….Arguably our physio-muscular imaginations can conceive of something cycling 200 times a second, but the frequencies of light are far outside any macroscopic physiological processes we can viscerally imagine. It’s also worth noting that while the frequency range of audible sound covers a factor of 1000, (about 10 octaves), the range of visible light covers only a factor of 2 (just one octave).
The differences between our mechanisms of perception of sight and sound are also quite striking. For example, although there is a rough analogy between the pitch of a sound wave and the color of a light wave (since both are related to the frequency of the wave), our perceptual mechanisms for discerning pitch and color are very different. Most people are capable of distinguishing two different accoustical tones, and deciding which of them has the higher frequency, but almost no one can hear an isolated tone and identify its absolute frequency in terms of the corresponding musical note. (This ability is called perfect absolute pitch, and is extremely rare, even among trained musicians). In contrast, nearly everyone has perfect “absolute pitch” for optical frequencies, in the sense that we can be shown a red object and identify it as red, without the need to compare it with any reference color. In other words, we aren’t limited to making comparative evaluations of light frequencies, we experience each color as an absolutely identifiable sensation, with no direction sensation of higher or lower light frequencies. If people are asked whether red has a higher or a lower frequency than blue, they probably don’t know (indeed they might guess red, because red seems like a “hotter” color), and yet they can very accurately recognize red and blue as absolute sensations.
….if we are very familiar with the sight of a red apple next to a green leaf in full daylight, and if we then view this scene in the orange glow of a sunset, both the apple and the leaf reflect different absolute spectra, but to some extent our visual processing infers the shift in illumination and compensates for it, so that we still perceive the apple as red and the leaf as green, even though their spectra at sunset are quite different from their spectra at noon. It’s tempting to make an analogy with how we recognize a familiar melody played in a different key, but in the case of color perception we are not shifting the whole frequencies, we are filtering out a common spectral component from all the elements of a scene.
…..Of course, it’s not strictly accurate to say that colors correspond to frequencies, because most perceived colors actually represent a continuous spectral density profile with non-zero energy over the entire range of visible frequencies, …for typical profiles [of] light that is perceived as the colors blue, green, and red.
These three colors constitute an effective basis for many other colors of visible light, meaning that many (though not all) other color sensations can be induced by some linear combination of these three. By superimposing all of them in equal amounts we get a spectral profile with energy distributed more or less uniformly over the whole visible spectrum, so it is perceived as white light. Other combinations give different color perceptions….
…the spectral density profiles we perceive as pure colors are not, in general, monochromatic. A monochromatic wave has all of its energy concentrated at just a single frequency and wavelength. (In practice it’s impossible to produce a perfectly monochromatic beam of light, but we can come very close.) The dominant wavelengths associated with common sources of blue, green, and red light are 430, 530, and 670 nanometers respectively. Monochromatic light of these frequencies induces the sensations of blue, green and red, even though they don’t have the full spectral densities of typical light with those colors. Moreover, experiments have shown that if we combine three monochromatic beams with those frequencies, the result is perceived as white, even though the energy is not uniformly distributed….. For example, the sensation of pure yellow can be matched by superimposing pure red and pure green, even though this superposition is not “actually” monochromatic yellow.
….the three types of cones are effectively “tuned” to respond to certain absolute frequencies. Thus the signals sent to the brain do not consist of raw amplitudes in time, nor even of frequencies, but simply of the degrees to which each of the three types of cones have been stimulated. As a result, although we have no sense of frequency of optical waves, we can recognize absolutely a range of frequencies (and mixtures) based on the excitation states of the S, M, and L cones. It follows that our sense of color is essentially three-dimensional, i.e., every color we perceive corresponds to some combination of three scalars, representing the degree to which each of the three types of cones is being excited.
…..Given the smallness of these wavelengths and the slight variations between one color and the next, it’s remarkable that the tuning works so well, and is so uniformly accurate over our central field of vision. (Color perception is much less accute in our periferal vision, where rods predominate over cones.) It has been reported that humans can distinguish wavelength differences as small as 0.2 nano-meters. How is it that “red” receptors in one region of our retinas are so perfectly correlated with “red” receptors in other regions of our retinas, and from one eye to the other? And how is it that this tuning remains stable and accurate for decades, and in all different temperatures? It seems clear that psychological compensation processes (like the process to compensate for different illuminants) must be involved.
If our ears contained just a few individual sensing elements, each tuned to one particular absolute frequency, we might all be able to recognize the absolute “color” of audible tones just as well as we can recognize absolute red. However, the ear needs to respond over a much larger range of frequencies, and the dimensionality of the “space” of audible sensation is much greater, i.e., we can distinguish a much greater variety of spectral characteristics of sound than we can of light. Roughly speaking, the coiled cochlea of the human ear has a varying elasticity along its length, so it can be regarded as a series of oscillators of different resonant frequencies, and these perform a fairly detailed spectral analysis of incoming sound waves, transmitting to the brain something a 3000 point spectral profile. The detailed mechanics of how the cochlea responds to stimuli are very complicated, and the study of this function is hampered by the fact that the mechanical properties change significantly if a cochlea is removed for study. Nevertheless, it seems clear that whereas the spectral analysis of optical stimuli has only three dimensions, the spectral analysis of aural stimuli has at least 3000 dimensions. It is not surprising that we (most of us) don’t memorize the absolute sensations associated with tones over ten octaves. Instead, perhaps for more efficient processing, we rely on relative memories of frequencies. The rarity of perfect absolute pitch may also be due partly to a greater variability in the resonance characteristics of our aural sense organs than of our optical sense organs, whose reception frequencies are determined by fundamental atomic absorption properties of certain specific molecules. In contrast, the frequencies of the cochlea are determined by the fluid pressure in the inner ear, and many other factors that could be sensitive to temperature, humidity, barometric pressure, and so on.