24

Different kinds of white light have a different spectrum.

Light from a white LED will have blue at the peak intensity while white light from a CFL or something else will have a different looking spectrum.

I don't understand how this works. Shouldn't pure white light have a unique spectrum no matter what?

For example, a certain white LED's spectrum looks like this:
$\hspace{100px}$.

Nat
  • 4,640
Black Dagger
  • 1,283

4 Answers4

41

Shouldn't pure white light have a unique spectrum no matter what?

White is not a spectral color. It's a perceived color.

The human eye has three kinds of color receptors, commonly called red, green, and blue.

$\hspace{100px}$

Note that there's no receptor for yellow. A spectral yellow light source will trigger both the red and green receptors in a certain way. We see "yellow" even though we don't have yellow receptors. Any spectrum of light that triggers the same response will also be seen as "yellow". Computer screen and your TV screen manufacturers depend on this spoofing. Those displays only have three kinds of light sources, red, green, and blue. They generate the perception of other colors by emitting a mix of light that triggers the desired response in the human eye.

What about white? White isn't a spectral color. There's no point on the spectrum that you could label as "white". White is a mixture of colors such that our eyes and brain can't distinguish which of red, green, or blue is the winner. Just as any mix of colors that trigger our eyes and mind to see "yellow" will be perceived as "yellow", so will any mixture of colors that trigger the same responses in our eye.

Aside:

There is a mistake in the above image in the labels "bluish purple" and "purplish blue". That should be "blue-violet" and "violet-blue" (or possibly "indigo"). Purple is a beast of a very different color. It is a non-spectral color. The spectrum is just that, a linear range. Our eyes don't perceive it as such. We view color as a wheel, with blue circling back to red via the purples.

Nat
  • 4,640
David Hammen
  • 41,359
  • Suppose the light is spectrally flat — what colour would we perceive? – gerrit Jul 30 '14 at 15:26
  • @gerrit, I guess gray – Antonio Ragagnin Jul 30 '14 at 15:42
  • 13
    @gerrit, that would depend on many things. The human visual system does not create a static mapping between a particular spectrum and a single perceived color. Instead as lighting conditions and spectra change, our visual system re-evaluates and changes perception. It allows us to see "white" things in broad daylight as well as evening twilight. As such, a "flat" spectrum would be interpreted depending on what else was seen close to it in time and space. There is no context-independent answer. – BowlOfRed Jul 30 '14 at 15:49
  • @David Hammen You mean to say that our eyes are insensitive to a range of light-combinations that it identifies all the combinations as white? – Black Dagger Jul 30 '14 at 16:03
  • @BlackDagger - Our eyes aren't spectrometers. There's only so much our eyes can do with but three kinds of color receptors. Look at the image you posted. That's anything but a white noise spectrum. You see that light as white because it excites each of those three kinds of receptors similarly to how light with a much flatter spectrum would excite them. There's also a lot of signal processing going on in your eyes and brain. Your nice summertime white clothes look white at noon and at sunset, but point a spectrometer at them and you'll see rather different spectra. – David Hammen Jul 30 '14 at 17:31
  • 12
    This is the basis of this classical optical illusion. Perceived colour is (strongly) context-dependent. – Emilio Pisanty Jul 30 '14 at 18:01
  • 7
    @EmilioPisanty - Here's another optical illusion of the same ilk. – David Hammen Jul 30 '14 at 18:47
  • 1
    @DavidHammen - I love that one. When you put your finger over 90% of the center, and then shift it slightly so that it covers the whole center and then not, you can almost watch your eyes reevaluating the colors in regards to each other. Or at least I can... – Bobson Jul 30 '14 at 21:11
  • 4
    White, gray and black are actually the same color. Just like bright red, red and dark red are "reds". Gray is just dark white and black is very dark white. You're just changing the luminosity. Not the hue or even saturation. – slebetman Jul 31 '14 at 04:12
  • 1
    If you're mathematical: take the curve that represents the spectrum of the light source, and multiply it at each wavelength with the curve that represents the sensitivity of a given kind of cone, then take the integral under the product curve. Repeat for each cone type, and you have the LMS tristimulus values. Two sources that give the same LMS will appear the same. The curves (spectra) don't have to be the same, just the integrals that come out of them. – hobbs Aug 01 '14 at 04:57
  • decent answer, but what about tetrachromes? – Code Jockey Aug 01 '14 at 13:01
22

White Light is Just an Illusion

Pretty much any light source which is emitting "white" light has a different spectrum. There are a few reasons: blackbody radiation and the method of producing light. When our eyes see radiation of several different wavelengths at once, we can combine them and make them "white" or "pink" or whatever color it is.

Blackbody Radiation

Filament bulbs (and stars!) produce light via blackbody radiation. It just so happens the curve for these things fits into the visible spectrum well, and we register it as "white" light. This is, however, why filament bulbs have that slight yellow quality.

Light of Varying Wavelengths Team Up to Make White Light

If you take the spectrum of different lamps, you'll see they are composed of different combinations of wavelengths. They all produce what is called "white" light, but they each have different wavelengths of varying magnitudes. This results in different light bulbs having different qualities of "whiteness" or "color" to them, and is something interior designers worry about. ("Yes, that fabric looks good now, but what about in artificial light?")

White LEDs are Hardly LEDs at All!

For instance, that LED isn't a "white" LED. It's a blue one which drives a filament. The actual LED is why you get the nice peak in the blue area, but then the filament is producing that "blackbody radiation" curve after it. It's why white light from an LED looks blueish; it is emitting more blue light than other light. A true white LED would have three spikes in its emission spectrum, and would likely be composed of three different LEDs.

Nat
  • 4,640
PipperChip
  • 2,208
  • 6
    It's not blackbody radiation around blue peak of the LED. It's luminescence spectrum of the yellow phosphor filling the rest of space around the crystal. – Ruslan Jul 30 '14 at 15:52
  • @Rusian It's what I meant, but I'll edit for clarity. – PipperChip Jul 30 '14 at 16:23
  • The color of blackbody radiation depends on the temperature of the emitter. A hotter filament will look bluer. A cooler one will only appreciably glow in the infrared. –  Jul 30 '14 at 17:12
5

When we talk about "white light" we're usually referring to sunlight, which has a blackbody spectrum with a temperature about 6000 kelvin. But since the human eye has only three types of color receptors, it's possible to produce the same "white" response in the eye with a carefully chosen combination of line spectra.

rob
  • 89,569
  • 2
    And at the store you can buy bulbs with different equivalent black body temperatures - they will look different (particularly to different people), but they are still "white". Human perception is more than just the response of the color receptor - there is a lot of processing as well going on, and it accepts a fairly broad range of "white". – Jon Custer Jul 30 '14 at 14:27
  • 1
    Note that blue light is preferentially scattered away from the sunlight reaching the Earth, so the colour of the direct sunlight we seeing from Earth is not the colour we would see from space. – gerrit Jul 30 '14 at 15:25
  • I have always considered sunlight to be slightly yellowish and I think a lot of people here feel the same. I guess we grew up with CFL light which changed our concept of "white". – slebetman Jul 31 '14 at 04:17
  • 2
    @slebetman Some sunny day, use a pinhole camera to make an image of the sun's disk. (If you feel extra-motivated, use a flat mirror to send the sunlight horizontally about 50 feet into a dark alcove or interior wall, and use a pencil-sized pinhole to get enough detail to see sunspots on a six- to eight-inch image.) Put your image on a white backdrop. I dare you to call that color yellow! Sunlight is white. – rob Jul 31 '14 at 12:47
  • 1
    +1. Sunlight is white when the Sun is not near the horizon. It's not a good idea, but you can briefly look at the Sun briefly when it is near the horizon. You will see a reddish or yellowish Sun at such low elevations. You can't look at the Sun once it has climbed into the sky. Reflexes force you look away. You would see a brilliant white ball (not red, not yellow, just white) if you could look at the Sun when it is high in the sky. – David Hammen Nov 17 '14 at 18:17
2

There is no such thing as "pure white light". White is what your brain perceives when you see a broad, reasonably balanced mix of wavelengths. White light is, by definition, a mixture of wavelengths so it cannot be "pure".

David Richerby
  • 656