem

[Filroden]

42 minutes ago, graemlourens said:

A reason not to leave the green overcast would be that in any case we're not obeying the relative mass of the atom clouds, as there is far more HA there than OIII & SII combined

I think again it depends on what you are trying to communicate with the image. As it stands, because the three narrowbands are being normalised, you're showing distribution but not intensity. If you want to show intensity, individual mono images or very contrasting colour mapping would be needed. If it's just to show how beautiful the area is, then tonal ranges and complimentary colours will play a bigger part in the decisions. All three have their merits. For me, there are compromises that can achieve one or more of them but they don't quite tell everything.

[CCD Imager]

Graem

I think others have summed your discussion well. The hubble colours are artistic and artistic alone and the correct colours are those that are most appealing to the majority of viewers. Personally, I believe a little green should remain, so as to give more colour depth to the image.

Adrian

[graemlourens]

11 hours ago, CCD Imager said:

Graem

I think others have summed your discussion well. The hubble colours are artistic and artistic alone and the correct colours are those that are most appealing to the majority of viewers. Personally, I believe a little green should remain, so as to give more colour depth to the image.

Adrian

Hi Adrian. Thx for your input!

I do agree that one should not remove all green, but i think that if one balances the image well, green is rarely visible as green (only if there is a total lack of SII in an are, and i have not come around to such a target yet) but more that it contributes to the yellow/gold. 

I suspect that as its in most cases obvious that there would be a green cast everywhere due to the HA dominance, one removes it and rather plays with the other shades because - as ken stated it could be - the green cast is visually not appealing to a majority of viewers.

I don't agree though that the hubble palette is purely artistic. I think a large role it plays is to visualize the different gases in a target in a way that is visually appealing. In my case i'm not only trying to make a pretty picture (anymore). The longer i'm at this hobby the more time i spend per target, and the more i'm trying to 'get it right' even if thats an endless task it seems!

I do have to do a little more testing though to play with the various results when considering that one does not want to loose the green information, without seeing green its self. Just running SCNR/HLVG over an image does not feel right anymore.

Kind regards, Graem