I’ve been around SDMB long enough to know that blue, red, and yellow, which I was taught were the primary colors of pigment are just one possible set, and not even a particularly good one. I know that the preferred set is cyan, magenta, and yellow, and I even know those are based on the complements to the primary colors of light, red, green, and blue, which in turn are based on the cone cells specific to human eyes.
So, what I don’t get is why yellow, which presumably is absorbing blue light and letting red and green be reflected, and blue, which is only reflecting blue I guess, when mixed together appear green. I can’t even come up with a reasonable theory.
Your eye has three types of color sensors. We can call them R, G, and B. They don’t sense just one wavelength of light. They sense a range of different wavelengths, but have their strongest response to the wavelengths we call red, green, and blue, respectively.
Note that you don’t have a sensor dedicated to yellow. When yellow light hits your eye, it makes the R and the G sensors both fire. That’s what yellow means to your brain – a particular ratio of signals coming from the R and the G sensors.
If you see a mixture of yellow and blue light, then the all the sensors fire. The R and the G sensor are mostly stimulated by the yellow. And the B sensor is mostly stimulated by the blue. However, because the G sensor responds to a range of wavelengths, it picks up some of the blue light as well. Your brain says “Hmmm … I’ve got a little R and a little B, but a lot of G … I must be seeing green!” A mixture of yellow and blue light produces the same pattern of sensor response as pure green light, so we see it as green.
I have, of course, read all of Cecil’s columns at one time or another, though I have been know to forget a few of the details.
However, that particular column only answers my question if you assume that what my elementary school art teacher and I called blue is actually cyan. (Cecil describes cyan as both a light blue and a mixture of blue and green. That lacks a certain scientific rigor, but I blame space limitations.) If I found a true blue paint, defined by Cecil to be one that only reflects blue light, and mixed it with yellow, would I then get black, or even brown?
Well, OK, but all the sensors respond to a range of wavelengths, right? Not just the G ones.
.. and you lost me. Why is there a lot of G, but only a little R and B? (cue joke answers) It seems like there would be a lot of B, and less R and G, but I guess that would depend on the exact shades we were using.
All you eye sees are different inputs from the three types of color sensor. Very different combinations of wavelengths can produce the same pattern of stimulus. We see yellow + blue as green because yellow + blue produces the same pattern of stimulus as green by itself.
The response curves of the different cones overlap. However, G is in the middle. Red light doesn’t have much effect on B. And blue light doesn’t have much effect on R. But red and blue light both affect G.
Say you have 10 units of yellow light. Because this wavelength falls in between the peaks of R and G, you might get a response like:
R = 6
G = 6
B = 0
Now add in 6 units of blue light. B responds the strongest, but G responds some too:
R = 6 + 0 = 6
G = 6 + 4 = 10
B = 0 + 6 = 6
Now you have a situation where the G sensor is responding much more strongly than R and B. But that’s similar to what happens when you see pure green light. So you see green.
That’s true. Green is in the middle, so your theory is plausible. Is it backed up by empirical observations though? Someone must have broken out the frequencies coming from green paint (mixed from blue and yellow) and shown something the 10 units of yellow and 6 units of blue that you describe.
Also, cyan is a mixture of blue and green light, so by your theory, mixing red and cyan paint should also produce green. That is counter-intuitive to me, but I don’t have any paints to experiment with at the moment.
You seem to be talking about additive colour while the op is talking about subtractive. If you mix pure blue and pure yellow paint the result would be a very dark grey. The blue paint absorbs all but blue and the yellow paint absorbs all but yellow, mix it all together and there is very little left to be reflected. The blue and yellow paints that mix to make green are not pure colours.
The OP isn’t quite sure what he’s talking about. There’s clearly two ways for colors to mix, additive and subtractive. If you took a giant blue and yellow chessboard with thousands of ranks and files, and viewed it from a long distance, that would be additive, even though we’re “mixing” pigment colors.
If you mix inks, like you do when printing, that’s subtractive. The inks are (mostly) transparent to the frequencies they reflect, and so what you see with mixed inks, is frequencies that pass through both.
What’s going on with the paint though? At first it seems like paint would be more like the ink, but then paint isn’t really transparent like ink. Also, adding white paint makes colors lighter, which supports the chessboard idea.
So does our blue and yellow chessboard look green from a distance?
Not exactly. Both R and G cones have their strongest response to a color that people would describe as yellow and greenish-yellow, respectively. It’s probably best to talk about the three types of cone cells as responsive to short, medium and long wavelengths to avoid confusion. S would correspond to wavelengths closer to blue, L for closer to red, and M is in-between, but actually has a peak pretty close to the L cone.
Yes to the ratio, no to the comment about the lack of a yellow receptor.
You are confusing additive and subtractive color theory. Yellow plus blue equals green in subtractive color theory. Yellow subtracts away short-wavelength responses and blue subtracts away long-wavelength responses, leaving medium-wavelength responses, i.e., green.
In additive color theory, when all three cone cells are reacting to the visual stimuli, people perceive white, not green.
It depends. Each paint will have a spectral response curve. Certain wavelengths will be absorbed by each paint in different amounts. If every wavelength is absorbed by one paint or the other, then you’ll perceive their mixture to be black.
The OP is on about subtractive color. Pigments not lights. You can think about that as the opposite - it filters out colors rather than adds them together. So if you apply all 3 filters 100%, no light shines through (black). The filters themselves: blue gets hit by white light, absorbs R+Y, bounces blue back out into your eyes, etc. Mixing blue and red means that all hues in the greenish range may be absorbed, but there will be a portion of light that can pass through, and that light is seen as purple.
CMY(K) is really just a printsetting standard. The colors our eyes respond to are “RGB” but the real peak wavelengths will not be the best representation of what you might consider green to be. Also it gets a bit more complicated in areas beyond the retina, even before the cortex. For opponent colors that our brain considers ideal, e.g., the “blue/yellow” pathway is described by me as “purple/baby puke green.”
Cool website. Move a slider to stop the randomization.
Also, look up on metamers, two lights/pigments/etc. which look the exact same to you, but have completely different spectra. Some people complain about fluorescent lights, but many don’t see them as a big change from incandescent. If you measure the spectral output though, they’re very different, and will continue to be so even if they invent “perfect” fluorescents.
For your blue paint, the “yellow” would make black if it absorbs everything that the blue doesn’t, or at least does good enough by our eyes. For additive mixtures, the yellow would have to be a complement of the blue to make what seems pure white.
Yes. Yellow + Blue = White in additive color theory. You shine a yellow light and a blue light over each other, you get white, which should make sense, as yellow is red + green, and blue is blue, so you get R+G+B = white.
Pure deep yellow and pure deep blue should make black in the subtractive process for the reasons the OP stated. Usually, though, you’re not dealing with a pure yellow and pure blue, so you get a muddy green of some sort when mixing pigments.
Now that the additive vs subtractive color confusion has been addressed, it may be interesting to mention that, by the time color information leaves retina, it has been transformed into a coordinate system similar to the YC[sub]r[/sub]C[sub]r[/sub] popular in image processing computer applications. Specifically,
[ul][li] Output of Red and Green cones are added to produce a (Daytime) Luminance signal (supplemented with Rod outputs for Nighttime). Note that Blue cone outputs do not participate here, a reason “pure” blue appears dark.[/li][li] The difference of Red and Green cone outputs produces a (Red vs Green) signal.[/li][li] The difference of Luminance (R+G) and Blue cone output produces a (Blue vs Yellow) signal.[/li][/ul]
Thus the idea of four psychological primary colors (red, yellow, green, blue) does derive from retinal circuitry. In a thread 3 years ago I linked to a paper which depicts part of that retinal circuitry in some detail.
The metamer pigments may look the same under one source of illumination, but different under another. This is one reason people are often disappointed when they match colors in a store but find they mismatch in daylight.
