Jump to content

  •  

CNers have asked about a donation box for Cloudy Nights over the years, so here you go. Donation is not required by any means, so please enjoy your stay.

Photo

How do clean-cut LRGB filters tell spectral orange from red?

  • Please log in to reply
59 replies to this topic

#1 5th Gin

5th Gin

    Explorer 1

  • -----
  • topic starter
  • Posts: 54
  • Joined: 24 Aug 2020

Posted 09 September 2020 - 03:00 AM

I'm new to astronomy and still purchasing my first telescope/camera, so this might be a very stupid question, but I noticed that most LRGB filters I found are very different from what I expected: very clean cutoffs except a slight overlap between G and B channel for OIII. Can they actually perceive spectral colors like orange, yellow, or violet?

ofil-bp-lrgb-2_1.jpg

 

 

As far as I know, human eyes and color cameras perceive a full spectrum of wavelengths based on the ratio among RGB. However, this information seems lost in those filters: an orange spectral color will trigger a full red & luminance channel response and nothing in other channels, just like red will do.

183QE1.jpg

 

 

Now of course, a mixture of red and green colors can also appear yellow or orange, which is how computer screens work, and can be captured by those filters. A continuous spectrum like sunlight might work also. But what about those distinct colors? Can they be hinted by those filters at all? (Outside of those popular emission lines which have dedicated narrowbands, of course. I mean - there are plenty of other space objects apart from emission nebulas, right?) Should we avoid those filters then (I did find an Astronomik set with some overlaps in response), or are those colors just so rare in the universe that we can just ignore and mix with others?

astronomik-lrgb-typ2c_trans.png

 


  • Deven Matlick and Simcal like this

#2 sharkmelley

sharkmelley

    Fly Me to the Moon

  • *****
  • Posts: 7,489
  • Joined: 19 Feb 2013
  • Loc: UK

Posted 09 September 2020 - 03:24 AM

You're absolutely right - sharp cutoff RGB filters are not designed for good colour reproduction.  For instance, if you take an image of a continuous spectrum using typical RGB astronomy filters you will not see a continuous variation of colour but 3 distinct "blocks" of red, green and blue.  To answer your question, they are unable to distinguish spectral orange from spectral red.

 

To obtain good colour reproduction, not only are large overlaps in RGB transmissions required (e.g. Astronomik Type 2c) but also their response should also ideally fulfil the Luther-Ives condition which means their response is a linear combination of the human eye response - this is what DSLR camera sensors try to do.

 

Having said all this, for astrophotography it's the colours of stars, interstellar dust and emission nebula that are important and sharp cutoff RGB filters do not do a particularly bad job of this - depending on what your goals are.

 

If you are interested in good colour fidelity then the processing sequence is equally as important as the sensor/filters used.  White balance needs to be applied followed by the colour correction matrix appropriate for the sensor/filters and then the gamma curve of the colour space being used.  Background subtraction needs to be performed while the data are still linear and any stretching operations need to preserve the hue and saturation of colours.  Very few people deliberately employ use such techniques.  Additional complications are caused by the fact that the colours found in some emission nebula and some very deep red stars are outside the colour gamut of the colour spaces and display devices being used.

 

Mark


Edited by sharkmelley, 09 September 2020 - 04:02 AM.

  • Deven Matlick, jdupton, BinoGuy and 2 others like this

#3 Drakkith

Drakkith

    Vostok 1

  • -----
  • Posts: 194
  • Joined: 06 Aug 2011

Posted 09 September 2020 - 04:39 AM

Note that except for emission nebulas, there are practically no sources of specific spectral colors in the sky. Stars, galaxies, and reflection nebulas are all broad-band sources that will be seen through all three RGB filters to some extent. When narrow band imaging you don't usually even try to reproduce accurate colors, as out of the standard SHO filters one is in the blue-green area of the spectrum and two are in the deep red.


  • Madratter, sharkmelley and bobzeq25 like this

#4 Alex McConahay

Alex McConahay

    Hubble

  • *****
  • Posts: 13,408
  • Joined: 11 Aug 2008
  • Loc: Moreno Valley, CA

Posted 09 September 2020 - 08:46 AM

One of the things that makes it difficult to understand how the filters are working is the illustrations you are using.

 

Note that the colors they represent show a spectrum that consists of ONLY pure blue, pure green and pure red. The area under the "blue" curve, for instance, is colored the same blue from one side to the other. There is no representation of yellow, teal, orange, etc. There is no recognition that blues range from violet blues to aquamarine blues.  

 

This makes it look like there are only three monolithic colors when there are really shades of all colors in there. In fact, the spectrum is continuous, and continuously changing. 

 

Were an actual continuous spectrum to be pictured, it would be easier to see how combining light from separate filters could represent intermediate colors.  

 

Alex


  • BinoGuy and bobzeq25 like this

#5 bobzeq25

bobzeq25

    ISS

  • *****
  • Posts: 34,871
  • Joined: 27 Oct 2014

Posted 09 September 2020 - 11:22 AM

Sharp color cutoffs do just fine in reproducing ALL colors.  They're all some combination of red, green and blue.

 

I think they make color processing easier.  The newer Astronomik Deep Sky RGB (which I have) have sharp cutoffs.

 

The Bayer matrix records some red and blue light as green.  There are techniques to address that, I don't find them as easy.  In any event, the Bayer matrix filter is designed for terrestrial use, it's inefficient for astro.

 

Bottom line.  You can make excellent images both ways.  A mono camera plus LRGB filters is more efficient, which is especially useful in light polluted skies.  But the equipment is way more expensive.  People spend that money for good reasons, this isn't like buying a Rolex.  <smile>

 

I've done extensive imaging both ways (one shot color or mono plus filters), use them both in the right circumstances.


Edited by bobzeq25, 09 September 2020 - 11:25 AM.


#6 sharkmelley

sharkmelley

    Fly Me to the Moon

  • *****
  • Posts: 7,489
  • Joined: 19 Feb 2013
  • Loc: UK

Posted 09 September 2020 - 12:13 PM

Sharp color cutoffs do just fine in reproducing ALL colors.  

Except they can't distinguish between the wavelengths 600nm and 610nm which the eye sees as orange and red respectively - which is the original question. 

 

Mark


  • RossW and 5th Gin like this

#7 OhmEye

OhmEye

    Apollo

  • *****
  • Posts: 1,146
  • Joined: 15 Sep 2019
  • Loc: Western NY Southern Tier

Posted 09 September 2020 - 12:21 PM

I don't find the graphs difficult to understand at all. Coloring the graph is just to identify which filter is being represented, not the "color" of light. The light is represented on an axis as wavelength. It's a simple bandpass transmission for wavelengths. The filter passes the percentage at the given wavelength. Constructing an RGB representation of the original wavelengths will never be accurate, it's just a convenience to approximate what human eyes interpret as color. We could use a filter for each and every "color" wavelength we care about, but it's conventionally easier to use RGB.



#8 jdupton

jdupton

    Fly Me to the Moon

  • *****
  • Posts: 5,642
  • Joined: 21 Nov 2010
  • Loc: Central Texas, USA

Posted 09 September 2020 - 12:22 PM

Mark & Bob,

 

Except they can't distinguish between the wavelengths 600nm and 610nm which the eye sees as orange and red respectively - which is the original question. 

 

Mark

 

   Exactly. And the Astrodon LRGB sets cannot even see 585 nM light which we perceive as yellow-orange. The same is true of most (sharp cutoff) LRGB sets; they block all light around that wavelength. [Added Edit: You cannot reproduce a color you never captured in the first place.]

 

 

John


Edited by jdupton, 09 September 2020 - 12:26 PM.

  • sharkmelley and 5th Gin like this

#9 Alex McConahay

Alex McConahay

    Hubble

  • *****
  • Posts: 13,408
  • Joined: 11 Aug 2008
  • Loc: Moreno Valley, CA

Posted 09 September 2020 - 01:03 PM

>>>>>>[Added Edit: You cannot reproduce a color you never captured in the first place.]

 

Color theory says you can mix primary colors with each other to produce every color in the spectrum, doesn't it. ?

 

My monitor has phosphors only for red, green, and blue. (or is it the inverse of those?) Yet, I can see yellow on the screen. And many other colors.

 

 

What am I missing?

 

Alex


Edited by Alex McConahay, 09 September 2020 - 01:13 PM.

  • Astrola72 likes this

#10 Alex McConahay

Alex McConahay

    Hubble

  • *****
  • Posts: 13,408
  • Joined: 11 Aug 2008
  • Loc: Moreno Valley, CA

Posted 09 September 2020 - 01:07 PM

>>>>>>600nm and 610nm which the eye sees as orange and red respectively

 

Does the eye see as "orange" and "red?" or different degrees of a combination of yellow and red? Orange, yes, but yellowish orange and reddish orange?

 

If it sees 610 as "red," then what does it see 640 as? Certainly not the same color. Yet, I think most of us would call it red before we would call 610 "red."

 

Alex


  • Astrola72 likes this

#11 jdupton

jdupton

    Fly Me to the Moon

  • *****
  • Posts: 5,642
  • Joined: 21 Nov 2010
  • Loc: Central Texas, USA

Posted 09 September 2020 - 01:33 PM

Alex,

 

Color theory says you can mix primary colors with each other to produce every color in the spectrum, doesn't it. ?

   It does and you can.

 

   For me, the keyword is "reproduce" rather than "synthesize or produce". In theory, you could take terrestrial images using narrow-band filters at the emission lines of Hγ (434 nM), Oiii (496 / 508 nM), and Sii (672 nM) and synthesize a rainbow. However, in my mind, that would not be reproducing an image of a rainbow.

 

   Without a lot of manipulation of the data, simply displaying such an image would show only blocks of deep blue, (blue-)green, and deep red in the rainbow. It cannot be "reproduced" from the data. A semblance of it could be synthesized from the data but that is not same from my viewpoint.

 

 

John


Edited by jdupton, 09 September 2020 - 01:36 PM.


#12 sharkmelley

sharkmelley

    Fly Me to the Moon

  • *****
  • Posts: 7,489
  • Joined: 19 Feb 2013
  • Loc: UK

Posted 09 September 2020 - 01:42 PM

>>>>>>[Added Edit: You cannot reproduce a color you never captured in the first place.]

 

Color theory says you can mix primary colors with each other to produce every color in the spectrum, doesn't it. ?

 

My monitor has phosphors only for red, green, and blue. (or is it the inverse of those?) Yet, I can see yellow on the screen. And many other colors.

 

What am I missing?

What you are missing is that the Astrodon R filter does not capture the 585nm wavelength, nor does the Astrodon G filter, nor does the Astrodon B filter. If you try to shoot a continuous spectrum with the Astrodon RGB filter set, the image will have a black gap at 585nm between green and red.

 

Mark


Edited by sharkmelley, 09 September 2020 - 01:43 PM.

  • RossW and 5th Gin like this

#13 Alex McConahay

Alex McConahay

    Hubble

  • *****
  • Posts: 13,408
  • Joined: 11 Aug 2008
  • Loc: Moreno Valley, CA

Posted 09 September 2020 - 01:43 PM

Alex,

 

   It does and you can.

 

   For me, the keyword is "reproduce" rather than "synthesize or produce". In theory, you could take terrestrial images using narrow-band filters at the emission lines of Hγ (434 nM), Oiii (496 / 508 nM), and Sii (672 nM) and synthesize a rainbow. However, in my mind, that would not be reproducing an image of a rainbow.

 

   Without a lot of manipulation of the data, simply displaying such an image would show only blocks of deep blue, (blue-)green, and deep red in the rainbow. It cannot be "reproduced" from the data. A semblance of it could be synthesized from the data but that is not same from my viewpoint.

 

 

John

Okay.....

 

Alex



#14 deonb

deonb

    Ranger 4

  • *****
  • Posts: 334
  • Joined: 16 Jul 2020
  • Loc: WA

Posted 09 September 2020 - 02:02 PM

>>>>>>[Added Edit: You cannot reproduce a color you never captured in the first place.]

Color theory says you can mix primary colors with each other to produce every color in the spectrum, doesn't it. ?

My monitor has phosphors only for red, green, and blue. (or is it the inverse of those?) Yet, I can see yellow on the screen. And many other colors.


What am I missing?

Alex

You can indeed produce all human visible colors using RGB only. However there are two ways a human can perceive the color "orange". One is as a distinct 585nm wave. The other way is as a 560nm wave on top of a 700nm wave.

Your monitor can't reproduce the 585nm wave but practically it doesn't matter - it can produce 560 + 700, and your eyes (brain really) can't tell the difference. However, a filter can certainly tell the difference...

And not all objects will reflect 560+700 - some will only reflect 585.

Hence the topic of this thread.

For additional fun - read up on impossible colors:
https://nesslabs.com/impossible-colors

Edited by deonb, 09 September 2020 - 02:15 PM.

  • Der_Pit and 5th Gin like this

#15 Alex McConahay

Alex McConahay

    Hubble

  • *****
  • Posts: 13,408
  • Joined: 11 Aug 2008
  • Loc: Moreno Valley, CA

Posted 09 September 2020 - 02:26 PM

One of the things that makes it difficult to understand how the filters are working is the illustrations you are using.

 

Note that the colors they represent show a spectrum that consists of ONLY pure blue, pure green and pure red. The area under the "blue" curve, for instance, is colored the same blue from one side to the other. There is no representation of yellow, teal, orange, etc. There is no recognition that blues range from violet blues to aquamarine blues.  

 

This makes it look like there are only three monolithic colors when there are really shades of all colors in there. In fact, the spectrum is continuous, and continuously changing. 

 

Were an actual continuous spectrum to be pictured, it would be easier to see how combining light from separate filters could represent intermediate colors.  

 

Alex

Just to illustrate what the above means, I took the originally posted image of the filters and their cutoffs, and I superimposed the visible spectrum. You can see how misleading the characterization of the filters as "blue," "green," or "red" is. In fact all filters pretty much smoosh over several what we would call "colors" and there are no really distinct colors until you really narrow the bandpass. 

 

I really understand that the point is that one has not actually captured the bandpass at the heavily light polluted High Pressure and low pressure sodium emission bands (yellow) around 590. 

 

The camera is not perceiving (much) from precisely 590 maybe. But color theory says that one can mix different colors to get to other colors. And there is enough yellow in the red, that when mixed with the proper amount of green, it tells the computer to display "orange." 

 

I think the bigger issue is whether there is anything out there that displays in only one "color." That is, is there anything out there that displays only in a very narrow bandpass. 

 

5th Gin (original post) expresses it like this:

 

>>>>>But what about those distinct colors? Can they be hinted by those filters at all? (Outside of those popular emission lines which have dedicated narrowbands, of course. I mean - there are plenty of other space objects apart from emission nebulas, right?) Should we avoid those filters then (I did find an Astronomik set with some overlaps in response), or are those colors just so rare in the universe that we can just ignore and mix with others?

 

and  Sharkmelly says:

 

>>>>>>If you try to shoot a continuous spectrum with the Astrodon RGB filter set you will see a black gap at 585nm between green and red.

 

JD Upton's example of trying to reproduce a rainbow from three narrowbands follows the same logic. 

 

And I must agree with them. Sharkmelly and JD Upton's examples.

 

But, except for narrowbands targets (basically certain structures in narrowband nebulae, etc), are there any targets that present such imaging challenges? That is is there a star, or structure in a nebula, or a galaxy that does not emit a sufficiently wide range of light that it cannot effectively be captured and later reproduced by using broadband filters that have sharp cutoffs and (some) slight gaps between the bands. And I think there are not. 

 

Now, whether this is capturing, synthesizing, reproducing, or what, I am not so concerned with. I mean, even our eye/brain systems in visual astronomy are reproducing/synthesizing/whatever from the photons that hit the back of our eyeballs. 

 

Now, back to the original question, bottom line--from the original post:

 

>>>>>>>>Can they be hinted by those filters at all?

Sure, we do it all the time.

 

>>>>>>>I mean - there are plenty of other space objects apart from emission nebulas, right?)

Yep, and nearly everything we see up there is emitting a broad range of light that generally ignores the simple divisions we see on the original image classifying light according to blue, green, and red filters. Yet, we reproduce it. 

 

>>>>>>Should we avoid those filters

No, not at all. IN fact they should be preferred. They give you more control over how to mix and match the colors. (ANd, for the 590 gap, cut through light pollution.) 

 

>>>>>>then or are those colors just so rare in the universe that we can just ignore and mix with others?

Generally, the one color in particular, around 590 is all over the place here on earth because it is the predominant color of highly light polluting streetlamps. But, if there is a star out there someplace with lots of sodium, emitting away at that color, it will also be emitting other colors, and will show up just fine as a yellow star.  

 

At any rate.....we are here in the beginners and intermediate area. It is an interesting discussion. But, really, I would prefer the first graph you have up there, with the sharper cutoffs and the graphs. I am sure I would never notice the gap between the red and green. 

 

And here is that graph, redone to show that Red, Green and Blue are not all that exclusively what they say. 

Alex

Attached Thumbnails

  • Spectrum-and-Filter-matches2.jpg


#16 klaussius

klaussius

    Surveyor 1

  • *****
  • Posts: 1,674
  • Joined: 02 Apr 2019
  • Loc: Buenos Aires

Posted 09 September 2020 - 03:07 PM

I'm sure I'm oversimplifying things here, but aside from what has been said already, one has to keep in mind that there are 2 categories of light emitting objects in the sky:

 

  1. Emission nebulae, that emit in a few well known narrow bands
  2. Broadband targets, be it stars or dust that reflects star light, which emit a blackbody spectrum.

The key point here is that we have a mix of narrow band light, in bands we know and can design around, and black body radiation. Black body radiation is broadband, and has lobes that make it look "blue" or "red" based on temperature, but it's not just 700nm light or just 450nm light. It's just got a peak towards the blue or red. Or yellow. The color of stars is just where the peak emission is, but they emit the whole spectrum in varying degrees.

 

So, given that, clear cut filters will sample a different part of that black body spectrum and produce the necessary ratios so that a monitor will output "orange" quite easily. Ie: it will produce more red than green, because there was just more emission in the red band vs the green band.

 

Narrow band light is tricky. A mono sensor with sharp cutoff RGB filters will not perceive NB light color accurately, in that if it falls in the R band alone, it will be reproduced as red. Even if it was slightly orange.

 

But LRGB filter sets are usually designed with the known emission lines in mind. Ha and SII fall squarely in red, they look mostly red and that's fine. OIII is placed at an overlap zone between G and B to produce teal. So where you can't in general reproduce narrowband colors accurately with sharp bandpass filters, you can design them to accomodate the important emission bands and that's usually what's done.

 

Aside from the most important emission lines, there aren't that many objects producing NB light in sufficient quantities to matter. So it all works out quite fine.


  • 5th Gin likes this

#17 sharkmelley

sharkmelley

    Fly Me to the Moon

  • *****
  • Posts: 7,489
  • Joined: 19 Feb 2013
  • Loc: UK

Posted 09 September 2020 - 03:10 PM

And here is that graph, redone to show that Red, Green and Blue are not all that exclusively what they say. 

 

I don't know where you obtained that colour spectrum but it is very badly wrong.  600nm is definitely not yellow. Also, think about 656nm which is H-alpha - an H-alpha filter is a very deep red (which I'm sure you know) but the colour in that diagram shows it to be orange.

 

[Edit:  The spectrum displayed at the following link is pretty accurate and the problem is discussed of how to display the out-of-gamut spectral wavelengths: http://www.techmind....ur/spectra.html ] 

 

Mark


Edited by sharkmelley, 09 September 2020 - 03:54 PM.


#18 Alex McConahay

Alex McConahay

    Hubble

  • *****
  • Posts: 13,408
  • Joined: 11 Aug 2008
  • Loc: Moreno Valley, CA

Posted 09 September 2020 - 03:54 PM

Mark, 

 

I know what you are saying, but google Visible Spectrum, ask for images, and you will see a display of images. Some have blue to the left, some blue to the right, but in general, their color scheme matches the one I am showing. They are generally more saturated. I had to limit the opacity of mine to display the original graph underneath it. As in mine, 600 is listed as yellow orange generally on the visible spectrum. 

 

We can quibble about the display a bit. But the point of it is just that the spectrum varies continuously from one wavelength to the next. It is broken into an infinity of colors, not the three that we talk about. It is safe to say that a red filter as originally displayed in the first post covers a wavelength (580-590) that many of us would consider yellowish/orangish more than "red." 

Alex

https://www.uib.no/e...gnetic-spectrum

Attached Thumbnails

  • Spectrum.jpg


#19 sharkmelley

sharkmelley

    Fly Me to the Moon

  • *****
  • Posts: 7,489
  • Joined: 19 Feb 2013
  • Loc: UK

Posted 09 September 2020 - 04:05 PM

I know what you are saying, but google Visible Spectrum, ask for images, and you will see a display of images. 

The vast majority are completely wrong - they don't even pretend to reproduce accurate colour but they serve merely as illustrations.

 

Mark



#20 Xentex

Xentex

    Vostok 1

  • -----
  • Posts: 158
  • Joined: 10 Nov 2015
  • Loc: Philadelphia Suburbs

Posted 09 September 2020 - 07:43 PM

Reading the responses, I think there may be a couple seasoned astrophotographers here that are missing the truth that the OP was asking about.

 

If you photograph an orange light with a telescope using those sharp cutoff RGB filters it will look red.  If you image a rainbow using those sharp cutoff RGB filters, you will only see three colors.

 

Imaging the orange light bulb will show up entirely in the red channel, and none of it will be in the green and blue channels.  If you put a red light bulb next to it, and image them both at the same time, you will not be able to distinguish the difference between them.

 

As it turns out, that really doesn't matter much in the astrophotography most of us do.  Which is why seasoned astrophotographers who never really thought about it might not initially believe it.


  • RossW and 5th Gin like this

#21 bobzeq25

bobzeq25

    ISS

  • *****
  • Posts: 34,871
  • Joined: 27 Oct 2014

Posted 10 September 2020 - 11:36 AM

Reading the responses, I think there may be a couple seasoned astrophotographers here that are missing the truth that the OP was asking about.

 

If you photograph an orange light with a telescope using those sharp cutoff RGB filters it will look red.  If you image a rainbow using those sharp cutoff RGB filters, you will only see three colors.

 

Imaging the orange light bulb will show up entirely in the red channel, and none of it will be in the green and blue channels.  If you put a red light bulb next to it, and image them both at the same time, you will not be able to distinguish the difference between them.

 

As it turns out, that really doesn't matter much in the astrophotography most of us do.  Which is why seasoned astrophotographers who never really thought about it might not initially believe it.

The orange light would have to be unnaturally monochromatic.  A standard lightbulb with colored glass isn't monochromatic. You could get colored lightbulbs from Home Depot, image that orange light bulb and a red one with sharp cutoff filters, and get completely different (and quite realistic) colors.  The same thing applies to rainbows.

 

The orange light from such a bulb is a blend of R, G, and B, just as orange paint is a blend of pigments.  As seen by this diagrams in this excellent article on color spaces.  No significant astronomical light source is monochromatic orange.

 

https://programmingd...d-color-spaces/

 

Bottom line.  Sharp cutoff filters are perfectly capable of capturing the colors of astronomical objects, just as they are capable of capturing the color of colored light bulbs.  The theoretical argument that they could not distinguish 600 and 610nm never actually comes into play.

 

Personally I think the problem of the Bayer matrix capturing red and blue light as green has more impact on our images.


Edited by bobzeq25, 10 September 2020 - 11:41 AM.


#22 bobzeq25

bobzeq25

    ISS

  • *****
  • Posts: 34,871
  • Joined: 27 Oct 2014

Posted 10 September 2020 - 11:41 AM

>>>>>>[Added Edit: You cannot reproduce a color you never captured in the first place.]

 

Color theory says you can mix primary colors with each other to produce every color in the spectrum, doesn't it. ?

 

My monitor has phosphors only for red, green, and blue. (or is it the inverse of those?) Yet, I can see yellow on the screen. And many other colors.

 

 

What am I missing?

 

Alex

Nothing.  <smile>



#23 sharkmelley

sharkmelley

    Fly Me to the Moon

  • *****
  • Posts: 7,489
  • Joined: 19 Feb 2013
  • Loc: UK

Posted 10 September 2020 - 01:34 PM

Personally I think the problem of the Bayer matrix capturing red and blue light as green has more impact on our images.

Capturing red and blue light as green - you do realise that your eyes are doing exactly that?

 

Mark



#24 Xentex

Xentex

    Vostok 1

  • -----
  • Posts: 158
  • Joined: 10 Nov 2015
  • Loc: Philadelphia Suburbs

Posted 10 September 2020 - 02:08 PM

The orange light would have to be unnaturally monochromatic.  A standard lightbulb with colored glass isn't monochromatic. You could get colored lightbulbs from Home Depot, image that orange light bulb and a red one with sharp cutoff filters, and get completely different (and quite realistic) colors.  The same thing applies to rainbows.

It's not clear to me if we're arguing semantics here.

 

If you image a nebula with a sharp cutoff RGB filter set you will not be able to distinguish Ha from Sii.  They will both appear entirely in the red channel and not at all in the G & B channels.  There is nothing you can do in post-processing to split them.

 

If you were to image a rainbow produced by a prism using those sharp cutoff RGB filters, you would have that exact same problem in all three of your channels.  All you would be able to create in post-processing is a three color rainbow.  There is no way to re-create what was lost.

 

I already agreed with you that this is largely irrelevant in astrophotography because the only place it shows up to any material degree for backyard imagers is emission nebula.  However, that is not at all true for terrestrial photography, your "Home Depot lightbulb" notwithstanding.

 

Nevertheless, the original poster's starting question was literally "can they actually perceive spectral colors?"

 

The definition I learned for "spectral colors" in science was monochromatic light at a specific frequency.  Not that Wikipedia is a global authority, but that's how it defines "spectral colors" as well.  And, yes, we all learned that you can fool the human eye by mixing other colors, but that doesn't make it the same thing.

 

The short answer to the OP's question is "no, sharp cutoff RGB filters cannot distinguish spectral colors."  The slightly longer answer is "but the only place that really matters in this hobby is when imaging emission nebula, and for that narrowband filters are used."

 

He mentioned emission nebula in his question, and added "there are plenty of other space objects apart from emission nebulas, right?"  Obviously, yes.  But as you stated, relatively few produce monochromatic light, so the sharp cutoffs aren't really problematic for those.

 

And, since he brought up DSLRs and their overlapping RGB bayer filters, it is correct to say that the sharp cutoff filters would not work well at all for DSLRs shooting terrestrial targets.  Because rainbows, whether created by water droplets or crystals, are a phenomenon that is about a natural object spacing out monochromatic light.  At their core, many LEDs are monochromatic for the same scientific reason that emission nebula are.  Sodium vapor lights are largely monochromatic.  Lasers are largely monochromatic.  Neon lights are largely monochromatic.  Much of terrestrial photography would look very wrong color wise if shot through those sharp cutoff RGB filters.


Edited by Xentex, 10 September 2020 - 02:11 PM.

  • 5th Gin likes this

#25 gezak22

gezak22

    Gemini

  • *****
  • Posts: 3,478
  • Joined: 15 Aug 2004
  • Loc: On far side of moon. Send help.

Posted 10 September 2020 - 02:18 PM

Assumption: We are talking about broadband (>50 nm) sources and narrow (<10 nm) gaps.

 

When there is a spectral gap, part of the spectrum of the signal will indeed be lost. But the other parts are still present in the respective filters, so the combined image will still have a reasonably accurate color.




CNers have asked about a donation box for Cloudy Nights over the years, so here you go. Donation is not required by any means, so please enjoy your stay.


Recent Topics






Cloudy Nights LLC
Cloudy Nights Sponsor: Astronomics