Minute Physics: little bits of science
January 10, 2012 10:44 AM   Subscribe

 
My favorite is Schrodinger's Cat.

Totally went to college with this guy! Didn't know him personally, though.
posted by dismas at 10:51 AM on January 10, 2012


There are currently 29 videos, and one new video is posted every Sunday night/Monday morning.
posted by filthy light thief at 11:10 AM on January 10, 2012


Pink != fuchsia.
posted by cmoj at 11:22 AM on January 10, 2012


And I'm apparently at a point in my life where I'm internet-pedantic about colors and I'm okay with it.
posted by cmoj at 11:23 AM on January 10, 2012


My favorite is Schrodinger's Cat.

Totally went to college with this guy! Didn't know him personally, though.


Purrrrdue?
posted by Mach3avelli at 11:25 AM on January 10, 2012 [1 favorite]


Learn why there are tides

They go in and they go out. You can't explain that.
posted by Dodecadermaldenticles at 11:54 AM on January 10, 2012 [2 favorites]


Feynman is rolling in his grave. The first video I happened to click on, "why there is no pink light," is an abomination. The statement "...we replace all that hidden grandeur with pink" is not correct in any sense whatsoever. And the color wheel illustration is deeply misleading. The videos are not made by someone with a real grasp of even basic concepts.
posted by radikai at 11:55 AM on January 10, 2012 [2 favorites]


The statement "...we replace all that hidden grandeur with pink" is not correct in any sense whatsoever.

Yeah, that was weird. I assumed it was a joke.

And the color wheel illustration is deeply misleading.

Eh, it gets the general gist. What do you want in a minute?
posted by Tell Me No Lies at 12:03 PM on January 10, 2012 [1 favorite]


RICK

RICK

RICK

PURDUE TAUGHT ME THIS RICK!
posted by wheelieman at 12:08 PM on January 10, 2012 [1 favorite]


I've watched a few of these, and the explanations are almost always... wrong. (*)

Apart from that - great stuff!

(* Example - Neutrinos can't be seen because they don't interact with light. Er, the only thing you _can_ see is light. Things other than light, you see because they interact with light or they generate light. Like neutrinos do, albeit through electron-excited Cerenkov radiation... ah, as it says. So you can see them. Or do they mean you can't see them _directly_... like everything apart from photons.)

(** Example 2 - the laser. Coins can be heads or tails, so there are four combinations of coins (HH HT TH TT) - and there's a picture. But photons can't be told apart (er, like two pennies can?) so here's a picture of up-up and down-down, but only one down-up, so there's one way to be different and two to be the same. Uuuuuhhhh... so where's the up-down? The what? But photons want to be like each other, so that's ok... Yeah, this is starting to hurt.)

And I don't _think_ this is one of theirs - the Twin Paradox - but it has the same delights. It's wrong on so many levels, without exploring what time dilation is about, that it's almost criminal. The way the clocks do or don't move is arbitrary, and the final statement that GPS needs relativity to work (when it merely compensates for it, because it has to) is really annoying.

Make it as simple as necessary, but no simpler, said Al.
posted by Devonian at 12:20 PM on January 10, 2012 [1 favorite]


The pink light question is one in which I have a special interest.

A misnomer in Signal Processing and related fields is "pink noise". It is very much like "white noise" except it has a different spectral shape. On a linear scale it looks like 1/x and on a decibel scale it is a downward sloping line from low frequency to high frequency. That is not pink, although I see that term daily. The spectrum of pink is white with a notch in the middle of the green part. So at the end when he says pink = white - green, that part is correct.

But:

magenta is not pink. Not even really close

And:

you can see pink light any night in downtown New York or Tokyo in a neon light where white light is emitted and there is a gas (I don't know offhand what gas--it is probably not pure neon) that absorbs green.

so to say there is no pink light is flatly incorrect.

I don't think I will be watching too many more of this guy's videos.
posted by bukvich at 12:27 PM on January 10, 2012


I don't think I will be watching too many more of this guy's videos.

I'm pretty sure you're not the target audience.

Since we have all of our physics experts present though: Explain relativity and why it is important in a way that your mother will understand. You have 75 words. Go.
posted by Tell Me No Lies at 12:30 PM on January 10, 2012


The speed of light is constant in any moving frame of reference = the special theory of relativity.

That one sentence revolutionized the field of physics. It's kind of hard to explain why. See Feynman on "why" questions.
posted by bukvich at 12:33 PM on January 10, 2012


Ok, now that pink is settled, why is there no brown stripe in the rainbow?
posted by ikalliom at 12:48 PM on January 10, 2012


I think these videos take a number of short-cuts in describing larger concepts, and in doing so, seem to confuse the topic for people who know more than the intended audience.

bukvich: you can see pink light any night in downtown New York or Tokyo in a neon light where white light is emitted and there is a gas that absorbs green.

The video says there is nothing that emits pink light, which your statement would agree with, as pink is made by absorbing green.

cmoj: Pink != fuchsia.

I think the point is that there is a gap between Red and Purple, between which lie pink, fuchia, magenta, and other similar colors. In the interest of time, Henry lumped all those "between" colors into the same discussion, or focused specifically on Pink.

Fun tangent: the archaic color Pinke is actually a shade of yellow that would blend with blues to make green.
posted by filthy light thief at 1:11 PM on January 10, 2012


If pink (light) is the absence of green, and green is made up of yellow and blue, does that mean pink is red?
posted by kilo hertz at 1:15 PM on January 10, 2012


Oh wait, this probably has something to do with difference between subtractive and additive methods of producing colour.
posted by kilo hertz at 1:18 PM on January 10, 2012


filthy light thief I will take your word for it but I did not hear any emit. I heard no pink. I swear I was paying attention.
posted by bukvich at 1:27 PM on January 10, 2012


bukvich, I added some words for him.

Transcript summary: there is no pink light, but it is made up from a mix of red and blue light, and our eyes make up pink. And pink should be called minus green, because pink is what is left when you have white light and take out the green.

Shorter summery: while you see colors of the rainbow, your brain makes up pink.
posted by filthy light thief at 1:36 PM on January 10, 2012


@Tell Me No Lies - your name forbids me to.
posted by Devonian at 1:42 PM on January 10, 2012


The eye/brain makes up other colors, too. Yellow, as the eye/brain sees it, can be either yellow light, with a wavelength of about 575-600 nanometers, or a mixture of green (500-575) nanometers and red light (600-750 nanometers). And the green and red making yellow doesn't have to have any yellow-wavelength light. And cyan (turquoise, blue-green) and be its own wavelength (around 480 nm) or a mixture of blue and green. But yeah, there is no magenta/pink wavelength. Gotta be a mixture. And colors like brown or olive or steel blue are complex mixtures of wavelengths.
posted by tommyD at 1:43 PM on January 10, 2012 [3 favorites]


The eye/brain makes up _all_ the colours. All there is 'really' is electromagnetic energy of various intensities and frequencies; the rest is a factor of our retinal chemistry and some rather neat processing, neither of which is by any means the same between individuals.

There is no monochromatic light that is commonly called pink, is about as much as you can say without getting into some seriously deep waters about perception. You can then (I'm sure) go through a bit of a canter through the shallows of some of those waters usefully, interestingly and reasonably accurately, even in 60 seconds. I'm not sure that this particular video does that.
posted by Devonian at 1:50 PM on January 10, 2012 [2 favorites]


The problem with the pink light thing is that the person who made the video doesn't understand that color is entirely synthetic (not just "the stuff in the gap") and so his explanation is a really good example of knowing just enough to get things wrong in a very bad way. Most importantly, this is not a physics question. It's a biology question.

The reason why all the different representations of color space (the wheel, the cube, whatever) are unsatisfactory is because how our brain decides what "color" something is arises from a complicated combination of signals (or their lack) from the three kinds of pigmented color receptors in the human eye. Each is tuned to a roughly bell-curve shaped region of the visible spectrum, and so for just that one type of receptor ("cone") there's no way to tell the difference between a low level of a wavelength on-center for that receptor and a high level of a wavelength off-center for that receptor. But since we have more than one type, then our brains can make a decent guess based upon how two adjacent (in sensitivity) cones signal. Or all three. Or the two that aren't adjacent...which couldn't correspond to a single narrow (relative to what the cones are sensitive) wavelength, but could happen with two and does happen in nature (as in Eath's flora and fauna) and it can be helpful for us to be sensitive to it.

So in this sense you could say that a particular color we see isn't on the spectrum because the spectrum, a rainbow, is always made up of all wavelengths varying at different angles relative to our eyes (which is why we see it as separate colors). However, to say that it's "not a real color" assumes that there are "real colors", which there aren't. No color is a real color. Color only exists in our brains. It's a total hack—not just to roughly locate where some reflected light is on the solar spectrum (as seen through the Earth's atmosphere), but to "tag" certain combinations so that we experience them qualitatively differently from each other when they're not actually qualitatively different. That's very helpful because it makes it faster and easier to distinguish between these combinations—it's extremely useful to immediately distinguish red from green (the former is blood or plants advertising themselves, the latter is random plant greenery).

Because our perception of color is wholly synthetic and exists exclusively in our brains as the result of an idiosyncratic interpretation of signals from variously pigmented receptors in our eyes, describing it exactly in a model—like a color wheel or cube or whatever—is very difficult or essentially impossible. It's not something physicists can do, it's something that only neurologists could eventually do. As it is, there's a lot of data that's been gathered about how people see color and, for example, I've read a paper where the author did a lot of difficult mathematics and found that the best dimensionality to describe human color space was something like 17 (I'm guessing; I don't recall exactly what it was...it was bigger than ten dimensions, I'm certain).
posted by Ivan Fyodorovich at 1:50 PM on January 10, 2012 [4 favorites]


sorry - by any means _always_ the same between individuals...
posted by Devonian at 1:52 PM on January 10, 2012


He's not saying that pink is not a real color. I think that particular video was more about wavelengths, less about perception and biology.
posted by filthy light thief at 2:45 PM on January 10, 2012


"He's not saying that pink is not a real color. I think that particular video was more about wavelengths, less about perception and biology."

Color is only biological. You can't talk about color and only talk about wavelengths because wavelengths don't have color. Light doesn't have color. He's saying that "pink" isn't in the spectrum, which is true, but he's saying it as if that makes it qualitatively different in reality from the other colors. But it's no more invented by our brains than the other colors are.

And he's completely, entirely, horribly wrong when he says that everything outside the visible spectrum (the "gap" on his color wheel) corresponds to that "pink". That's not true in any sense. That "pink" doesn't exist anywhere on the spectrum, visible or the entire EM spectrum. That's because it's the result of the two extremes of our visible spectrum being stimulated while the middle isn't. There's no single range of wavelengths on the spectrum, no matter how big or small, that corresponds to that.

Insofar as we perceive light that's in that "gap", then by definition it's not in that gap. What he's saying is nonsensical. And when a range of wavelengths peaks outside the visible range but its edge weakly stimulates one of our kinds of receptors, then it's either a very deep and dim violet or a very deep and dim red. (This is complicated, however, by our night vision receptors, which are sensitive to a different range of wavelengths and they do contribute in a very small way to our color vision, noticeable under certain extreme conditions.)

What is going on with this guy shows why it's totally wrong to talk about color as wavelengths, as you say he is doing, rather than biology. Because where I think he went wrong is that he has some knowledge of color theory from an artistic perspective, not biology, and so he has the color wheel in mind, which he then attempts to connect with a little bit of physics. That's why he wants to somehow place what he's calling "pink" in that "gap" and why he then wants to place everything outside the human visible spectrum in that "gap". It's just wrongheaded.

Anyway, he's in good company—that's what Newton did. But they're both wrong. And Newton didn't have the advantage of, you know, actually having all this stuff scientifically known and well-understood.
posted by Ivan Fyodorovich at 3:21 PM on January 10, 2012 [2 favorites]


This is the trouble with that approach to the spectrum - there's no fuchsia in it.
posted by Devonian at 3:30 PM on January 10, 2012 [2 favorites]


Ahoy metafilter!

Henry Reich here of MinutePhysics, glad you've had such a fun discussion and I'd like to respond to some of your comments.

First: neutrinos can't be seen directly because they don't scatter light. Other things (like most matter) does scatter/emit light, which I would argue is the definition of being able to 'see' it. With a neutrino all you can see is the result of it crashing into stuff, but not the neutrino itself. The difference between seeing the ripples in a lake from a rock and seeing the rock itself.

Second: two coins are distinguishable at a microscopic level (so HT & TH are different). Two photons are not (so HT is the same as TH). Hence, bose statistics. This is why lasers and superconductors work, so if you don't believe it, just play a CD.

Third: The color pink!! Everyone really likes to talk about this one. Sometimes they start by saying I really meant hot pink, magenta or fuschia and not pink... but I believe that was addressed in the video, and anyway, do a google image search for "pink" and tell me you get anything but magenta.

As for the color wheel issue, I realize I have made some error in communication because so many people think that I mean we "see" all the other light (radio, IR, UV, etc) as pink. This is horrendously wrong. What I meant was that in order to compactify the electromagnetic spectrum into a wheel, there's inherently a gap, and our brains fill that gap with a made-up color.

And I'm fully aware that there are many nuances in the biological&neurological perception of color, but it's interesting to think that, even though we have only three different types of cones and so in some sense we just see "RGB", most colors that we see CAN be generated by a corresponding "color" of monochrome EM light (and by color I mean frequency), while pink/magenta/fuschia is not associated to a monochromatic color of light.


AND in response to your challenge about Special Relativity, I think that saying SR is just "speed of light is constant in every inertial frame" is a horrible definition. It may be the thought process that Einstein used to discover SR, but here's my <75-word take:

Special Relativity is the idea that space and time are just different directions in the same flat "space-time". In the same way that turning 90° to the right can "rotate" x into y (with large distances in x corresponding to large distances in y), you can "rotate" x into t… except that large spatial distances correspond to small temporal ones. The speed of light is just a units conversion factor between x and t.
posted by minutephysics at 1:37 PM on January 11, 2012 [8 favorites]


**by google image search for pink I meant "color pink". "pink" gives a lot of images of the artist
posted by minutephysics at 2:04 PM on January 11, 2012


Hiya, Henry! Welcome aboard.
posted by cortex at 3:41 PM on January 11, 2012


The difference between seeing the ripples in a lake from a rock and seeing the rock itself.

But all we see of anything is the ripples.
posted by empath at 3:50 PM on January 11, 2012


Yeah, there's a lot more to go into than the purely biological stuff on Pink, (which I'll go ahead and refer to as Magenta now because that's more my background with it.) But Ivan, sure, light has no intrisic color, we get it through our sensory perception. Fine. What you perceive in a pure visual spectrum of colors does not include magenta.

As was mentioned above, yeah, it's a question of additive verse subtractive color. Add Red, Green and Blue together and and you get White light. Subtract Red from White and you get Cyan, subtract Blue from White and you get Yellow. Subtract Green from White and you get Magenta, for the reasons explained in his video. Green is in the center of the visual color spectrum, and subtracting us gives us (roughly) a mix of red and blue. without the colors between them in the spectrum.

Magenta isn't particularly noteworthy in this regard, really. Most colors aren't on the pure spectrum (brown, as someone said, is a good example) but are rather mixtures of other "pure" colors. What makes Magenta interesting is the color wheel aspect of it. That our brains replace the gap in the non-visual part of the spectrum with this non-spectrum mix of red and blue.

Also, as a fun-fact. Fluorescent lights give off a nasty green spike when caught on film or video, so gaffers and designers will often cover them with a light pink gel before shooting to cancel this out. Most gaffers I've worked with have been pretty insistent that the gel color not be called "pink" or "magenta" but, in fact, "minus green."
posted by Navelgazer at 4:11 PM on January 11, 2012 [2 favorites]


Navelgazer: Most gaffers I've worked with have been pretty insistent that the gel color not be called "pink" or "magenta" but, in fact, "minus green."

Interesting, and an example of a "Minus Green" filter gel sheet.
posted by filthy light thief at 4:57 PM on January 11, 2012


What I meant was that in order to compactify the electromagnetic spectrum into a wheel, there's inherently a gap, and our brains fill that gap with a made-up color.

Isn't the color wheel an artificial model in color theory, based not on the electromagnetic spectrum, but on how cone cells in the retina are distributed in a majority of people — in other words, the color theory that's taught in art class is based on how red, blue, green and additive mixtures of them are perceived by most people's eyes?
posted by Blazecock Pileon at 6:19 PM on January 11, 2012


I had really hard time understanding the explaining about tides. I still don't get why there's a second tide. It didn't make sense to me watching the video.

Still love watching them though, even if they all go over my head a little bit.

And welcome to metafilter minutephysics!
posted by royalsong at 6:32 PM on January 11, 2012


I'd often read that renormalization was a kind of mathematical "cheating", but until I saw the "Adding beyond infinity" video I'd never really understood what kind. So thanks for that.

I am a little worried, though. It seems to me that if that method is sound, then I should be able to use it this way:

1 + 2 + 4 + 8 + ... = 1 (1 + 2 + 4 + 8 + ...)
= (4 - 2 - 1) (1 + 2 + 4 + 8 + ...)

= 4 + 8 + 16 + 32 + ...
- 2 - 4 - 8 - 16 - ...
- 1 - 2 - 4 - 8 - ...

Most of the second line cancels out the first line just like in the video, and we can group the third, leaving

1 + 2 + 4 + 8 + ... = -2 - (1 + 2 + 4 + 8 + ...)

Cancel the infinite sum on both sides and we're left with

0 = -2

Maybe this kind of thing is doable in FORTRAN but I'm sure it would cause trouble in the real world.
posted by flabdablet at 7:02 PM on January 11, 2012


I still don't get why there's a second tide.

That's probably because you think of the Moon going around the Earth, rather than thinking of the two of them as dancing around each other.

As well as the Moon lifting the water on the near side of the Earth away from the centre of the Earth, it's lifting the centre of the Earth away from the water on the far side.
posted by flabdablet at 7:07 PM on January 11, 2012 [1 favorite]


"What I meant was that in order to compactify the electromagnetic spectrum into a wheel, there's inherently a gap, and our brains fill that gap with a made-up color."

But this just doesn't make sense. In the color wheel model, it's not the EM spectrum which is compactified into a wheel, it's the human brain's response to the visible spectrum that's compactified into a wheel. By definition, there is no gap. The only way that the assertion that there's a gap makes sense is if you assume that the color wheel model is actually representative of the EM spectrum itself. But it's not, so it doesn't. The color wheel and the EM spectrum are not isomorphic. The whole notion of the "gap" reflects a deeper misunderstanding that's misguiding how you think of color.

"And I'm fully aware that there are many nuances in the biological&neurological perception of color, but it's interesting to think that, even though we have only three different types of cones and so in some sense we just see "RGB", most colors that we see CAN be generated by a corresponding "color" of monochrome EM light (and by color I mean frequency), while pink/magenta/fuschia is not associated to a monochromatic color of light."

As others have mentioned, there's many more colors than just fuschia that are not in a rainbow; that is to say, cannot be be produced in our perception via a single frequency of light. Brown is a frequently mentioned example. Metallic colors are others. I anticipate someone claiming that these are not "pure" colors; but that's begging the question. It assumes that "pure colors" exist in reality and, furthermore, it bears the hidden assumption that the "pure colors" are the colors in the rainbow.

This is all caused by the inevitable incorrect assumption made when it was learned that white light could be decomposed into a spectrum with a prism. The assumption is that the colors of the spectrum are "pure colors" and that those colors exist in nature, not just in our brains, and that our perception of them is a one-to-one correspondence. None of those things are actually true. Prior to this, there were lots of experiments with color, both additive and subtractive (though the distinction wasn't always fully understood) by artists and philosophers and such and various models to account for all this were proposed and used. But it was only when light itself was scientifically examined, primarily by Newton, and then later with the idea of EM, that the two things were connected in the way that has led to this misconception.

When you talk about how a single frequency can cause the perception of color as if going the other direction—talking about how a color can or cannot be stimulated by a single frequency—is meaningful, you're making a fundamental error.

Here's a very similar example: the human nervous system has separate nerves for sensing heat and cold. Now, does that imply that heat and cold are two different things?

Let's say, for the sake of argument, that in addition to the perception of "hot" and "cold", we also had a perception of a third qualitatively distinct feeling...we'll call this "frot". And let's say that we find, through experimentation, that a single probe of a certain temperature never elicits the "frot" perception. It can be elicited, however, by two probes, one that feels "hot" and one that feels "cold". And then, later, the ideas of heat and temperature are developed and we learn to think about temperature as something along a single spectrum. Now, someone comes along to tell us something interesting. He says, isn't it interesting that "frot" isn't on the temperature spectrum the way that "hot" and "cold" are? "Hot" and "cold" are on the spectrum, but "frot" isn't.

Well, it is interesting in what it tells us about the human nervous system. It isn't interesting in what it tells us about the temperature spectrum because it doesn't tell us anything about the temperature spectrum. In fact, "hot" and "cold" aren't distinct things on the spectrum, either. Those qualitatively different experiences we have are all in our heads, not in the real world. In this sense, "frot" is no different from "hot" and "cold". None of those things can correctly be applied to something on the temperature spectrum.

Just as "color" doesn't apply to the EM spectrum. Color exists in any meaningful sense only in our heads. To attempt to talk about it outside the context of human biology is meaningless.
posted by Ivan Fyodorovich at 8:11 PM on January 11, 2012 [2 favorites]


flabdablet: Cancel the infinite sum on both sides

You lost the sign there. The infinite sums don't cancel but reinforce, yielding the conclusion
1 + 2 + 4 + 8 + ... = −1
just like in the video. This is actually a correct result, in suitable contexts.

(In brief: When you define what it means to add up infinitely many numbers, you need to adopt some notion of limits and convergence and all that jazz which they do in first-year calculus. The usual notions yield 1+2+4+8+...=∞, like you'd expect, but there are other useful ways to set things up, and there's a particularly nice one that gives the result above (and has applications in number theory). The Wikipedia article on p-adic numbers has a decent introduction.)

(In briefer, if you have a comp sci background, you might be satisfied with this: the sum 1+2+4+... is the 2's complement representation of −1 on a computer with infinite word size.)
posted by stebulus at 8:52 PM on January 11, 2012


You lost the sign there.

D'OH!

Thanks.
posted by flabdablet at 9:39 PM on January 11, 2012


the sum 1+2+4+... is the 2's complement representation of −1 on a computer with infinite word size

And it works just as well on a computer with any finite word size, as long as there's no overflow detection.
posted by flabdablet at 9:41 PM on January 11, 2012


Just as "color" doesn't apply to the EM spectrum.

It feels like you are overreaching here. Based on the perceived colors of two different laser lights I can say quite accurately which regions of the electromagnetic spectrum they are operating in. It's not as if color gives no information about wavelength.
posted by grog at 11:05 PM on January 11, 2012 [2 favorites]


"It's not as if color gives no information about wavelength."

Well, no, because that's the point, right? Just as our perception of "hot" and "cold" gives some information about temperature. But "hot" and "cold" don't directly have anything to do with temperature, those qualia exist in our brains and it doesn't make sense to say that one region of temperature is "hot" and another is "cold" in an absolute sense—that's only meaningful for what hot and cold mean to us.

More to the point, you missed my point about isomorphism. The relationship between the visible EM spectrum and the way our eyes detect light and our brains interpret that data is not bi-directional. That is to say, the transformation from EM to our experience isn't reversible. All sorts of different combinations of wavelengths will produce the same experience of a color, so our experience of color is not a reliable means of measuring the wavelength of EM. In can be in some cases. In others, it's not at all. Trying to map out our experience of color directly to the EM spectrum can't work because the EM spectrum is a one-dimensional continuum and our experience of color is not. This is what is fundamentally wrong with minutephysics's treatment of it.
posted by Ivan Fyodorovich at 11:32 PM on January 11, 2012 [1 favorite]


Ivan Fyodorovich, how much trouble do I get myself in by asking where you fall on the science-religion spectrum?

For if the smallest most fundamental building blocks of this universe can be described in a mathematical & scientific method, I don't believe that there are ANY features of the universe that cannot ultimately be traced back to that form. That is to say, I believe it to be a scientific fact that all of the greater complexities of the world, life, human perception, consciousness, the weather, love, etc.etc.etc. arise from the simple and fundamental laws of the universe. Never mind that we do not know the exact details yet of how to explain these phenomena in fundamental terms - the chief aim of science is to uncover those relationships bit by bit.

As you have mentioned, temperature is a wonderful example – at first we had things called thermometers, and when it was hot the mercury rose, and when it was cold the mercury fell. And then we discovered statistical mechanics and were able to "explain" how the motion of atoms could cause the mercury in our thermometers to rise and fall. Did this merely tell us about how thermometers worked? No - it told us what "temperature" really was, and how the fundamental laws of physics could lead to perceived feelings of "hot" and "cold".

So when you say "color is only human perception", I say "of course, but first the photons of light must enter the eye, and there excite molecular transitions in the rods and cones, some of which are more sensitive to photons of a certain frequency than others, and then of course a message is sent to the brain. And for every monochromatic frequency between 400 and 790 THz, the brain is able to assign a color, which we may draw along a line. But there are combined frequencies of light to which the brain assigns other labels, and yet is able to perceive to some degree how "close" these frequencies are to the monochromatic ones. And one of these combined frequencies we happen to call "hot pink" and it is perceived to be "close" to both the ends of the spectrum at 400 and 790 THz, but not at all close to the middle (which isn't surprising since it is composed of light from both ends but not the middle), so from a topological perspective this color may be thought of as the one point which draws the ends of the line together into a circle."
posted by minutephysics at 11:59 PM on January 11, 2012


I believe it to be a scientific fact that all of the greater complexities of the world, life, human perception, consciousness, the weather, love, etc.etc.etc. arise from the simple and fundamental laws of the universe.

Why?
posted by flabdablet at 12:43 AM on January 12, 2012


I believe it to be a scientific fact that all of the greater complexities of the world, life, human perception, consciousness, the weather, love, etc.etc.etc. arise from the simple and fundamental laws of the universe.

Why?

Because the whole point of physics and science in general is to try to describe the underlying laws of everything in the universe. Or, to put it another way, where else would that complexity come from?

A more concrete example: information theory has a lot to say about the way we store and represent data, but since information is ultimately a physical thing that must be recorded somewhere in the universe, then ultimately information is quantum mechanical, because the universe is quantum mechanical. This is the origin of the field of quantum information.
posted by minutephysics at 1:15 AM on January 12, 2012


Well, to answer your question with its hidden assumption, I'm an atheist materialist with academic training in science (initially physics), then later the philosophy and history of science. So I share your materialist epistemology.

Your discussion of temperature is provocative because it's not unlike this discussion. You might find it beneficial to look into the actual history of the investigation of heat and temperature. (What we directly sense is the relative rate of heat transference. Heat is fundamental, temperature is synthetic and expresses a relationship.)

I'm not saying that "color" is metaphysical. I am saying two, related things. First, that color distinctions are, in our experience, qualitative distinctions but those qualitative distinctions don't exist outside our brains. In exactly the same way that because we have separate hot and cold receptors, we experience them as qualitatively distinct but the physical process that excites those receptors is identical in both cases, except in quantity, not quality.

Second, and more to the point, I'm saying that our experience of color cannot be mapped onto the EM spectrum. This is essentially because the EM spectrum is a linear continuum and our color perception isn't. When you talk about a color wheel and topology—and it's good that you're talking about topology, because this will lead you in the right direction—your mistake is in thinking that the color wheel correctly approximates human color vision. It doesn't. You're attracted to it because although it's a loop, topologically it's one-dimensional like the EM spectrum, so you want to map the two to each other. Note that this is an isomorphic mapping: any given experience of color can be mapped to a region on the EM spectrum, including the one color you say cannot, with it being a special case that maps to the region in the "gap".

But the color wheel doesn't approximate human color vision. The color wheel is just one among many useful-for-some-purposes-but-not-accurate-for-all models of human color vision. The color cube is thought to be more accurate, but human color vision isn't three-dimensional, either.

You keep talking about your "pink", but you haven't addressed the issue, raised above by others, of brown and metallic colors and the numerous other colors that are not found in a rainbow. Color is something that our brains produce that has a very complicated relationship with the physical properties that give rise to it and thinking about color as if it can be meaningfully laid along the EM spectrum is a fundamental category error.

Frankly, I feel like you at some point learned that what you're calling "pink" is produced by the two cones which lie at the opposite ends of our sensitivity, you assumed all the other colors exist in-between, you thought of connecting those two ends into a loop, and then maybe you heard of the color wheel, and you thought you'd found an accurate mapping of human color vision with the visible range of EM. Maybe I'm being ungenerous.

There's numerous resources available where you can learn about human color perception. If you want to talk about how our color perception relates to the physical properties of light, then first you need to actually know quite a bit about human color perception. Knowing that there are three types of pigmented receptors just isn't enough. It's enough to lead one to making incorrect assumptions, however.
posted by Ivan Fyodorovich at 1:17 AM on January 12, 2012 [1 favorite]


I feel a little bad for hitting on you so hard on this because, otherwise, I'm all in favor of what you're doing. Indeed, I've thought in the past about doing pretty much the same thing—making videos to explain science and math concepts to people in an accessible manner.

It just happens that color vision is a longstanding pet interest of mine. I think it's really a quite remarkable evolutionary achievement—and by "it" I don't mean developing the ability to discriminate between relatively different ranges of frequencies of EM, I mean the cognitive adaptation of us experiencing these different ranges as qualitatively distinct.

A similar adaptation with hearing would be if different octaves seemed to us to be very distinct things. Where, say, 650Hz to us was like "red" and sounded completely different than 10.4kHz, which was like "blue"...even though it's 650 x 2^4. In fact, we naturally experience those two frequencies as being similar, because of the log relationship...which makes sense because of harmonics. Our perception of sound, unlike light, is isomorphic with the physical phenomena. Both are one-dimensional. (Not accounting for amplitude.) Our perception of sound is "monochromatic", so to speak.
posted by Ivan Fyodorovich at 1:46 AM on January 12, 2012 [1 favorite]


By the way, I really like your explanation of Heisenberg Uncertainty because a pet-peeve of mine is when it's characterized as a practical measurement problem, e.g.: "by measuring something you disturb it" and not as a something fundamental, which it is.
posted by Ivan Fyodorovich at 2:02 AM on January 12, 2012 [1 favorite]


No worries at all, I enjoy the discussion.

I still think you're misinterpreting my understanding of color perception - I certainly make no claim of an isomorphism between the EM spectrum & perceived color. Just that there's a map from the EM spectrum to perceived color (but not the other way round), and there's a perceived color which ISN'T in the image of the EM spectrum but which can be viewed as compactifying the EM spectrum.

With regards to colors such as brown and "metallics", well, as far as I'm aware browns are typically in the red-orange range of the spectrum and are merely perceived as "brown" by their darkness relative to the white in their surrounds. But please correct me if I'm wrong. Metallic colors, on the other hand, are actually much more strongly determined by their characteristic reflections of incident light - though I believe these reflections are typically a fairly constrained portion of the EM spectrum as well (whether red, orange, yellow, or 'white' i.e. everything equally, etc.), determined by transition levels of the metal.

I know color is super complicated, I just don't have a big problem with sharing the idea that there is no monochromatic light that maps to our perception of pink. I also agree that it's interesting to note that color perception is qualitatively different from sound perception (where there is an isomorphic response) - but at the same time we are able to distinguish high pitch from low pitch sounds and we just happen to "compactify" the "sound wheel" with a periodicity that is much smaller than the frequency response range of our ears. Would you find it inconceivable to imagine a world where our perception of "red to blue" went not just from 400 to 790THz but then began again with a new "octave" or red at 800THz?
posted by minutephysics at 2:44 AM on January 12, 2012


"...EM spectrum but which can be viewed as compactifying the EM spectrum.

That's where I part ways with you. But, anyway.

This was interesting:

"...but at the same time we are able to distinguish high pitch from low pitch sounds and we just happen to 'compactify' the 'sound wheel' with a periodicity that is much smaller than the frequency response range of our ears. Would you find it inconceivable to imagine a world where our perception of 'red to blue' went not just from 400 to 790THz but then began again with a new 'octave' or red at 800THz?"

I'm not entirely sure what you're asking. The periodicity we perceive in sound reflects the actual physical properties of the sound waves; i.e., the mathematical relationships between the frequencies. Each octave represents a doubling of frequency. I assume that we perceive the entire audible spectrum as qualitatively a continuum of the same "thing" (that is, it's qualitatively one thing) because there's been no real adaptative benefit in perceiving different ranges as qualitatively distinct things, as we do with light. I've thought about this a bit, and my highly speculative thoughts are that our aural environment is relatively much less information rich than our visual environment and therefore we don't need shortcuts for singling out bits of information that are especially important.

That's begging the question, of course: is the natural environment itself less saturated with sound information than it is light information? I suspect that it is. But even if it's not, I also suspect that for the purposes of primates, the visual environment is more important and therefore our emphasis on sensory apparatus for sensing it means that it's relatively information rich for us. Which then also leads to needing ways to quickly sift through all that information. I also speculate that our powerful color vision is related to us being omnivores.

In the case of the visible spectrum, there's no opportunity for the same kind of periodicity we perceive in sound because the visible spectrum doesn't span magnitudes of two. It ranges only from 380THz to 780THz. (Incidentally—although I'm sure you know this, I mention it for others—that's very close to where the Sun's spectrum peaks and is pretty much exactly where the solar spectrum that reaches the Earth's surface peaks, which is why that's our visible range. Diurnal animals may have somewhat smaller or larger ranges, but they're all pretty much centered here.)

That visible light is a relatively small range of very high frequencies and audible sound is a relatively large range of low frequencies has a lot to do, also, with the differences between how we sense light and sound.

In truth, we sense sound not unlike how we sense light—there are different kinds of sensors tuned to respond to a specific range of frequencies. These take the form of very small hairs connected to nerves. It's not that difficult to have many different kinds of these hairs, each tuned to a relatively very small range of aural frequencies. So, what we have is a whole bunch of these that each account for our ability to discriminate between two different frequencies. As far as I know, we don't interpolate a frequency from the combination of stimuli, although we might.

It's not biologically practical to discriminate frequencies of light in this way because the range is relatively small and the frequencies are relatively large (huge). So animals developed two or three or four different kinds of sensors sensitive to different ranges, and then also interpolated greater resolution of frequency via combined signals from them.
posted by Ivan Fyodorovich at 3:53 AM on January 12, 2012 [2 favorites]


No no no no no magenta is not even close to pink.

Pink is a green notch. All colors with one pure chroma removed.

Magenta is a red spike + a blue spike. The sum of two pure colors.

Are you color blind? They don't even look like one another. You might want to look at an artist's book about color sometime. This one ain't too bad.
posted by bukvich at 6:05 AM on January 12, 2012


Although in my prismacolor box they got pink at # 929 and magenta at # 930 so you are not the only person who makes this mistake.
posted by bukvich at 6:29 AM on January 12, 2012


human color vision isn't three-dimensional, either.

Sure it is; there are three color opponent processes. Red vs. Green, Blue vs. Yellow, and White vs. Black.

Pink is a green notch. All colors with one pure chroma removed.

Nope, pink has many metamers. The representation of pink on any CRT, for example, is mostly just three spikes (with the green and blue spikes shorter than the red spike).

Here's my breakdown of how physical spectra relate to human color vision.
posted by a snickering nuthatch at 8:06 AM on January 12, 2012 [1 favorite]


> pink has many metamers

Crikey I knew somebody would feel the need to go 3D on me. Yes it is true that on the cube or on the cylinder there are many pinks. I was talking about a simple linear representation which I have never seen fail unless you want to get into textures, less uncanny human models, and stuff you need a buttload of bits to render.

Since I don't have the energy to enter a debate which will probably not end in our lifetimes, I will just say peace (Peace!) and invite your comment on the illusion of the pink dots:

new scientist link.

Do you think that guy knows what he is talking about?
posted by bukvich at 9:11 AM on January 12, 2012


Because the whole point of physics and science in general is to try to describe the underlying laws of everything in the universe. Or, to put it another way, where else would that complexity come from?

See, I kind of come at this from the opposite direction. Seems to me that the whole point of physics and science in general is to try to make descriptions and predictions that (at least in principle) work equally well for everybody; that is, to produce a consensus representation of those parts of reality available to human senses as augmented by instruments, insight and reason.

And I'm fully convinced that no Theory of Everything from "fundamental" physics will ever be of much use, in practical ways, to help us understand systems as complex as human cognition. It's a forest-for-the-trees thing. Systems have levels of organization, and applying a theory applicable only to the finest-grained details of a system's operation is often not very profitable. When I'm debugging rogue code, there's very little point getting down to the nitty-gritty of what's tunneling through which transistor inside the CPU and why.

From where I sit, complexity is a given. It's the starting point. And it seems to me that scientific theories (including the more reliable and general ones we tend to label "underlying laws") are more akin to information compression techniques than to fundamental features of reality. And sometimes, lossy compression is more useful.

Nature is not and never has been constrained by the laws of physics and mathematics that we invent and/or discover. We just get better at finding workable metaphors for various aspects of its behavior as the years go by. Sometimes those metaphors are quite simple (e.g. point particles and all-pervasive fields) and yet can encode astoundingly complex behavior. Sometimes we find the complexity so overwhelming that we just choose to ignore it, and satisfy ourselves with statistical models instead (e.g. thermodynamics). And sometimes, as is the case with quantum mechanics, we do both things at once - and then convince ourselves that our inability to assign definite quantum states to systems between measurements says more about nature than it does about the process of recognizing quantifiable things.

We're so hung up on the idea that nature ought to be repeatable and predictable that even when our very best theory shows that it is neither, we start inventing idiocies like multiple-worlds to "explain" the "collapse" of wave functions in ways that let us cling to the illusion that our "laws" do constrain and "explain" behavior. Personally, I'm with Albert: the most incomprehensible thing about the world is that it is comprehensible.

I mean, I like that it's comprehensible. But I don't believe for a second that comprehensibility is anything more than evidence of our brains being well adapted to the complexity they're embedded in. The idea that the universe has "underlying" laws is one I find completely and utterly wrong-headed.

Who breathes fire into the equations? Wrong question. The equations are what's left when we've finished sucking the fire out of reality.
posted by flabdablet at 10:25 AM on January 12, 2012 [1 favorite]


OK two more comments on pink and then I swear I am going to stop.

1.) There may be a cultural or generational divide here. I am old enough that my elementary school was not filled floor to ceiling with a bunch of Mattel pink Barbie stuff and I never had any special attraction to that color so I may well see it differently from others in a way similar to the Himbas with five words for color see differently from the anthropologists who worked with them. When I see something like this, it is almost physically painful for me to look at it.

2.) I did an experiment where I took my prismacolor 929 pink and 930 magenta and swatched them on white paper under sunlight illumination. Those two colors look only slightly closer to me than red and green. They look as distinct to me as green and blue.
posted by bukvich at 11:05 AM on January 12, 2012


By the way, I really like your explanation of Heisenberg Uncertainty because a pet-peeve of mine is when it's characterized as a practical measurement problem, e.g.: "by measuring something you disturb it" and not as a something fundamental, which it is.

How do you know? You have an apparatus configured which reports measurements. That's all you know. You have no idea what's actually happening in the apparatus, only the measurements it records. I don't think there's anyway to distinguish if it's a measurement problem.
posted by empath at 11:22 AM on January 12, 2012


Actually, there are definitely perceived sounds that originate from a combination of stimuli: the best example is "phantom bass", where the sound is a series of harmonics based on a fundamental but lacking the fundamental (for example 440, 660, 880, 1100) and in many cases we will "hear" the fundamental (220Hz) anyway. An example (LOUD - careful with your volume):
falling phantom bass
Actually, you can hear several phantom pitches in that sound file, both a falling and a rising pitch well below the actual frequencies of the component sounds: the file is just two notes, a constant 1760Hz and a tone slowly rising from 1760 to 2000 Hz.
posted by minutephysics at 1:33 PM on January 12, 2012


And Ivan, any chance you'd be interested in helping write a video explaining human color perception in all its glory?
posted by minutephysics at 1:34 PM on January 12, 2012


"Sure it is; there are three color opponent processes. Red vs. Green, Blue vs. Yellow, and White vs. Black."

No, it's not that simple. I don't understand why people (who should know better) want to describe color vision as a simple physical process when, in fact, it's a product of complex cognitive processing.
posted by Ivan Fyodorovich at 6:12 PM on January 12, 2012 [1 favorite]


minutephysics, here's a nice article that deals with some of the things I've been saying in a rigorous manner. The final sentence of the conclusion: "Spectral spaces are not color spaces. They lack specific relationship to human color vision and the term color space should not be used for them.

I would love to help make such a video. Unfortunately, I don't think it would be possible to do succinctly, the subject is too complex. This thread demonstrates it.
posted by Ivan Fyodorovich at 6:24 PM on January 12, 2012 [1 favorite]


No, it's not that simple. I don't understand why people (who should know better) want to describe color vision as a simple physical process when, in fact, it's a product of complex cognitive processing.

I'm not sure what processing you are referring to over and above what I described in my linked comment. There's projecting the full spectrum down to RGB and transforming RGB into the opponent processes. What other processing am I missing here? Sure, there's a bunch of stuff that's going on that might affect the perceived color in the end like lateral inhibition, temporal filtering, and habituation, but the result of all of that still falls into the three dimensions formed by the color opponencies. Yes, as a result of your complex processing, you see a square (that would otherwise be gray) that is currently surrounded by blue to be yellow. But you don't see it as gurple or any other color not otherwise describable by the three opponencies.
posted by a snickering nuthatch at 8:37 PM on January 12, 2012


Sorry, wasn't intentionally trying to get your goat by calling the output of short, medium, and long-wavelength sensitive cones "RGB"; that was just a shorthand.
posted by a snickering nuthatch at 8:38 PM on January 12, 2012


there's many more colors than just fuschia that are not in a rainbow; that is to say, cannot be be produced in our perception via a single frequency of light. Brown is a frequently mentioned example.

Given that brown is just a really dim yellow, and given the availability of single-frequency yellow light generators, is brown actually a good example?
posted by flabdablet at 9:16 PM on January 12, 2012


"I'm not sure what processing you are referring to over and above what I described in my linked comment. There's projecting the full spectrum down to RGB and transforming RGB into the opponent processes. What other processing am I missing here?"

I'm pretty sure that the opponent processes are not as simple as you describe. It really sounds like you're describing a fifty year-old model of color vision. For the model of color perception to be three-dimensional on the basis of those three opponent processes, they each have to be independent of each other. But they aren't.

Even if they were, that wouldn't guarantee that human color perceptual space is three-dimensional any more than the dimensionality of EM determines the dimensionality of color perception. That is to say, you can model EM, you can model the neurological response of the cones, you can model the opponent system, and you can model the final result of human color perception and none of those are required to have the same dimensionality.

This isn't actually as complex of a cognitive process as it might be, and so it's very likely that the dimensionality of color perception will, in the end, closely reflect the dimensionality of its most direct sensory neurological precedent. But that this is the case shouldn't cause you to make the assumption that it necessarily must be the case. The dimensionality of a full and accurate model of something is not a physical property. It is a property of the model which most fully describes that physical process.

Yes, this leads into tricky philosophical areas of physical realism and in the domain of, for example, physics, I'm inclined to accept that the mathematical models which (seem to, anyway) fully describe physical processes reflect, in their structure, something true about reality. Again, those are deep waters and I'm not too interested in wading into them because, as it happens, for almost all purposes I'm a naive realist in the way that science is and most scientists are, but when I put my (amateur) philosopher hat on, I am not quite so naive.

But, really, when we're deep into the squishy and speculative domain of human perception at high levels of cognition, I am much less inclined to either assume that a good model represents anatomy, nor that the level of anatomy that we're presently doing is good enough to hand us that model on a platter. That is to say, at the moment, I'm interested in extensive psychophysical experimentation on color perception, and whatever mathematical model most fully describes it. It can be informed by biology—indeed, if nothing else, current biology can tell us what models won't and can't work. It can gives us clues to what models are likely to work. Sure. But, basically, I don't feel like it's useful right now to think that the dimensionality of human color perception must in some respect correspond to the dimensionality of some already well-known neurological process. (Though, when we eventually know enough it will.)
posted by Ivan Fyodorovich at 5:30 AM on January 13, 2012


I'm inclined to accept that the mathematical models which (seem to, anyway) fully describe physical processes reflect, in their structure, something true about reality.

I completely fail to understand the respect so frequently given to the process of treating this completely tautological observation as if it were something deep.

Again, those are deep waters

Only if we believe them to be so.

It is demonstrably the case that we can construct testable theories about reality, that people who do indeed test these while paying sufficient attention to experimental design will either get compatible results or reasons to construct more useful theories.

It is also demonstrably the case that any fool can make up any wild-ass guess about how things are, and find other fools for followers.

It is demonstrably the case that each of us carries around a personal set of subjective opinions and beliefs about the world and our own relationship to it, and that some of these belong to the first category above, others belong to the second, and others still are purely personal and idiosyncratic.

And it seems perfectly clear to me that this is about as deep as the philosophical waters around this issue ever need to get, and that people who try to treat them as any deeper are indulging in mere intellectual masturbation: a pleasurable pastime surely, but worthwhile for no better reason than that, and vastly unlikely to add anything to our store of useful knowledge. Why am I wrong?
posted by flabdablet at 8:03 AM on January 13, 2012


I understand your point better, Ivan, but a quick quibble:

I'm pretty sure that the opponent processes are not as simple as you describe. It really sounds like you're describing a fifty year-old model of color vision. For the model of color perception to be three-dimensional on the basis of those three opponent processes, they each have to be independent of each other. But they aren't.

You appear to be switching between two (related) definitions of independence here.

The 3 opponent processes are linearly independent (i.e. given two of the values, you cannot with certainty derive the third value), which is sufficient to say that the space they span is 3 dimensional.

The way in which the opponent processes are not independent is that they lack the property of statistical independence (e.g. you can't have a percept that is fully red, fully blue, and fully black). But that doesn't mean that the volume of colors that can be perceived isn't three dimensional; it just means that it doesn't form a cube or octant.
posted by a snickering nuthatch at 8:48 AM on January 13, 2012


Given that the rod cells have a spectral response that peaks about halfway between that of blue cones and green cones, and given that the slope of this response is quite steep in the spectral region around 550nm where the responses of red and green cones are roughly equal but there's virtually no response from blue cones, and given that there are many more rods in the outer parts of the retina, I would expect to be able to handcraft spectra with twin narrow peaks either side of 550nm that show Red vs. Green, Blue vs. Yellow and Black vs. White color opponencies identical to those for a spectrum with a single narrow peak right at 550nm, and find that these are in fact visually distinguishable given a sufficiently large visual field.

Does anybody know whether this has been done? Because if it has, and if it works, that would mean a topological map of color space needs four dimensions.
posted by flabdablet at 7:57 PM on January 13, 2012


I would expect to be able to handcraft spectra with twin narrow peaks either side of 550nm that show Red vs. Green, Blue vs. Yellow and Black vs. White color opponencies identical to those for a spectrum with a single narrow peak right at 550nm, and find that these are in fact visually distinguishable given a sufficiently large visual field.

Does anybody know whether this has been done? Because if it has, and if it works, that would mean a topological map of color space needs four dimensions.


That's tricky. We know that information from the different kinds of cones gets munged together right away in the thalamus, but I don't think that this information gets integrated with that from the rods until after V1. I would suspect that this would result in a qualitative difference between distinctions derivable from the information from cones and distinctions that also rely on the information from rods. I'm not sure- someone should do this experiment- but your two spectra might be visually distinguishable experiences without people labeling the difference as one of "color".
posted by a snickering nuthatch at 7:00 AM on January 14, 2012


might be visually distinguishable experiences without people labeling the difference as one of "color"

I'm wondering whether it could be the umami of vision :-)
posted by flabdablet at 4:23 PM on January 16, 2012 [1 favorite]


It would also be interesting to see how any such visual distinction varies with illumination. I'd expect it to disappear for light bright enough to saturate the rods.
posted by flabdablet at 4:27 PM on January 16, 2012




« Older Clipart Covers   |   ...act now before Canada becomes too similar to... Newer »


This thread has been archived and is closed to new comments