# What does the "true" visible light spectrum look like? [closed]

When I google "visible light spectrum", I get essentially the same image. However, in each of them the "width" of any given color is different.

What does the "true" visible light spectrum look like, then? It can't be that each and every image search result is correct.

I could not find any information about this on the web, so I turn to the experts.

• It is not clear what you mean by the visible light spectrum. Do you mean the spectrum that comes from the sun? Or are you referring to the sensitivity of the human eye to the wavelength spectrum? Feb 11, 2021 at 2:41
• An important concept here is gamut Feb 11, 2021 at 3:11
• en.wikipedia.org/wiki/Cone_cell#/media/… Maybe this image? nything with a wavelength x axis in nanometers should be good. A lot of it is taken up by red. Feb 11, 2021 at 3:55
• Seems like the author of that image was careful to try to normalize things. OP should find that image useful. Feb 11, 2021 at 4:15
• Even without getting into things like 'your red is my green' a brief perusal of color-based optical illusions (what color is this dress!) shows that there's a ton of things going on behind the scenes in the brain beyond just relaying the raw data. It's quite possible the answer to this question is slightly different for each person -- I doubt you can find two people who react exactly in the same way to identical optical stimulus.
– eps
Feb 11, 2021 at 17:14

Most computer monitors aren't capable of displaying any spectral color. Some of the RGB monitors could display at most three of them: some red wavelength, some green and some blue. This is because the gamut of the human vision is not triangular, instead it's curved and resembles a horseshoe:

In the image above, the black curve represents the spectral colors, with the wavelengths in nm denoted by green numbers. The colored triangle is the sRGB gamut, the standard gamut that most "usual" computer monitors are supposed to have.

As you can see, the black curve doesn't even touch the triangle, which means that sRGB monitors can't display any of the corresponding colors.

This doesn't mean that you can't see any good representation of the visible spectrum. You can e.g. display what the spectrum would look like if you took a gray card and projected the spectrum onto it, thus getting a desaturated version. CIE 1931 color space, via its color matching functions, lets one find, for each spectral color, corresponding color coordinates $$XYZ$$, which then can be converted to the coordinates in other color spaces like the above mentioned sRGB.

The inability of the sRGB monitors to display spectral colors manifests in the fact that, after you convert $$XYZ$$ coordinates to sRGB's $$RGB$$ ones, you'll get some negative components. Of course, negative amount of light is not something a display device can emit, so it needs some workaround to display these colors (or something close to them). Displaying the spectrum as projected on a gray card is one of these workarounds.

Here's how such a desaturated spectrum (with a scale) would look:

To get this (or any other, actually) image to display "correctly", ideally you need to calibrate your monitor. Some of the consumer devices have better color rendering out of the box, others have quite poor color rendering and show visibly wrong colors. If you don't calibrate, then just be aware of this nuance.

Also, if you happen to be a tetrachromat (virtually never happens in males, rare in females), then the above image will look incorrect to you in any case.

How to see an actual spectrum, without the workarounds discussed above? For this you should use not a computer monitor. Instead you need a spectroscope. These can be found in online stores like AliExpress quite cheap, some using a diffraction grating, others a prism. The ones with grating will give you almost linear expansion in wavelengths, while the ones with a prism will have wider blue-violet part and thinner orange-red part of the spectrum.

• Comments are not for extended discussion; this conversation has been moved to chat. Feb 13, 2021 at 0:40

If you're really curious, buy a cheap prism, and take it outside in sunlight.

You'll be dispersing the frequencies present in sunlight, and in addition, your eyes are more or less sensitive depending on the frequency, but that's a good start for being able to see what a "real spectrum" of visible light is.

A monitor does not produce all frequencies of light, but rather tricks human perception by sending different proportions of red, green, and blue light. A color-calibrated, wide-gamut display can reproduce the effect of broad spectrum light, but it won't be the real thing.

• An even cheaper option is a simple water prism, which you can make with a shallow dish of water and a mirror small enough to lie in the dish at an angle, so that the water above the mirror forms a wedge shape. Water has a lower refractive index than glass, but it has higher dispersion (i.e., a greater relative difference between the speeds that red & violet travel through the medium). It's easy to make quite large spectra with a water prism. It can be helpful to have another mirror to direct the sunlight into the water at a good angle. Feb 11, 2021 at 3:25
• What, then, is the best computer representation of the visible light spectrum?
– Alex
Feb 11, 2021 at 3:51
• First you have to calibrate your monitor. photographylife.com/the-basics-of-monitor-calibration At that point, colors within the gamut triangle will be accurately represented. Then, generate an approximation using e.g. methods in stackoverflow.com/questions/3407942/… Feb 11, 2021 at 4:04
• With a prism you'll get quite a crude representation of the spectrum: basically a convolution of the actual spectrum with the box function corresponding to projected width of the prism. Better use a spectroscope. Feb 11, 2021 at 7:38
• @PM2Ring ..., add a case and thus get a DIY spectroscope :D Feb 11, 2021 at 9:32

When you look at images on your computer, you’re seeing light emitted from the LEDs (most likely) in the pixels of your screen. Those pixels generate colors by combining the output of LEDs with only a few different center wavelengths. Want purple? Mix blue and red. Want a slightly different purple? Adjust the proportions of blue and red. Your eyes and brain interpret these combinations of colors as different single colors.

The point is: Your screen is not giving you pure colors of the rainbow. It simply cannot. It’s not designed to. It’s designed to generate colors with linear combinations of other colors. Any generation of a color in the visible spectrum will be an approximation. However, one benefit of this is that your screen can generate colors which do not exist in the rainbow, like pink.

So all the images of spectra you see are different approximations of the actual spectrum you’d see if you put white light through a prism. That’s why they don’t agree. It’s impossible to find exactly what you’re looking for coming from your computer screen.

• What, then, is the best computer representation of the visible light spectrum?
– Alex
Feb 11, 2021 at 3:51
• @Alex is this question just a curiosity, or is there some application behind it? If you really need an accurate rainbow, generate one on your desk directly with some sort of prism. Then take a picture of it and adjust the color values pixel by pixel so that the one on your monitor best matches the one on your desk. Feb 11, 2021 at 5:08
• "Want purple? Mix blue and red." No, that's mixing paint, not light. Feb 11, 2021 at 22:17
• @philipxy purple is absolutely a mix of blue and red light [in fact, it does not exist at all as a pure wavelength]. You may be reaching for a distinction between "purple" and "magenta", but purple is really just dark magenta [and brown is dark orange], and in terms of light it's simply... less light. Feb 12, 2021 at 7:04
• @philipxy Rather than just disagree, here's a nice question/answer from this SE with more detail: "So the way to make light that appears purple is to mix the 400nm violet light (or even blueish light with a slightly longer wavelength) with the 700nm red light (or even orange light with a slightly shorter wavelength)." physics.stackexchange.com/questions/122601/… Feb 13, 2021 at 0:38

Others already pointed out the effects of sensors and pigments (or emitters) that cannot perfectly mimic the response of the (standard) human eye.

So one would need to look at a real spectrum. The excellent answer by Jonathan Jeffrey is to point to a prism. But there one has the problem that the dispersion increases towards the ultraviolet, and that relative widths of colour bands would be slightly different for different materials of the prism.

Maybe more "true" would be to look through a diffraction grating. Cheapest solution is to use the reflection in a CD-disk (or the transmission when the metallic layer of a recording CD is removed).

• Just use a grating spectroscope, or you'll get anything from white with color fringes on the sides to a poor-resolution spectrum... Feb 11, 2021 at 9:02

However, in each of them the "width" of any given color is different.

That is because the images were recorded with different methods and materials.

What does the "true" visible light spectrum look like, then? It can't be that each and every image search result is correct.

"True" needs the measurement of the frequency on the x axis and the intensity on the y. It is the intensity that makes the difference you observe in the apparent widths of the spectra. That intensity is a function of the lattice spaces of the atoms of the spectrum analyzer used in taking the picture.

You are looking at the spectra with your eyes, and color perception, the color observed by our mind and defined as a given one, is a complicated effect. See the answers to this question.

The spectrums in addition to being recorded on different media, are perceived by you brain too, and that depending amplitude of the different frequencies with the method the spectrum was recorded, will also introduce a width in you perception of the colors.

...the "width" of any given color is different

This is to be expected, since the unit of the horizontal axis might be different. Two separate bands that are equally wide on a frequency (energy) scale will always have different width on a wavelength based chart.

When you look at a spectrum, you're using some kind of optical effect to spread out the different frequencies of light. Most commonly this is done by a triangular prism, where the difference between the speed of light in air and the speed of light in the prism material ('refractive index') causes different frequencies of light to bend by different amounts. Generally, lower-frequency red bends least and higher-frequency violet bends most, with the other colors spreading to varying amounts between the two.

Depending on the refractive index difference, the amount of bending changes. One kind of prism (say, clear plastic) might be rather weak so that the difference in angle between red and blue are very small, producing a very narrow rainbow, while another prism can use a better material (such as a diamond) and spread red and blue onto very different angles, giving you a wide rainbow. Both of those are real and valid spectra. Neither is more "correct" or "real" than the other; it's just different materials spreading out the light to different amount.

And then of course the actual size of the spectrum produced depends on how far the target wall is from the prism. A prism that produces a very wide spectrum with a wall that's only two inches away will have a narrower resulting image than a less powerful prism projecting on a wall ten feet away.

This question seems bit like looking at an 18" television and a 50" television and asking "But which one is showing the real TV signal?" The question doesn't really make sense.

Look at a piano keyboard. The space that each octave takes up (a C key to the next C key) is the same. This is actually logarithmic spacing. If you spaced the keys linearly according to their frequency, the keys would be spaced much further apart in the higher octaves. Each C note actually has twice the cycles per second as the previous C note. So some of the images use a linear scale and some use a logarithmic scale. All the images that show colors as rectangular blocks are actually artist's conception as the frequency of the color varies continuously. So these images are really only suggestions to help in understanding the concepts.

Light comes to the eye as photons. Every photon has a specific frequency. Your eye contains multiple types of cells which each respond differently based on their frequencies. A given object will send photons at various frequencies, often represented as a graph with wavelength as the x-axis and intensity on the y-axis. (See https://en.wikipedia.org/wiki/Spectrophotometry) When the combination of frequencies hits the cells in the eye, it is converted to an intensity for each of the types of cells. (red, green blue, and one covering the whole visible spectrum.) Multiple combinations of frequencies can yield the same four values for the nerve cells in the eye. Each combination of the four values is associated with a "color" by the brain. Therefore, multiple combinations of frequencies can appear to be the same to a person.

So

• Visible light spectrum is simply the range of electromagnetic frequencies that can be detected by the eye. The illustrations are mostly artist's conceptions to help understand the concept.
• You can take a beam of light and use a spectrum to split it according to frequencies. Each dot on the image generated by the prism will be composed of photons at a single frequency.
• Human eyes, photographic film, and electronic cameras have receptors of multiple types with each type having a sensitivity that varies with the frequency of the photons. The values from each of the types of sensors is sent to the brain, where it is converted to a "color".
• If you take an image of the spectrum produced by splitting light
through a prism, and then examine the image on a monitor, the color
of each point will appear as a combination of red, green, and blue.
• However, if the frequency sensitivity distribution of the types of cells in the human eye match the distribution for the types of pixels in the camera and everything is perfectly aligned, the color that the human brain obtains by using the prism and viewing the electronic image should be the same.
• If the frequency response of the types of cells in the eye did not match the frequency response of the pixels in the camera, the colors would appear to be wrong. Butterfly eyes have different response curves, so a computer display's image would look wrong to them.