Light's properties (Why isn't the sky violet?)

[Salles]

Why isn't the sky violet? Since violet has a shorter wavelength even than blue, why doesn't it dominate then?

 

[hmmm27]

Our sun doesn't toss out much violet (also, I think oxygen absorbs some)

 

[Doc Al]

Also, our eyes are more sensitive to blue.

 

[Drakkith]

It's a combination of several things. In addition to what has been said above, plenty of light of longer wavelengths is scattered. Blue happens to occupy a unique position in the spectrum where it is both emitted in very large amounts, is scattered very well, and has a decent response from our cone cells.

Besides. The sky is not blue. At least not spectral blue. It ranges in hues and saturation depending on the time of day and air quality. The only way for it to range in hue and saturation is for there to be appreciable amounts of other colors scattered as well as blue.

 

[Andy Resnick]

Why isn't the sky violet? Since violet has a shorter wavelength even than blue, why doesn't it dominate then?

Part of the reason is that violet is not a spectral hue. Wavelengths shorter-than-blue (ultraviolet spectra) are indeed scattered more than blue, but your photoreceptors don't respond to UV. "snow blindness" occurs from the large amount of UV light scattered from ice/snow.

 

[Drakkith]

Part of the reason is that violet is not a spectral hue.

What do you mean? Violet is 'spectral' as much as blue is if I'm not mistaken.

 

[Vanadium 50]

I don't want to put words in Andy's mouth, but we are moving away from physics and into perception. A fine subject, studies by physicists like Helmholtz abd Young, bit more phisiology than physics.

Color is perceived (over simplifying) as the relative insities of three difference rexeptors at three different wavelengths: kind of yellow-orange, kind of blue-green and kind of blue-violet. Just as it is not possible to have a green blackbody, it may not be possible to have a violet sky.

 

[hutchphd]

What do you mean? Violet is 'spectral' as much as blue is if I'm not mistaken.

We don't see violet very well. But if you are not careful, I will start into a discussion of "indigo" as a color and whether our perception in this region is because of the odd color response as indicated in the CIE standard curves. I find it interesting (consider this a threat !)

 

[Drakkith]

We don't see violet very well. But if you are not careful, I will start into a discussion of "indigo" as a color and whether our perception in this region is because of the odd color response as indicated in the CIE standard curves. I find it interesting (consider this a threat !)

Hold your fire! I just want to know what Andy means by 'spectral hue'!

 

[Andy Resnick]

What do you mean? Violet is 'spectral' as much as blue is if I'm not mistaken.

I mean that no points on the violet-purple line on the CIE chromaticity diagram correspond to electromagnetic frequencies. Alternatively, the electromagnetic spectrum does not contain purples/violets (or pinks and browns, for that matter...)

 

[hutchphd]

But it is listed in the official acronymic name "Roy G.Biv".
So it must be a color, 'cause that is the officially recognized acronym.

Arguing colors is llike nailing Jello

 

[berkeman]

Alternatively, the electromagnetic spectrum does not contain purples/violets

https://www.timeanddate.com/astronomy/rainbow.html

 

[Andy Resnick]

But it is listed in the official acronymic name "Roy G.Biv".
So it must be a color, 'cause that is the officially recognized acronym.

I'm assuming this is a joke?

 

[berkeman]

No, it's a rainbow.

 

[Drakkith]

I mean that no points on the violet-purple line on the CIE chromaticity diagram correspond to electromagnetic frequencies. Alternatively, the electromagnetic spectrum does not contain purples/violets (or pinks and browns, for that matter...)

Violet is absolutely a spectral color, meaning that it can be generated with a small, continuous section of the EM spectrum. Wavelengths from about 380 - 435 nm are generally considered to be in the violet range of the spectrum.

Purple, on the other hand, is not a spectral color. It is a mix of red and blue.

 

[hutchphd]

I'm assuming this is a joke?

I would prefer to call it hyperbole.
We had a previous thread about why are there seven colors.....

 

[cjl]

I mean that no points on the violet-purple line on the CIE chromaticity diagram correspond to electromagnetic frequencies. Alternatively, the electromagnetic spectrum does not contain purples/violets (or pinks and browns, for that matter...)

Violet is absolutely a spectral hue. I have a 405nm laser pointer, and it is absolutely, 100% violet (and anyone looking at it would describe it that way, though it's also kinda wild to see because you don't see pure light in that frequency range that often - most monitors and displays are completely incapable of replicating it). Purple and magenta is non-spectral, as it requires a mixing of blue/violet and red, but violet is recognizable and producible by a single light frequency in the low-400nm range.

 

[Andy Resnick]

Violet is absolutely a spectral hue. I have a 405nm laser pointer, and it is absolutely, 100% violet (and anyone looking at it would describe it that way, though it's also kinda wild to see because you don't see pure light in that frequency range that often - most monitors and displays are completely incapable of replicating it). Purple and magenta is non-spectral, as it requires a mixing of blue/violet and red, but violet is recognizable and producible by a single light frequency in the low-400nm range.

I wouldn't- that wavelength simply appears as an intense blue to me. This diagram is pretty clear- violet does not correspond to an electromagnetic frequency.

https://www.ledtronics.com/html/1931ChromaticityDiagram.htm

There's often a lot of noisy "discussion" about color due to lack of attention to how photoreceptors and brain work. In my experience, the noisiest people are those that never considered color blindness.

 

[Andy Resnick]

Violet is absolutely a spectral color, meaning that it can be generated with a small, continuous section of the EM spectrum. Wavelengths from about 380 - 435 nm are generally considered to be in the violet range of the spectrum.

Purple, on the other hand, is not a spectral color. It is a mix of red and blue.

If the hue is generated by a continuous section of the EM spectrum, it's not a pure spectral color- by definition.

 

[cjl]

I wouldn't- that wavelength simply appears as an intense blue to me. This diagram is pretty clear- violet does not correspond to an electromagnetic frequency.

https://www.ledtronics.com/html/1931ChromaticityDiagram.htm

There's often a lot of noisy "discussion" about color due to lack of attention to how photoreceptors and brain work. In my experience, the noisiest people are those that never considered color blindness.

Then you would disagree with everyone I've shown it to.

Also, that diagram to me pretty clearly shows violet extending out to the bottom left corner, which is a pure spectral hue at ~400nm. Blue only extends down to ~440nm or so, before that diagram starts labeling it as "purplish blue" then "violet" (you can clearly see this since they show the boundary between blue and purplish blue with a convenient solid line).

There's a reason we call shorter wavelength light "ultraviolet" and not "ultrablue".

 

[Andy Resnick]

Then you would disagree with everyone I've shown it to.

That's fine- I am color-blind, after all. How about this- given the sensitivity curves of human cone cells, explain how red hues are is perceived when the retina is excited by 405 nm light:

https://askabiologist.asu.edu/sites/default/files/cones_graph.gif

There's a reason we call shorter wavelength light "ultraviolet" and not "ultrablue".

Yeah, yeah.... whatever. Bees see 'ultrayellow'.

https://www.americanscientist.org/article/the-doors-of-animal-perception

 

[Drakkith]

I wouldn't- that wavelength simply appears as an intense blue to me. This diagram is pretty clear- violet does not correspond to an electromagnetic frequency.

I don't see anything on that chart that says that violet does not correspond to an EM frequency.

If the hue is generated by a continuous section of the EM spectrum, it's not a pure spectral color- by definition.

Forgive me, I was being too precise. There are no single-wavelength light emitters in real life, not even lasers, so I was trying to be as accurate as possible. I should have simply said 'single wavelength'.

Anyways, like I said before, and which is so well supported that I won't bother to make a list of references, violet is absolutely a spectral color. If you think otherwise please provide a reference supporting your position.

 

[Drakkith]

That's fine- I am color-blind, after all. How about this- given the sensitivity curves of human cone cells, explain how red hues are is perceived when the retina is excited by 405 nm light:

https://askabiologist.asu.edu/sites/default/files/cones_graph.gif

The graph is probably wrong, that's why.
In fact, the number of different graphs on the response of cone cells is high enough that I don't know which ones to believe. Here's a few I quickly found from google images, all different:

As you can see, all three of my graphs are different and all three are different from yours as well. However, almost all of them show that the red photoreceptor still responds, at least a little bit, to light all the way down to around 400nm, and at least one graph (the last one in my post) shows a small 'bump' from 380-430nm where the red photoreceptor apparently regains some sensitivity after dropping to near-zero.

 

[hutchphd]

This diagram is pretty clear- violet does not correspond to an electromagnetic frequency.

The reason this is tricky stems IMHO from the fact that the "red" CIE response curve has two humps. One of them is in the violet . I believe this drives the other CIE color map
OOPS I see @Drakkith has beaten me to the punch here

 

[hutchphd]

And do not forget metamers

 

[berkeman]

If the hue is generated by a continuous section of the EM spectrum, it's not a pure spectral color- by definition.

Sorry, but why does my rainbow picture not show clearly the ROYGBIV colors? Or is there some other definition of color?

 

[Motore]

Wikipedia:
Violet is at one end of the spectrum of visible light, between blue light, which has a longer wavelength, and ultraviolet light, which has a shorter wavelength and is not visible to humans. Violet encompasses light with a wavelength of approximately 380 to 435 nanometers. Violet objects often appear dark, because human vision is relatively insensitive to those wavelengths.[citation needed] The reason that to (typical trichromat) humans violet light appears a bit reddish compared to spectral blue (despite spectral red being at the other end of the visible spectrum) is because the S-cone type (i.e. the one most sensitive to short wavelengths) contributes a bit of red to the red-versus-green opponent channel (which at the longer blue wavelengths gets counteracted by the M-cone type).[19] Computer and television screens, using the RGB color model, cannot produce actual violet light and instead mimic it with a shade of purple, combining blue light at high intensity with red light at less intensity.

https://en.wikipedia.org/wiki/Violet_(color)#Optics

 

[pinball1970]

A pf thread from 2014 had a very similar discussion

https://www.physicsforums.com/threads/why-the-sky-is-blue-and-not-violet.783337/

And this one minute physics post by @A.T.

 

[Andy Resnick]

The graph is probably wrong, that's why.
In fact, the number of different graphs on the response of cone cells is high enough that I don't know which ones to believe. Here's a few I quickly found from google images, all different:

Huh- I'll be a monkey's uncle, you are right! I found this reference, very careful measurements:

https://www.sciencedirect.com/science/article/pii/S0042698900000213?via=ihub

(see especially Figs 4-6 and 9)

Clearly showing the L-cone has sensitivity all the way down to 400nm. Violet is indeed a real color!

I always wondered why my (elementary school) art teacher used 2 different words for the same color...