Imaginary Mark

Good point. A color can't be represented by a single number representing a point on the spectrum. It is the intensity at each point on the visible part of the spectrum. It can be represented as a graph with electromagnetic frequency on the X axis and intensity on the Y axis.

Christopher Lord

Our three colour-sensitive cell-types have overlapping sensitivity. When a certain wavelength arrives (say orange), all three cell types activate to some degree or another. RGB displays exploit this property of our eyes to fake the appearance of the full EM spectrum. Interestingly, most people can see into the infrared and the UV since our red and blue cells, respectively, are slightly sensitive in those ranges. consider what a black light looks like, for instance. or consider what a stove element looks like when it first starts to glow.

As to the original question, yes there are many (nearly infinite) frequencies in our visible range. but as has been mentioned, the question is actually the number of discernibly distinct colours. 32-bit RGB displays are really good in this respect. In a given brightness context, the visible difference between two adjacent colours is tiny. Since we are most sensitive to green, expect to be able to see colour steps best in this spectrum. Since RGB displays use a trick to mimic the full EM range, it can not represent fringe cases -- for example, near-infrared or near-UV. Again, consider photographs of stove elements or UV lamps. Each photograph would not do the original scene justice at all.

But 256 steps per channel is insufficient when displaying images with wide dynamic range. The human eye can adjust exposure, so we can see fine by moonlight, daylight, or anything in-between. We can even see such vast contrast differences in one scene, all at once (a sunset, for example). But a monitor displaying a photograph of a scene with vastly different dark and light regions looses the ability to show detail in both the light and the dark region. This is because all the dark stuff has to cram down near values 0-10, and all the bright stuff has to get pushed up into 240-255. So the image has only 10+15=25 different colours.

This is particularly bad for photographers, who frequently adjust colours after shooting. If we had only 25 colours in a photograph, there is very little room for adjustments (for example, to bring out detail in the shadows). So we typically shoot in a much higher bit-per-colour mode (I shoot 24-bit-per-channel, then work in 16-bit-per-channel), then downsample to the final colour range when everything looks good.

This is all to work around the limitations of 8-bit-per-channel limit of most computers. I wish everyone had 16-bit-per-channel monitors. it would make for games with really realistic transitions between indoor and outdoor scenes, for one thing. Imagine stepping out into the sunlight and having to wait for your pupils to dilate!

Karalora

I remember when I first realized that I could see UV light. I was looking at some really dark purple flowers and realized that they were glowing despite being only a few shades shy of black. The glow was identical to what you get from a high quality black light.

youngalexander

What is the frequency of 'white' light?

Colour is a complicated and partially subjective area of study. True, there is the physics of electromagnetic radiation in all its detail. The colour of an object results from the reflection of incident light. It will depend upon the frequency mix of that light and the absorption. It is also important to realise that color is a matter of impression. For instance, some colours will appear brighter than others when projected with equal intensity.

Colour has three properties; brightness, hue and saturation. Mageth has already mentioned mixing of frequencies and Chris Lord television & RGB. A very good exposition is given by Wiki Color

The answer to the question re White is illuminating.

Nostalgic Pushhead

AFAIK a new Halflife 2 add-on/extension/mod/etc Lost Coast has exactly this function- "High Dynamic Range Lighting".

fließendes

Another fun way to test yourself: color converter. (Note: you must be a nerd in order to go to this site for entertainment).

I think the 24-bit colors are easiest to discern between red (R=255,G=0,B=0) and yellow (R=255, G=255, B=0). I can't tell the difference one increment at a time:_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

Or even at increments of 5:
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

If it were printed color, it might be different.

I can just barely perceive the change when you get to increments of 10:
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

And I consider myself pretty attuned to visual nuances. But maybe it's a test of how evolved you are. :Cheeky:

Malachi151

First of all, there is no such thing as "a color", and the "visible light spectrum" is literally different for every single living individual organism in existence.

All that there are is wavelengths of light. Our cells detect those wavelengths and then send signals to our brain where the brain then creates a perception to represent the wavelength in a model.

The "real world", however, is devoid of color, and every other sensation for that matter. There is no such thing as smell or sound or feel, these are all representations in our unique mental models. Every single organic being in existence has a different model and a different representation for external phenomena.

"Red" is not a property of the light wave, it is a legend in our mental models, and it is relatively common among human individuals because we share the same evolutionary history and thus mostly the same genetic code, which means that we mostly models the world in the same ways, but due to the slight variations, every model is just a little different.

Same with smell. "Why does shit stink?" It doesn't stink. Molecules from shit, like every volatile substance, become airborne where our noses are able to detect them. When the molecules bind to receptors in our noses the nose sends a signal to the brain telling it what receptor has been triggered. The brain then creates a sensation to model the data.

Natural selection has "guided" (I use this term loosely here) the evolution of humans such that humans in which a negative sensation was generated when "shit molecules" were detected were selected for because this negative sensation had the effect of repelling the individuals away from the shit. Since shit carries disease this led to a higher survival rate than among those individuals who were ambivalent towards shit or who were attracted to it.

Conversely, among flies, organism that are not susceptible to the same diseases that reside in the fecal matter of vertebrates and who can survive on the existing organic molecules in vertebrate fecal matter, the molecules emitted by shit likely cause a pleasant sensation in flies in order to attract them to the resource.

Santas little helper

So does anyone know if there are infinitely many frequencies in the visible spectrum ?

doc_simon

Does the encoding from the retina to the visual part of the brain limit our colour resolution in anyway? AFAIK it's based on frequency of the spiking output from neurons - which would suggest that for small points of light our ability to distinquish colours would be limited.

Any thoughts from the neuro-phys people?

ninewands

Computer monitors are limited to roughly 24-bit color depth because of the limited color resolution of the phosphors used in the screen. This works out to about 16.7 million colors that can be rendered on a CRT screen. Film, on the other hand, has a limiting color depth of approximately 256 bits per primary, or 768 bits of color depth (for a maximum number of discrete colors of about 7.76*10^230 [if I didn't miscount the digits in bc] ). Film's higher color resolution comes from the fact that the "pixels" are not fixed size, thus varying the concentrations of the dyes that substitute for the metallic silver during development over a wider range.

I don't know if any statistical study has ever been done but I seriously doubt that the color resolution of the human eye is anywhere near the resolving power of film. I suspect that it is, at best, no more than about twice the resolving power of a CRT ... maybe 8 times max (192 bit color depth, total). I know that I can just BARELY discern a difference of plus or minus 1 in the intensity of colors on an 8 bit per primary scale (24-bit color)).