Waveform and the perception of light.

[NanoStuff]

I'm unclear on how light is perceived in terms of a pure wave as opposed to a harmonic.

Considering sound, a pure tone has a distinct, agreeably "pure" sound across the spectrum, with recognition to the change in frequency. A sound can also have a harmonic, in this case you have two separate tones that cannot be represented by a single frequency. Such a sound is distinguishably different than any pure tone.

With light however, I am unclear. It would seem to me that a change in wavelength would determine the hue of the color, however this seems to be contradicted by the fact that a combination of red and green produces yellow. These are two separate wavelengths that if I'm correct, together form an impure sine.

In fact the entire visible spectrum can be reproduced using only... seemingly three wavelengths, red, green and blue, but what about all the wavelengths in-between? Does combining these colors not introduce harmonics but instead shifts the wavelength? And if so, how does this explain saturation? White is a combination of "all" colors so they say, but what is the waveform of all colors? You cannot, for example, fit every audio frequency in a sound, there are certain smallest-intervals between two frequencies. How would white be represented on a spectrum analyzer?

One thing I'm sure of is that the solution is very simple, nevertheless it does elude me because there is surely one little thing I'm not taking into proper consideration.

 

[NanoStuff]

Perhaps I should have checked Wikipedia first:

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

Seems to explain it fairly well, thought there's a bit to get your head around. It apears the eye is incapable of distinguishing between a pure wavelength and a combination of wavelengths between which the equivalent pure wave would sit... more or less. This sure opens my eyes, can't begin to imagine what sensory perception we are missing by not being able to distinguish between the two.

 

[CycloWizard]

It's due to the way phototransduction happens. There are only a few photoreceptive molecules, each of which has an absorption spectrum different from the others. I'm not that familiar with color perception, so I'll use the example of rods instead of cones.

Rods are covered with proteins connected to light sensitive molecules (an 'opsin'). When a photon hits the opsin, it isomerizes and releases energy. The energy is not necessarily that of the photon, as the isomerization is a form of transduction: energy equal to h*nu (Planck's constant multiplied by the frequency) is input by the photon, then less than that amount of energy is output in the form of an electrical potential increase. If enough photons hit a single cell in a short enough time period, the minute increases in voltage are temporally summed and depolarize the rod's membrane, resulting in an action potential.

In humans (and most animals), the maximum sensitivity in rhodopsin (the opsin found in rods) corresponds to 'green' photons (can't recall the wavelength specifically), so it takes fewer green photons than any other color to induce an action potential.

In cones, three different opsin molecules with more wavelength specificity (i.e. a narrower absorption band) allows differentiation between the wavelengths. If cone A is packed with red-sensitive opsins and is struck by a purple photon, not much happens. The brain correlates the output from certain cones with a certain color, dictating how you perceive that color.

Most of this is probably explained better in the Wiki article, but hopefully that helps a little.

 

[MrDudeMan]

Originally posted by: CycloWizard
...
In humans (and most animals), the maximum sensitivity in rhodopsin (the opsin found in rods) corresponds to 'green' photons (can't recall the wavelength specifically), so it takes fewer green photons than any other color to induce an action potential.
...

which is exactly why digital camera sensors are all starting to include 2 green photodiodes to every 1 red and blue

 

[CycloWizard]

Originally posted by: MrDudeMan
which is exactly why digital camera sensors are all starting to include 2 green photodiodes to every 1 red and blue

Right. There is a new one that includes a receptor for each color (RGB) on each pixel now too - Spartan-3 IIRC (unrelated, but cool ).

 

[Paperdoc]

Aside from the receptors, the light source characteristics are vital. Almost all colored light sources emit light over a broad range of frequencies - for example, the three phosphors used in TV picture tubes. "Green" is not just one green, but a large range of frequencies over a somewhat bell-shaped distribution curve. Likewise, an object with a color actually has absorbers in it that absorb selected ranges of frequencies, leaving in the reflected light the other frequencies we "see" as its color. What we see is the combined impact of the broad range of light (not necessarily balanced white) source, modified by the absorbers in the objects. The perception is our brain's interpretation of the broad overlapping color curves. We will label one single color with a simple word (like, green) or with more slective word groups (like, Emerald Green), but in fact we can distinguish very small differences in appearance that are based on small differences in the details of the whole color spectrum - that is, in the details of the graph of light intensity versus frequency.

If you were to try to duplicate all visible colors with only a very few pure single-frequency light sources (say, red green and blue lasers) you would find it impossible to create every color. But using broad-frequency-range light sources (or absorber dyes), you can reproduce a huge range of colors, even though it will NOT be ALL visible colors.

 

[NanoStuff]

So it would seem that an RGB monitor, for example, would be limited in it's ability to present a pure Cyan tone because the Green light would interfere with the Red color receptor. So the end result would be a color of reduced saturation.

Has anyone invented a full-gamut diode yet?