That image comes from an 1855 book, and shows purple as a prismatic color.
Opticks, 4th edition by Sir Isaac Newton, Knt. is available on-line. It is pretty clear on Newton’s use of purple
You must not get very bright rainbows where you live. The brighter they are, the more colors you can see. I’ve seen rainbows where I could pick out several distinct shades in between green and yellow and between yellow and orange.
“The originall or primary colours are Red, yellow, Green, Blew, & a violet purple; together with Orang, Indico, & an indefinite varietie of intermediate gradations.”
With spelling like that, you trust Ike? 
A spectral color is a single wavelength of light that stimulates one or more of the cone cells in the retina. A composite color is a mixture of two or more wavelengths striking the retina together, giving the impression of a single color.
To use a musical analogy, violet (a spectral color) is a pure note, and purple (a composite color) is a chord.
Yellow is a spectral color that stimulates the red- and green-sensitive cones. However, a mixture of red and green light can also stimulate those cones in the same way, creating a perception of “yellow” in the brain even though no yellow light is present. Red and green spotlights are combined into yellow lighting in theatre productions, and red and green pixels create the illusion yellow on video screen. So yellow can also be a composite color.
Opticks 4th edition, published posthumously 1730, contains no such quote (the word “gradation” appears not at all). Your quote comes from a 1672 Draft of ‘A Theory Concerning Light and Colors,’ fifty years earlier.
He was mixing colors in 1672 but may not have stumbled onto red+violet=purple so didn’t need to use ‘purple’ in a special sense then. Setting up prisms in just the right way was tedious — even for this genius, perspiration may have been key to inspiration.
[QUOTE=Isaac Newton in his 1672 draft]
The same colours in Specie with these primary ones may be also produced by composition: ffor a mixture of yellow & blew makes green, of red & yellow makes Orang of Orang & yellowish green makes yellow. And in generall if any two colours be mixed, which in the series of those generated by the Prism are not too far distant one from another, they by their mutuall alloy compound that colour which in the said series appeareth in the midway between them. But those which are situated at too great a distance doe not soe. Orang & Indico produce not the intermediate green, nor Scarlet & green the intermediate yellow. [my emphasis]
[/QUOTE]
(I’ll guess that his Orang-Indico mix gave him a grayish brown he deemed uninteresting — he hadn’t yet discovered purple.)
Some of the confusion about “purple” might go away if magenta were used as the name for red+violet. Perhaps Doper cmyk should weigh in! :rolleyes:
I find this question baffling. Are you claiming that the spectral order of colors is common knowledge? I’d be surprised if one person in ten among the general population could accurately name them in order. Why would the ROY G BIV mnemonic be necessary if everyone knew this?
–Mark
I’ve asked the question about why red+blue results in the sensation of a more engergetic color than you started with. I didn’t quite get a definitive answer, but this is what I took away from it:
Spectral violet actually somewhat activates the red cones in our eyes. The best guess for why is an “octave effect.” An octave is a doubling of the wavelength of sound. And sounding a higher octave, will cause a string of the lower octave to sympathetically vibrate. Something similar may happen with the eyes, as violet correponds to almost double the frequency (half the wavelength) of red.
Of course, we can simulate this sympathetic vibration by adding a little red to something that is blue. And that is how systems such as RGB produce purple.
You’d think that this would have to be built into cameras, so that they would convert violet to purple. And the CIE chart, with its extra red hump towards the deepest violets, seems to indicate that this is the case.
I doubt it is an octave effect - although it is a neat thought. It isn’t on an octave, and for the most part optics doesn’t really work that way.
I don’t think I have seen this built into cameras - although it certainly could be. We survived for many years with colour film, and that has all sorts of issues.
One thing is - as we have discussed earlier - there is very little actual violet out there. You would need to find a light source that was close to monochromatic violet. Apart from a rainbow you are not going to find one outside of a lab (and as Chronos notes, even the rainbow is significantly mixed.) So cameras cheerfully just deliver purple when they see purple, and very pure blue in the incredibly rare times they see pure violet.
Indeed, any sort of monochromatic light almost always causes cameras to fail in their colour representation. The best known examples are white light sources that are made from a set of line spectra. Cameras have a terrible time trying to get the colours of things illuminated with such sources to match what you eye perceives.
The thing about the CIE chart is that it represents colours as we see them. Colour is an eye/brain artefact. There is really no such thing as an intrinsic colour outside of this.
And white vs. black or luminance, which happens to rely on the “red” AND “green” content, but not so much the blue.
I’m certain that I could not identify an indigo plant, but I like the word because using it makes me feel like a pirate. We have a shipment of rum and indigo, matey!
Snip. Some color systems are cyndrilical, and it sounds like you stumbled across that empirically. For example, HSV/HSL (subtle difference) puts Hue on the circular part, Saturation is out from the center (as if cutting it pizza-style), how mixed it is with gray, and cut down the cylinder and its Value or Lightness, how bright it is. If you go into a drawing program, even the basic MS Paint, you can pick colors in either RGB or HSL, the latter is useful because you can say, keep brightness constant and change the hue (at least if your monitor was calibrated to do so), while with RGB it isn’t intuitive how to do this.
Just checking: are you referring to this: CIE 1931 color space - Wikipedia ?
These don’t directly model your retina’s response, but represent an experiment from the 1930s where observers had to mix three lights to match the range of wavelengths. So they had to add a little red to the blue in order to reproduce violet. Thecones’ responses are rather Gaussian, with a little heaviness and asymmetry in the tails, but there isn’t an extra hump. But beyond the very first few layers of the retina, color is encoded as Francis Vaughn describes.
Oh, and red, green, blue are shorthand, but really the peak response of the cones is respectively yellowish, yellow-green, and violet, so in color science they’re normally called the L, M, and S cones (long, medium, short wavelengths). And the opponent color system doesn’t respond to red vs. green and blue vs. yellow as we normally understand them. I characterize the B vs. Y as purple vs. baby puke green (left figure).
Sharp was claiming for awhile that they introduced a fourth phosphor. We had a thread about it and didn’t really solve it, but I understand it was mostly a marketing gimmick. If you were to add a “yellow” as they claim, it wouldn’t add much to the gamut if you look at where yellow is in the CIE charts vs. the sRGB triangle.
It would only work if they were able to move the green phosphor well up the into the more pure shorter wavelength greens. But doing that would deliver most of the benefits anyway. The real problem is that it is hard to create really efficient narrow band phosphors. If you could create three monochromatic emitters that sat at the top edge and two bottom corners of the CIE chart you would get 90+% of colours. This seems to be what the laser based projection systems are getting towards. But then you need source material, and that all gets hard - at least the is a wide gamut standard.
The sRGB triangle and most of the similar gamuts are all based upon the emission spectra of the generally available phosphors. It is, in some ways, a tail wagging the dog problem.
I have an old CRT based home theatre projector (Barco Graphics 808s). It has three tubes, one of each colour, and three projection lenses. One of the really common upgrades to this projector is to install lenses that have colour filters. Filters are placed in the red and green, and although they reduce the light output, they shift their effective colour closer to the CIE chart corners. It makes a really noticeable difference. Blood red looks like blood and green grass looks properly green. Sadly you need to recalibrate the projector once you do this. The extension of the gamut isn’t huge, but it is enough to make it worthwhile hunting down the lenses.
You don’t even need a camera to get this effect. It’s not too hard to engineer two pigments and two light sources, such that the light sources look the same when viewed directly, and the two pigments look the same when viewed under one of the light sources, but different when viewed under the other.
For an extreme example of this, there are some (extremely expensive) projection systems that use an (apparently) black screen. The screen reflects only in one very narrow band each in the red, green, and blue range, and since those bands are so narrow, they represent only a negligible fraction of a true white spectrum, so under most light sources, it appears black. But those three bands where the screen reflects are exactly the three narrow bands produced by the projector. The advantage is that, since you almost completely eliminate the glare from normal light sources, you can use the system in a well-lit room, or even outside, and still clearly see the screen.
Indeed, if you could somehow rig a display so that the channels could produce both positive and negative values, you could reproduce the entire gamut with three channels, or even only with two. Maybe a white backlight, with colored filters that you could somehow make more or less transparent?
I had speculated that instead of adding a yellow they renamed the old green to yellow and added a new green further out. But apparently it’s even less useful and a gimmick.