Why the sky is blue and not violet?

[kmm]

I have some confusions about this that I'm not finding clear answers for. First, looking at the color violet, it looks like a sort of "purple". The answer as I've understood it to why the sky doesn't appear violet or this "purple" is that it has to do with a combination of the fact that there isn't as much violet produced by the sun and it's absorbed higher in the atmosphere. In addition to that is while violet appears sort of purple since it stimulates the blue cones and the red cones less, the net effect from the different frequencies of light coming into our eye is that the blue cones are stimulated most. So we see the sky as blue. What's confusing to me is that while thinking of violet as sort of a purple, I've come across some explanations that regard violet as a deep blue, not this "purplish" color. To add to this, if you look at a spectrum produced by a prism, it doesn't appear to show violet as this "purple" but a deep blue, however when you look at a rainbow you can see violet as this sort of purple. So I'm not sure what violet is actually perceived as. If violet is perceived as a purple, why doesn't this show up with a prism? If violet is perceived as a deep blue, why do we see "purple" where violet should be in a rainbow? Which is it?

 

[A.T.]

 

[kmm]

This claims that violet is perceived as a deep blue. Is this video correct though? Pretty much everything I see regarding violet shows it as a "purple" not a deep blue. For example: http://upload.wikimedia.org/wikipedia/commons/5/5f/CIE-1931_diagram_in_LAB_space.svg

 

[Andy Resnick]

Violet/purple are not spectrally pure colors, but a product of our visual system (so are brown and pastel hues). Color perception allows a surprisingly rich analysis: for example, mapping the 'color line' (the usual spectral representation) to a 'color wheel' (the usual visual representation) involves a non-trivial mapping that introduces a singularity.

 

[Khashishi]

All the rainbow colors are spectrally pure, but typical human vision has just three dimensions of color. The space of possible spectra is much larger than the space of colors that we can see, so many different spectra will give the same visible color. So purple (which is generally not spectrally pure) and violet (which may or may not be spectrally pure) look similar.

 

[cjl]

This claims that violet is perceived as a deep blue. Is this video correct though? Pretty much everything I see regarding violet shows it as a "purple" not a deep blue. For example: http://upload.wikimedia.org/wikipedia/commons/5/5f/CIE-1931_diagram_in_LAB_space.svg

I own a 405nm (violet) laser pointer, and I can tell you, I definitely perceive it as purplish, rather than blue. It does look a little different than a traditional "purple" though. It's kind of hard to describe without seeing it for yourself. That having been said, I wouldn't say it's just a deep blue.

 

[kmm]

Violet/purple are not spectrally pure colors, but a product of our visual system (so are brown and pastel hues). Color perception allows a surprisingly rich analysis: for example, mapping the 'color line' (the usual spectral representation) to a 'color wheel' (the usual visual representation) involves a non-trivial mapping that introduces a singularity.

I don't quite understand this. I thought violet has a well defined wavelength range (380nm - 450nm). I understood purple to be a mixture of blue and red though. So I'm interested in how violet is actually perceived. In most depictions of violet, it looks like sort of a purple, I assumed because it stimulated blue cones and red cones. However, I don't see why a prism doesn't show this but a rainbow does. I assumed because a rainbow doesn't produce a pure spectrum.

All the rainbow colors are spectrally pure, but typical human vision has just three dimensions of color. The space of possible spectra is much larger than the space of colors that we can see, so many different spectra will give the same visible color. So purple (which is generally not spectrally pure) and violet (which may or may not be spectrally pure) look similar.

Are you implying that a prism is not spectrally pure? If you're saying violet looks similar to purple, which is what I'm trying to determine, then I have to wonder why this apparently is never seen with a prism.

I own a 405nm (violet) laser pointer, and I can tell you, I definitely perceive it as purplish, rather than blue. It does look a little different than a traditional "purple" though. It's kind of hard to describe without seeing it for yourself. That having been said, I wouldn't say it's just a deep blue.

Interesting. I've read some other people saying the same thing. I wonder why the video said violet is perceived as deep blue then.

 

[A.T.]

I don't see why a prism doesn't show this but a rainbow does

At around 1:10 min:

 

[kmm]

At around 1:10 min:

OK now I understand.

What I wonder now is, why is violet often portrayed as a "purplish" color? And as cjl and others I read have said, why does a 405nm laser pointer is perceived as a more purple?

 

[sophiecentaur]

OK now I understand.

What I wonder now is, why is violet often portrayed as a "purplish" color? And as cjl and others I read have said, why does a 405nm laser pointer is perceived as a more purple?

It's "portrayed" that way because you cannot "portray' spectral violet any other way than with a spectral source. Why would you be surprised that the eye is fooled / confused when presented with a very rarely seen colour? The 'violet' that's shown on film, TV displays and printing with a pigment is much the same as spectral violet - certainly near enough for that small range of colours on the bottom left of the CIE chart to be called violet. And it is only called violet. It's time to accept that classification can be very wooly and particularly when it doesn't really matter. Never, in human evolution, was spectral violet seen (the rainbow is far too desaturated to be classed as a set of spectral colours); we were not 'designed' to distinguish that colour. I very much doubt whether most readers of this thread have actually seen a good, spectral violet (the laser pointer thingy would be the only likely example).

 

[kmm]

It's "portrayed" that way because you cannot "portray' spectral violet any other way than with a spectral source. Why would you be surprised that the eye is fooled / confused when presented with a very rarely seen colour?

OK. I wasn't confused that the eye would be fooled when presented with violet, I just wasn't sure what the eye would actually perceive in that case. It's just that some people were saying deep blue and others a "purplish".

The 'violet' that's shown on film, TV displays and printing with a pigment is much the same as spectral violet - certainly near enough for that small range of colours on the bottom left of the CIE chart to be called violet.

On film and tv etc violet appears to be a "purplish" color. So in that case are you saying that is close to what spectral violet would be perceived as? Not a deep blue?

 

[cjl]

I would say this image is pretty close to how I perceive my 405nm laser pointer (though my monitor can't quite replicate the true effect):
http://www.ledmuseum.candlepower.us/netg405.jpg

(Admittedly, with the caveats thrown in that the way you view that image will depend on your monitor and its calibration/settings).

 

[kmm]

I would say this image is pretty close to how I perceive my 405nm laser pointer (though my monitor can't quite replicate the true effect):
http://www.ledmuseum.candlepower.us/netg405.jpg

(Admittedly, with the caveats thrown in that the way you view that image will depend on your monitor and its calibration/settings).

Ok interesting. This is something like what I would actually expect our eyes to perceive when presented with violet. Our cones aren't as sensitive to violet, so the red and blue cones would be stimulated yet stronger in the blue giving the illusion of this purplish color. Although, the video in post #8 seems to give a good explanation why this isn't actually the case. Maybe the 405nm laser pointer isn't pure 405nm light then.

 

[dipole]

The 410 nm spectral line of Hydrogen certainly looks "purple" to me.

 

[Andy Resnick]

I don't quite understand this. I thought violet has a well defined wavelength range (380nm - 450nm). I understood purple to be a mixture of blue and red though. So I'm interested in how violet is actually perceived. In most depictions of violet, it looks like sort of a purple, I assumed because it stimulated blue cones and red cones. However, I don't see why a prism doesn't show this but a rainbow does. I assumed because a rainbow doesn't produce a pure spectrum..

A few comments- first, human color vision is based on three populations of cone cells:

http://wearecolorblind.com/wp-content/uploads/graphs.jpg

And thus, any 'pure spectral color' excites multiple cones. This is actually the reason why we can perceive so many hues, even though we only have (colloquially) 'red, green, and blue cones'.

So deep blue light (380 - 450nm or so) excites multiple cones- and as you can see from the above graph, the red and blue cones respond more strongly than the green in that waveband, so that's why we describe purple as a mixture of blue and red.

In any case, the more fundamental point is that 'color' as an attribute of electromagnetic radiation can be most easily quantified either as a linear scale or as a plane surface, and these different representations are not equivalent to each other. The first is used when quantifying the electromagnetic field in terms of it's frequency, while the latter is used to describe how color is perceived by the brain as a result of visual stimulation.

http://en.wikipedia.org/wiki/CIE_1931_color_space

It's not easy to describe 'saturation' in terms of the linear scale: all 'spectrally pure' light is maximally saturated, and mathematical descriptions of desaturated colors requires statistical models (correlation). That is, spectrally pure light is completely coherent, while desaturated colors (pastels, for example) are partially coherent and neutral grey is completely random.

Does this make sense?

 

[kmm]

A few comments- first, human color vision is based on three populations of cone cells:

http://wearecolorblind.com/wp-content/uploads/graphs.jpg

And thus, any 'pure spectral color' excites multiple cones. This is actually the reason why we can perceive so many hues, even though we only have (colloquially) 'red, green, and blue cones'.

So deep blue light (380 - 450nm or so) excites multiple cones- and as you can see from the above graph, the red and blue cones respond more strongly than the green in that waveband, so that's why we describe purple as a mixture of blue and red.

In any case, the more fundamental point is that 'color' as an attribute of electromagnetic radiation can be most easily quantified either as a linear scale or as a plane surface, and these different representations are not equivalent to each other. The first is used when quantifying the electromagnetic field in terms of it's frequency, while the latter is used to describe how color is perceived by the brain as a result of visual stimulation.

http://en.wikipedia.org/wiki/CIE_1931_color_space

It's not easy to describe 'saturation' in terms of the linear scale: all 'spectrally pure' light is maximally saturated, and mathematical descriptions of desaturated colors requires statistical models (correlation). That is, spectrally pure light is completely coherent, while desaturated colors (pastels, for example) are partially coherent and neutral grey is completely random.

Does this make sense?

Yes, I understand that. My only question at this point is how is violet perceived? "Purplish" or deep blue? Up to now I've seen that a prism shows it as a deep blue but others say they see violet lasers as a purple. As of now though, I'm inclined to say it looks a deep blue as in post #9. I'm suspecting that violet lasers may appear as a deep sort of purple due to some combination of an illusion due to color context and impurities in the light (not entirely composed of 405nm light say).

 

[sophiecentaur]

OK. I wasn't confused that the eye would be fooled when presented with violet, I just wasn't sure what the eye would actually perceive in that case. It's just that some people were saying deep blue and others a "purplish".
On film and tv etc violet appears to be a "purplish" color. So in that case are you saying that is close to what spectral violet would be perceived as? Not a deep blue?

I think you are trying for an answer that is far too definite, here. Colour is only a psychological response and it's 'learned' from experience and the general consensus. Historically, the rainbow was seen and the colours named, by popular opinion. I think that it is very likely to find that the appreciation of spectral violet will be different from observer to observer. That's not surprising as it is right on the band edge of our biological optical receivers. Looking at the far end of the rainbow and try to match it with a colour on a card or RGB display, many people will probably match it with a colour that's well inside the CIE curve. Their brains will be doing what all brains do and that is to find an answer to an unanswerable problem (i.e. matching two unequal things).
There is confusion here because there really is confusion in our colour senses. the same thing even happens with the perception of white and greys and short term colour memory can give us all sorts of confusing colour sensations. If you accept that and don't demand a metrological answer then you can stop worrying about it.
I guess that there will be specialist papers on 'The psychological perception of the colour Violet'. If you can find one, then you may get more sense than PF can yield. Personally, I reckon that anyone who's done a lot of work on colorimetry will not be inclined to investigate this problem because it is too wooly to nail down (as we have found here).

 

[kmm]

I think you are trying for an answer that is far too definite, here. Colour is only a psychological response and it's 'learned' from experience and the general consensus. Historically, the rainbow was seen and the colours named, by popular opinion. I think that it is very likely to find that the appreciation of spectral violet will be different from observer to observer. That's not surprising as it is right on the band edge of our biological optical receivers. Looking at the far end of the rainbow and try to match it with a colour on a card or RGB display, many people will probably match it with a colour that's well inside the CIE curve. Their brains will be doing what all brains do and that is to find an answer to an unanswerable problem (i.e. matching two unequal things).
There is confusion here because there really is confusion in our colour senses. the same thing even happens with the perception of white and greys and short term colour memory can give us all sorts of confusing colour sensations. If you accept that and don't demand a metrological answer then you can stop worrying about it.
I guess that there will be specialist papers on 'The psychological perception of the colour Violet'. If you can find one, then you may get more sense than PF can yield. Personally, I reckon that anyone who's done a lot of work on colorimetry will not be inclined to investigate this problem because it is too wooly to nail down (as we have found here).

I was starting to figure this, so I have to agree with you here. Thanks for all the responses.

 

[sophiecentaur]

I had another thought about this 'violet confusion' and I think it may well be due to the fact that all three sensor responses are well down for spectral violet. (See Fig 1 in this link - and many others). Assessment of colour is made on the basis of the ratios of the three responses. As the 'blue' response curve is dropping fast with decreasing wavelength and the other two are dropping very slowly, the way that the ratios are changing will be subject to a lot of error (small - probably unmeasured - variations of the skirts of the response curves). The nervous system does the best it can with the three values it has to deal with but - for evolutionary reasons - it has not (needed to have) good discrimination in that region. We know that evolution seldom gives us abilities that we don't really need.
Otoh, it is relevant and important to know about what is going on in the red regions of the spectrum, so the three analysis curves seem to be tailored much better to discriminate between reds and near- infra reds. Only yesterday, I was using a blow torch to heat a piece of steel up and I could detect a hint of 'red heat', without confusing it with a deep pink that a pigment could have produced.

 

[DrDu]

An imo inportant point is also that the eyes sensitivity is much lower in the violet than in the blue.

 

[sophiecentaur]

An imo inportant point is also that the eyes sensitivity is much lower in the violet than in the blue.

It's also low at near IR. There is something else, more, to it and I suggested it's probably to do with the two other responses in the tristimulus analysis. I am not aware of a similar confusion around the long wavelengths. I

 

[jim mcnamara]

The simplest answer to color perception questions is that they are cultural. What a native Navajo speaker linguistically classifies as blue or red does not correspond to what most Western cultures perceive. Example: many shades of green are sky color - our idea of blue. (This is a personal observation having learned Navajo). There is a strong color perception component in culture.

Himba people can perceive colors we English speakers literally cannot distinguish. This ability is found only in native Himba speaking people. The reverse is also true, we can perceive some hues they cannot differentiate.

http://en.wikipedia.org/wiki/Himba_people

Try the Anthropological section -- 4.1

I would agree with Andy Resnick - color perception and 'naming' is cultural.

 

[sophiecentaur]

I wonder if much work has been done to relate these different perceptions to the CIE colour chart, trying to make things as objective as possible.

 

[Manwithoutaplan]

Hello,

I want to point out what was said in comment #14
And I remember doing some practica with old optical spectrometers and I too remember violet lines as purpleish.

Secondly, this is not really an issue of cultural confusion.
Dark blue is what you get by mixing black paint with blue paint.
Purple is what you get by mixing red with blue paint.
Whatever the culture, they can call all colors blue, have eyes like mantis shrimps but if they want to make a painting, they will chose one or the other.

I for one, will mix blue and red (and a bit of white but that doesn't matter) to paint violet spectral lines seen trough a spectroscope. And I have become a bit curious where the red comes from.

An other thing, I remember that people with their lenses removed and switched by artificial ones, can see ultraviolet as blue.
For example Monet started painting white water lilies as whitish blue.

 

[kmm]

So it sounds like regardless of what anyone perceives violet to appear as, the explanation for why the sky appears blue and not violet remains the same, correct? The intensity of the violet produced by the sun is less than blue. The high atmosphere absorbs more violet, and our eyes just aren't as sensitive to violet. With the blue cones being stimulated much more strongly, we see blue. So it seems that we would expect the sky to appear blue regardless of whether people perceived violet as a "purple" or deep blue.

 

[sophiecentaur]

So it sounds like regardless of what anyone perceives violet to appear as, the explanation for why the sky appears blue and not violet remains the same, correct? The intensity of the violet produced by the sun is less than blue. The high atmosphere absorbs more violet, and our eyes just aren't as sensitive to violet. With the blue cones being stimulated much more strongly, we see blue. So it seems that we would expect the sky to appear blue regardless of whether people perceived violet as a "purple" or deep blue.

Your point is absolutely spot on. The sky is, in fact, only 'Bluish" rather than 'Blue as all wavelengths arrive at the eye from any point.
The thread has opened out and, consequently, other questions have come to light.
'Naming' and 'perceiving' are two different things. Obviously, in parts of the world where the sea and sky (etc) appear different, so will the local appreciation and naming of other objects by relating their 'colour' to that of natural things.
We can only resolve the question about culture and colour perception from some hard evidence about results of objective colorimetric tests. It's not beyond the realms of possibility that there are differences between the colour sensitivity curves of different races because evolution may (as in the case of skin pigment) have produced differences if there are significant advantages.
In a world where people still (even on PF!) talk in terms of the 'Wavelength' of a particular 'Colour', there is clearly a lot of confusion around. . . .

 

[voko]

I agree that our perception of color is to a large degree degree cultural. Not just socially cultural, but also individually cultural. I am pretty sure that a professional painter would see many colors in the chunk of the CIE chart that I would just call "blue". And I am not even sure that I truly know what violet should really look like, simply because my eyes (and brain) have never been trained for that task. My younger kid has a fascination for painting and pencils in particular (they all have this phase, as far as I can tell), and she demands ever-larger sets of pencils, and then she wants to know the names of new colors. With 5-pencil sets; that was en easy task. Now, as we are above 7, this has become difficult - if the names of colors are not specified on the box!

 

[sophiecentaur]

I agree that our perception of color is to a large degree degree cultural. Not just socially cultural, but also individually cultural. I am pretty sure that a professional painter would see many colors in the chunk of the CIE chart that I would just call "blue". And I am not even sure that I truly know what violet should really look like, simply because my eyes (and brain) have never been trained for that task. My younger kid has a fascination for painting and pencils in particular (they all have this phase, as far as I can tell), and she demands ever-larger sets of pencils, and then she wants to know the names of new colors. With 5-pencil sets; that was en easy task. Now, as we are above 7, this has become difficult - if the names of colors are not specified on the box!

Introducing the idea of colour pigment mixing into this has made it many times harder. Subtractive colour mixing is so much harder to calculate than additive mixing (TV etc). The problem with subtractive mixing is that the Luminance goes down and down, the more saturated you want the colour to be. For starters, at least, it is far better to sort things out in the world of RGB mixing, although we see a lot of coloured objects that are painted with pigments and dyed.
Objective colourimetry at least tries to eliminate some of the problems in this business.
Interesting as it may be, I think that a it's not a good idea to include what 'painters' do with their palette because they do a lot of 'firiggery' with their brushes and often produce fields of colour with stripes of pigments that merge together in a combination of additive and subtractive mixing - very hard to figure out what they are actually doing, except that they can do a good job of fooling your eye!

 

[kmm]

I thought I would add to this and say I got a prism for christmas and tried it on some sunlight shining into my house. I got a good ray of sunlight for the prism to capture and I got a very clear spectrum of colors cast onto a dark wall. On the violet end, it was very clear that violet appeared "purple-ish". Perhaps not everyone perceives it this way but my wife was watching this also and she saw the violet as a "purple". That's two data points anyway :)

 

[sophiecentaur]

I thought I would add to this and say I got a prism for christmas and tried it on some sunlight shining into my house. I got a good ray of sunlight for the prism to capture and I got a very clear spectrum of colors cast onto a dark wall. On the violet end, it was very clear that violet appeared "purple-ish". Perhaps not everyone perceives it this way but my wife was watching this also and she saw the violet as a "purple". That's two data points anyway :)

Home experimenting can be real fun and the subjective effect of a good spectrum in a well shaded place is blisteringly impressive. If you can get hold of a paint box, with a good assortment of watercolour tablets, you could try to get a suitable mix of paints that will give you a good match to your spectral violet. That would be very instructive as you could take the result outside and judge what colour you would call it. You will appreciate that the paint is not producing a spectral wavelength. (Your prism could show you that but the set up would be difficult). Everything's a bit confused around that particular region of colour space. My take on that is that it is not very relevant to survival so we just haven't evolved with the ability to discriminate those colours.

 

[DrDu]

Violet light stimulates both the red and blue receptors in the eye. So while spectral colours include pure blue and red, there is no spectral colour where red dominates over blue. The latter colour would be my definition of purple, which would be one of the few colours which do not occur as spectral colours.
There is plenty of space for experiments, e.g. comparing the spectral colour with some rgb colours on the screen.

 

[sophiecentaur]

Violet light stimulates both the red and blue receptors in the eye. So while spectral colours include pure blue and red, there is no spectral colour where red dominates over blue. The latter colour would be my definition of purple, which would be one of the few colours which do not occur as spectral colours.
There is plenty of space for experiments, e.g. comparing the spectral colour with some rgb colours on the screen.

Mostly Blue end receptor but also Middle and Red end receptors. This is not displays we're discussing. Red dominates over Blue in Spectral Red, btw.
Yes. Displays could be used but they cannot produce colours outside the phosphor gamut.

 

[mainliner]

I have some confusions about this that I'm not finding clear answers for. First, looking at the color violet, it looks like a sort of "purple". The answer as I've understood it to why the sky doesn't appear violet or this "purple" is that it has to do with a combination of the fact that there isn't as much violet produced by the sun and it's absorbed higher in the atmosphere. In addition to that is while violet appears sort of purple since it stimulates the blue cones and the red cones less, the net effect from the different frequencies of light coming into our eye is that the blue cones are stimulated most. So we see the sky as blue. What's confusing to me is that while thinking of violet as sort of a purple, I've come across some explanations that regard violet as a deep blue, not this "purplish" color. To add to this, if you look at a spectrum produced by a prism, it doesn't appear to show violet as this "purple" but a deep blue, however when you look at a rainbow you can see violet as this sort of purple. So I'm not sure what violet is actually perceived as. If violet is perceived as a purple, why doesn't this show up with a prism? If violet is perceived as a deep blue, why do we see "purple" where violet should be in a rainbow? Which is it?

this is a botanical question:)

its something to do with the blue hue hitting the atmosphere and dissipating over plant foliage during summer months, the sun is higher and the plants like blue light during vegetative stage, when the sun goes lower with the coming of autumn the atmosphere let's in more red when the flowers appear ... Its the magic of photosynthesis:)

im not a botanist put its something like this :)

indoor hydroponics growers use 7000k blue for veg and 2000k red for flower ... IM not sure what the mn is .

you've heard the saying " red sky at night Sheppard's delight" :)

that's all iv got hope this helps :)

 

[mainliner]

You allready know this i presume:)

 

[DrDu]

Red dominates over Blue in Spectral Red, btw.
Yes.

Certainly, but in spectral red, green receptors will be stimulated even more than blue receptors, so you won't get the impression of purple.