Which broadband RGB filter lets through OIII the most ?

[glowingturnip]

Just an observation really - I'm currently processing some Thor's Helmet data, strong in OIII.  I've got about 5 hours each of Ha and OIII, and I just thought that I'd maybe add to the general star colours, so got about half an hour each of broadband R, G and B.

I was quite surprised to see that although the R and G channels weren't showing much more than a smudge of the nebula each, the B channel was clearly showing some structure to it (nowhere near as much as the narrowband, but still)

 

 

Checking the numbers, OIII is actually two very close bands at 500.7nm and 495.9nm in a roughly 16 parts to 5 parts mix, so call it an average of 499.6nm.

Checking my filter specifications (they're Astronomik Deep Sky RGB):

so if the graph is accurate, that 500nm band is right in the overlap between the B and G filters, the B is letting it through 100%, but the G is almost bang on the line, so is presumably attenuating a lot of it.  Hence it makes sense that in my results I'm seeing most of the structure in B.

 

However, I did a bit more digging - the red on our computer screens is about 675nm, green is about 510nm and blue is about 440nm.  On that basis 499nm mapped to computer RGB space should be about (0,255,146) so definitely greener than it is blue.  There's a very nice little calculator here (https://academo.org/demos/wavelength-to-colour-relationship/) that gives this:

 

So all in all, I wonder if my broadband B filter lets through too much green, and my G filter not enough ?  Either that, or I'm picking up another element that's properly emitting in the blue spectrum - I think I'll call it Stuartium :-)

 

 

[Demonperformer]

I follow the first part of your argument, but think the second part is a bit of a red herring?

You have captured your three frames, but each of those is being displayed in b/w on the screen, so the exact wavelengths of the colours on the screen don't come into it? Once you mix them to form a colour image then maybe ...

Unless I am missing something?

[Shibby]

As you've noticed, OIII should be turquoise and sit right between green and blue. I think, in reality, there's probably also some reflection nebulosity going on. This tends to be blue in colour due to young stars being the hottest & brightest, so perhaps there is some of this contributing to your blue channel?

[Shibby]

Sorry, I may have missed the actual point of your post. I just looked at the Baader graphs and it looks to me like the G filter does pass all of the O-III. It's a bit difficult to tell on the Astronomik chart to be honest - the green line is a bit too thick to see clearly what's happening at 500nm!

[glowingturnip]

5 hours ago, Demonperformer said:

I follow the first part of your argument, but think the second part is a bit of a red herring?

You have captured your three frames, but each of those is being displayed in b/w on the screen, so the exact wavelengths of the colours on the screen don't come into it? Once you mix them to form a colour image then maybe ...

Unless I am missing something?

sorry, here's what it looks like when channel-combined and colour-balanced so the surrounding stars have the correct spectral colours - definitely gives me a blue bubble  :-)

[glowingturnip]

2 hours ago, Shibby said:

As you've noticed, OIII should be turquoise and sit right between green and blue. I think, in reality, there's probably also some reflection nebulosity going on. This tends to be blue in colour due to young stars being the hottest & brightest, so perhaps there is some of this contributing to your blue channel?

actually, that's a good point re the reflection nebulosity.  I guess most reflection illumination won't survive past the narrowband filters, and there must be quite a bit in there.

I might try using a bit of that blue for the nebula - it highlights the wrinkly bits of the bubble a bit more than the OIII does so might give a nice bit of colour contrast.  So Ha to red, OIII to the turquoise colour above, and drop a little bit of the B into the blue channel (just for chrominance, the narrowband luminance is much better than the above).

I'm not at the stage where I can start combining stuff yet, so you'll have to wait for the results of my little experiment  :-)

[Demonperformer]

Yes, but as from your initial RGB results the blue image showed the most detail, you would expect the result to have a blue bias.

It seemed from the second part of your arguement that you were saying it should be more green because that is the colour that is closest to the 500nm point on the screen.

Let me summarise my understanding of what you are saying (in case I have got the wrong end of the stick): The blue filter captures more OIII because the OIII line falls mainly within it (and just a bit in the green filter) [which I accept], but the screen should display the combined image as more green because that is the colour that is closest to the OIII line [which I think is erroneous].

Once you have the three greyscale images (RGB) where the OIII line falls is irrelevant, you are mapping those three images to Red Green and Blue in the final combined image, irrespective of from what or where the original signal came.

[Shibby]

2 minutes ago, Demonperformer said:

but the screen should display the combined image as more green because that is the colour that is closest to the OIII line

I think he's just questioning why the green filter doesn't seem to cover the O-III wavelength.

[glowingturnip]

yes, exactly - it just struck me as strange that my blue filter seems to pass much more of that OIII line than my green filter does.  Given how key that wavelength is for astro, you'd think that Astronomik would make sure that both the G and B filters pass it (as it seems the Baader ones do per Shibby), rather than having the OIII right on the edge of the G passband as the graph suggests.

[Demonperformer]

"wrong end of the stick" it is then.

[glowingturnip]

no worries mate  

interesting experiment though, I thought - even if the conclusion is that my green filter is rubbish...