> On 8-Jan-2004, Tak To <ta...@alum.mit.edu.-> wrote > in message <9YadnWhZDuRMAGCiRVn...@comcast.com>:
> > LEE Sau Dan wrote:
> > > And given that the blending done by our brain is not degenerate, the > > > colour space we perceive is 3-dimensional. It's just a matter of > > > changing the basis.
> > The issue is that the actual signals are not like mutually orthogonal > > coordinates: R+G, R-G, G-R, (R+G)-B, B-(R+G). They do _not_ form > > a basis in the vector field sense.
> Mathematical nitpick: Yes they do. A "basis in the vector field > [sic] sense" need not be mutually orthogonal, and in fact a > basis exists for every vector *space* (modulo the Axiom of > Choice for infinite-dimensional spaces), even for spaces with no > meaningful definition of orthogonality. All that's necessary > (and sufficient) to be a basis is that every vector in the space > be expressible1 as a unique linear combination of the basis > vectors.
> 1 Interesting. It's rare anymore that my spell-checker > legitimately flags a word whose spelling I was sure of. I would > have bet the farm it was *"expressable". Is there a useful rule > for <-able> vs. <-ible>?
I actually did once see a rule of thumb for it, but was never able to find it again. But I think you were supposed to know the etymology of the root your were attaching it to, so it could be of limited usefulness. -- Peter T. Daniels gramma...@att.net
> >>>No he crecido allá [en México]. ¡Núnca! Aun una vez. > "No crecí allá. ¡Ni he ido! (?)" (Don't quite get what you were > trying to say with "aún una vez".)
Since I can still remember, I can see now that "no he crecido" was what I meant after all, even though the "he comido" in the next line must have been contamination from French or German. "I've never been born there. Never! Not even once" was what I was trying to say.
Curiously, my Spanish dictionary spells "aun," but pronounces it "aún."
B >>>Quizas es porque no he comido la verdura...
P> "Quizás es porque nunca comí la verdura."
B> > I realized within a couple hours of posting that what lousy B> > Spanish that is. ...
I have occasional problems with English, too. :-(
> ...
P> "... yo tampoco la comía."
I'm surprised I didn't come up with "tampoco," though I guess I would have misused it anyway ("No la [comía] tampoco."
Hard to believe I took a semester (2nd semester, in fact) of college Spanish in 1958!
In article <b0461a9.0401082252.3fd93...@posting.google.com>, Javier BF says...
>> Nope. It is easy to apply a transparent red glaze over a solid green, or a >> transparent green over a solid >> red.
>And the result is yellowish brown, not "red-green". >Unless the glaze features such a distribution of >transparency levels as to render some spots in >reddish brown and some others in olivish brown, >so that some reddish and greenish tinges can be >observed here and there within the same colour >field - which all in all is the same effect of >'chromatically-textured' surfaces that I talked >about when I mentioned my "reddish-greenish" >brown shirt.
You don't know what you're talking about. What you claim is impossible is an everyday experience for every artist that knows how to mix paint.
>> Perhaps you are not aware that oil paints are not surface-reflective >> colors, they are transparent >>and get much of their color effect from light refracting off internal surfaces >> inside the paint. But I >> achieve my desired reddish greens by merely blending the appropriate colors >> together. The colors >>can't be photographed because they are outside the color gamut of photographic >> film, let alone >> anything an RGB monitor could represent. This is a fairly common problem with >> photographing >> paintings, and other transparent/translucent intensely-colored objects like >> certain gemstones.
>OK, go tell the guys at the University of Utah >about your discovery of "red-green" colour. I'm >pretty sure they'll be extremely interested in >your practical refutation of the neurophysiological >mechanism of red-green contrast.
I don't have to, the MIT/Harvard team does such a good job.
Let me see, who should I trust, the Retinex theorists at MIT and Harvard, America's most prestigious schools and research institutes, or University of Utah, one of Utah's finest state schools and.. um..
>> Color theory has come a long way since the Pointillists. Check into it.
>And it hasn't refuted the fact that surfaces with >small points of different colours are seen as motleyed >multicoloured surfaces when seen from a close point, >but become blended when from afar.
Which has nothing to do with what I'm arguing. You're arguing that reddish greens can only be seen in discontinuous steps like pointilism. I say it isn't so.
>That's a perceptual >fact. Or is your comment aimed at negating that >observable fact that everybody can check (and that >I'm sure everybody here has already checked more >than once)?
I don't see anyone agreeing with you, except that other nutball from sci.lang.
>> >This kind of complex effects in colour fields are like >> >the chromatic equivalent of those paradoxical images >> >that one second look like a vase and the next look like >> >two faces and the next like a vase and the next like a >> >two faces...
>> Um... no. Those illusions are completely unrelated to color perception.
>Can you understand English? Where did I say that >those illusions 'are related to colour perception'? >All I said was that _the paradoxical effect_ of >surfaces with reddish and greenish tinges "is >like the chromatic equivalent of" _the paradoxical >effect_ of those images that can be interpreted >in contradictory ways.
Apparently I can't understand English. Either that, or you aren't speaking English. Your restatement was pure gibberish. What does it mean to be "like the equivalent of" something?
>> >and when two adjacent spots, one red >> >and the other green, are seen from a distance far >> >enough as to not allow the eye to perceive them >> >as separate spots anymore, the resulting blended spot >> >does not look 'red-green' but 'yellowish/brownish'.
>>Well now I KNOW you're either a troll or an idiot, because everyone knows that >> complementary colors >>don't add up to yellow/brown, they add up to neutral grey. If you get a brown or >> yellow, it's because >> they're not true complements. In art school, we used to call that color >> "Painting 101 Brown."
>Need I remind you of your previous mnemonics for >remembering what the complementaries of red and >green are?
Red + green in their most intense mix form a subordinate tertiary black, slightly on the yellow side. You ought to read Arnheim, every artist has.
>Everyone knows that red and green paints do not >add up to grey but to some dull brown, like blue >and yellow aren't complementaries and their paints >add up to green. But it seems you are unable to >distinguish between complementaries and opponents. >So if there's a troll or an idiot here, that's >evidently you.
>> >> >Cyan is the 'RGB complementary' >> >> >of red (i.e. its retinal afterimage or complementary >> >> >RGB value) but _not_ its 'qualitative opposite', >> >> >because RGB values are not the basis of how we >> >> >'think' colours at the cerebral level: the RGB >> >> >output of cones goes through a network of retinal >> >> >nerves that turn it into an opponent-process signal >> >> >{white-to-black, red-or-green, blue-or-yellow} and >> >> >_this_ one is the signal that is finally sent to >> >> >the brain and _not_ the initial 'raw' RGB signal.
>>>> I think you better bone up on your neurology, that model of color vision is >> >> obsolete. Go study >> >> Edwin Land's Retinex theory.
>> >I did not claim to be telling the whole story of >> >how single colour spots then interact in complex >> >chromatic fields like that of gold or phenomena >> >like colour constancy or colour interaction and >> >contrast in adjacent surfaces.
>> Right, so you simplify it beyond any reasonable explanation.
>Weren't you expecting me to post a thorough formal >explanation of the neurophysiological mechanism >of human vision here, were you?
No. At this point, I merely expect you to obfuscate the theory of vision sufficiently to push your agenda, whatever the hell that is.
>And besides, what's that that I supposedly simplified >beyond any reasonable explanation? The thing about >colour opponents doesn't require any more elaboration >in order 'not to be simplified beyond any reasonable >explanation'. In fact, it shouldn't need any >explanation at all, because it is such an obvious >perceptual fact that white/black, red/green and >blue/yellow are opponents and each pair makes for >the best possible chromatic contrast.
Riiight. Which perfectly explains why it is accepted as an obvious perceptual fact that writing on yellow legal pads provides the best contrast for viewing handwritten black ink.
>All further >elaboration would be meant at explaining _why_ this >_fact_ is like it is, but a thorough explanation of >that is out of place here: go check the link to >the University of Utah I provided; they have more >information on their website about the mechanisms >of human vision than most people interested in >the subject are likely to ever care to read >from beginning to end.
Read it. Now go read about Retinex, which completely debunks all that crap.
>> Or more simply stated, these are simultaneous contrast effects. You ought to >> read Leonardo's Treatise >>on Painting, these effects have been known and used for centuries, and requires >> little reference to >>neuropsychology. However, for a full explanation, you need to read about Retinex >> theory, which was >> specifically developed to account for "color constancy" problems such as you >> describe.
>So? >Of course those effects have been known and used >for centuries and everybody I know is aware of them. >I'm sure we all here have seen at least half a >hundred times those optical illusions that illustrate >them. And I'm sure most are also familiar with the >work of well-known Op-Art artists like Vasarely.
>> >Or perhaps you have a wonderful different theory >> >that explains why the only colours that _actually_ >> >_do not look_ like mixes of any other colours are >> >precisely white, black, red, yellow, green and blue. >> >But first, have a look at what is already known >> >about the general correctness of the red-green >> >and blue-yellow opponents. E.g. here:
>> I have no idea what you're blathering on about here.
>How so? It's *you* who's trying to challenge such >obvious facts as the opposite nature of the red >and green perceptions claiming that you do see >"red-green" and saying that the opponent-channels >theory is "obsolete" and all that.
>> >"The Opponent Color Theory of the 19th century physiologist Ewald >> >Hering (Hering, 1964; Hurvich, 1981) derived by the analysis of >> >subjective human color vision is in general correct, although the idea >> >of opponent colors was described earlier by Goethe and Schoepenhauer. >> >Certain colors are not perceived together, i.e. they do not mix. We >> >never see bluish-yellows or reddish-greens. This is consonant with the >> >neural comparisons described previously. The yellow detector is always >> >inactive when the blue detector is active and vice versa. A similar >> >situation occurs for the neurons responding to red or green."
>> Right, like I'm going to waste my time debunking a discredited 125 year old >> theory developed by >>poets, especially a theory that contradicts the evidence of my own eyes. Go read >> Edwin Land.
>So now I know you're either a complete idiot or >you come from Mars, since that theory, first, has not >been discredited but there's already quite an amount >of scientific knowledge about the neurophysiological >mechanisms underlying the red-green and blue-yellow >contrasts, and second, it only explains what is >fairly obvious to any human being that features >normal vision abilities.
>Go tell the guys at the University of Utah that >they're idiots for supporting a "discredited theory".
>> >It's you who isn't understanding anything, because >> >I'm not talking about _mixing paints_. If you mix
> JBF> NCS (Natural Color System) places the 3 pairs of > JBF> opposed percepts orthogonally along each of the > JBF> three 3D axes. The result is among the most > JBF> 'intuitive' colour space arrangements I've seen: > JBF> > JBF> http://www.ncscolour.com/engelsk/main.asp?menu=1&main=meny1/page1.asp
> Red in on the opposite of green, hence not orthogonal to it.
I guess you meant "Red is not the opposite of green".
But red _is_ (perceptually) the opposite of green. If you place white and black side by side, you will perceive a perfect contrast between both colours, and the same perfect contrast you will get if you match red with green or blue with yellow.
I've already pointed out half a dozen times that we do NOT _think_ colours in terms of the [R,G,B] composition of stimuli (and that's why complementary colours, like red and cyan, do not appear as perfect contrasting opposites, like red and green, except for the achromatic gamut where RGB complementaries and perceptual opposites actually match), so the orthogonality or not of those opposites in terms of their RGB components (i.e. white [1,1,1], black [0,0,0], red [1,0,0], green [0,1,0], blue [0,0,1], yellow [1,1,0]) is utterly _irrelevant_ to their actual orthogonality in perceptual terms. Just define colours in terms of their _actual_ perceptual components and their vectors will feature perfect orthogonality: white [1,0,0], black [-1,0,0], red [0,1,0], green [0,-1,0], blue [0,0,1], yellow [0,0,-1].
I think all the above it's pretty obvious and shouldn't require any explanation: Can you tell me offhand what is the RGB code for ochre or for lilac or for olive? I can tell you offhand that ochre is a "dark reddish yellow", lilac a "light reddish blue" and olive a "dark yellowish green", but offhand I have no idea what RGB composition will produce those colour sensations that I can so easily define in 'intuitive' terms, I need to go and check a chart listing the 'unintuitive' RGB compositions corresponding to them (*).
> But the octahedron arrangement is indeed intuitive.
Of course, because it is based on the basic qualities that humans _actually_ perceive in colours (light, dark, reddish, greenish, yellowish, bluish), not on the RGB composition of the stimulus that will produce a certain psychological colour sensation.
Cheers, Javier
(*) Just checked those RGB compositions: A shade of ochre is RGB [CC,99,00], i.e. lots of red, quite an amount of green and no blue (Can you detect any green component in that shade of ochre? I certainly cannot, all I detect there is mixture of red, yellow and black perceptions). A shade of lilac is RGB [CC,99,FF], i.e. just like ochre but adding a ton of blue (Do you perceive lilac to be the result of adding blue to ochre?). And a shade of olive is RGB [99,99,33], i.e. take some red out of ochre and replace it with a bit of blue (Would you say that olive has the appearance of a bluish colour?).
> JBF> NCS (Natural Color System) places the 3 pairs of > JBF> opposed percepts orthogonally along each of the > JBF> three 3D axes. The result is among the most > JBF> 'intuitive' colour space arrangements I've seen: > JBF> > JBF> http://www.ncscolour.com/engelsk/main.asp?menu=1&main=meny1/page1.asp
> Red in on the opposite of green, hence not orthogonal to it.
Umm... ah, now I see - sorry for my previous misunderstanding of I what you meant here: You didn't mean that "Red is not the opposite of green", you meant that "Red is (placed) on the opposite (side) of (the same axis as) green" and thus red and green are not orthogonal to each other, am I correct?
Well, of course, red and green are not orthogonal to each other, because they are opposites (like +1 and -1) - it's blue/yellow what are orthogonal to red/green.
> Jim Heckman wrote: > > 1 Interesting. It's rare anymore that my spell-checker > > legitimately flags a word whose spelling I was sure of. I would > > have bet the farm it was *"expressable". Is there a useful rule > > for <-able> vs. <-ible>?
> I actually did once see a rule of thumb for it, but was never able to > find it again. But I think you were supposed to know the etymology of > the root your were attaching it to, so it could be of limited > usefulness.
The closest I have ever seen resembling a rule of thumb for this has to do with the conjugation of the Latin verb to which the suffix is being added. If the verb was in the 1st conjugation, then the suffix is "-able", otherwise, "-ible".
It appears that we are using "orthogonal" differently. Please note the use of "orthogonal" in our previous message.
Tak To wrote:
TT> [...] TT> The term "primary color" in perhaps confusing to people TT> unfamiliar the terminology in opponent-channel theory. Most TT> people tends to assume that primary colors are orthogonal TT> to each other, whereas these signals are not. A better term TT> might be "cardinal colors" (cf cardinal vowels).
I meant that people tends to assume that primary colors are like the (x,y,z) coordinates in a 3-D space. R,G,B in would be like orthogonal primary colors in the stimulus space.
Javier BF wrote:
JBF> Yes, the term "cardinal colours" appears to be JBF> less confusing. Although those cardinal colours JBF> can actually be arranged in an orthogonal fashion, JBF> because we do not 'think' of them in terms of JBF> those types of signals you mention. E.g. the NCS JBF> (Natural Color System) places the 3 pairs of JBF> opposed percepts orthogonally along each of the JBF> three 3D axes. The result is among the most JBF> 'intuitive' colour space arrangements I've seen: JBF> JBF> http://www.ncscolour.com/engelsk/main.asp?menu=1&main=meny1/page1.asp
TT> Red in on the opposite of green, hence not orthogonal to it.
JBF> I guess you meant "Red is not the opposite of green".
No. I meant Red and Green are on the opposite ends of the same axis, so they cannot be orthogonal to one another.
> It appears that we are using "orthogonal" differently. > Please note the use of "orthogonal" in our previous message.
No, actually we weren't using the term differently because I said "NCS [...] places the 3 pairs of opposed percepts orthogonally along each of the three 3D axes". Note that I said "the 3 pairs" and not "the 6 colours", i.e. I was referring to the orthogonality of the pairs of opposites B/W, R/G, B/Y to each other, while each individual cardinal colour defines only one half of an axis and so, of course, halves of the same axis are not orthogonal to each other but perfectly parallel.
> I meant that people tends to assume that primary colors > are like the (x,y,z) coordinates in a 3-D space. R,G,B in > would be like orthogonal primary colors in the stimulus > space.
Though you're right that people tend to think of pure hues as orthogonal to each other, instead of as pairs of opposites as they actually are. Probably that's influenced by the fact that most usually when people are taught about colour they are told about how colours are reproduced on TV, in printing and how artists get their palette by mixing pigments.
> TT> Red in on the opposite of green, hence not orthogonal to it.
> JBF> I guess you meant "Red is not the opposite of green".
> No. I meant Red and Green are on the opposite ends of the same > axis, so they cannot be orthogonal to one another.
Yes, sorry. I realized of that later; see my other message.
>>>>> "Kevin" == Kevin Wayne Williams <niho...@paxonet.kom> writes:
Kevin> Phil Healey wrote: >> LEE Sau Dan wrote: >>> - There does not exist a multiple of 6 that does not have "6" >>> in the unit digit (using decimal notation). >> Coincidence arising from notation system.
Kevin> I guess I don't understand this one. Where are the "6" unit Kevin> digits in 18, 24, 30, 42, 48, etc.?
I wanted to say "a (positive) power of 6", not a "a multiple of 6". :P
>>>>> "Peter" == Peter T Daniels <gramma...@worldnet.att.net> writes:
>> >> - There does not exist a multiple of 6 that does not have >> >> "6" in the unit digit (using decimal notation). >> Peter> What about 12, 18, 24, and 30, and so on through the Peter> integers? >> Sorry, that's a mistake. I wanted to say powers of 6. :P
Peter> All right then!
To be more accurate: positive integral powers of 6. (So, forget about negative powers and fractional powers.)
Peter> The legend for a diagram I saw just last night (in a book Peter> on the history of Grand Central Terminal) says that the Peter> public walkways are shown in light black, the rail lines in Peter> dark black. >> It's hard to understand how they define "black".
Peter> Not surprisingly, the shades in question were gray.
>>>>> "Phil" == Phil Healey <com.hotmail@psa_healey> writes:
Phil> LEE Sau Dan wrote: >>>>>>> "Charles" == Charles Eicher <ceic...@inav.net> writes: >> Charles> Since you seem to be a bit weak on this whole logic Charles> thing, let me remind you that it is almost impossible to Charles> prove a negative, i.e. the nonexistence of something, as ........................................^^^^^^^^^^^^ Charles> you are attempting to do. >> That's possible, Charles.
...
>> - There exists no dried-water. (c.f. dried tomatoes, dried >> squids)
Phil> There are also no married bachelors or round squares. That's Phil> not proving a negative. It's playing with contradictory Phil> definitions. If you ever bothered to take an introduction to Phil> philosophy class, you'd know all about this.
So, you still haven't got it?
You can prove the non-existence of something by showing that its existence is a contradiction. Prove by contradiction is a very basic proving technique in pure mathematics. It is typically employed to prove that something can't exist.
>>>>> "Peter" == Peter T Daniels <gramma...@worldnet.att.net> writes:
>> 1 Interesting. It's rare anymore that my spell-checker >> legitimately flags a word whose spelling I was sure of. I >> would have bet the farm it was *"expressable". Is there a >> useful rule for <-able> vs. <-ible>?
I've never seen "expressable".
Peter> I actually did once see a rule of thumb for it, but was Peter> never able to find it again. But I think you were supposed Peter> to know the etymology of the root your were attaching it Peter> to, so it could be of limited usefulness.
Do you mean: "-able" <==> active, "-ible" <==> passive?
>>>>> "Tak" == Tak To <ta...@alum.mit.edu.-> writes:
Tak> LEE Sau Dan wrote: >> And given that the blending done by our brain is not >> degenerate, the colour space we perceive is 3-dimensional. >> It's just a matter of changing the basis.
Tak> The issue is that the actual signals are not like mutually Tak> orthogonal coordinates: R+G, R-G, G-R, (R+G)-B, B-(R+G). Tak> They do _not_ form a basis in the vector field sense.
A basis need not be orthogonal (and hence need not be orthonormal). The vectors in a basis only need to be linearly INdependent.
>>>>> "Tak" == Tak To <ta...@alum.mit.edu.-> writes:
Tak> And I did *not* say "orthogonal basis". While the members of Tak> a basis need not be orthogonal to teach other, they must be Tak> _linearly_independent_, and these signals are not. This Tak> should be pretty obvious given that the dimension of the Tak> color vector space is 3 and there are 5 signals.
I was thinking about that red-green and blue-yellow issue. Let's forget about the intensity for the moment, Then, you get a 2-D colour-space which can be visualized as a colour wheel. The center of the wheel is white (or a kind of grey). Distince from the centre is proportional to the degree of saturation. Angle with the X axis determines the hue. Take 0 degrees to be red, 120 degress to be green and 240 (i.e. -60) degrees to be blue.
Now, assume yellow is RGB(100%,100%,0). Then, is the colour on the radial line making an angle of 60 degress with the X-axis. The opposite, blue = 240 degrees) is exactly 180 degrees away. So, they are complements of each other. We take this line to be one axis. If you want an orthogonal axis in this 2D space, you must choose the one in the direction -30 degree or 150 degree. Call the -30 degree "red" and 150 degree "green" (only 30 degrees away from the RGB primaries). This forms another axis. And this "red"-"green" axis is perpendicular to the blue-yellow axis.
Now adding intensity back. It is just the Z-axis, that is perpendicular to the 2D plane we have above. So, we have a vector space with an orthogonal basis: red-green, blue-yellow and black-white (i.e. intensity). Of course, not all points in this space represent meaningful and unique colours. The typical approach with the HSB model is to define the colours in a cone in this 3D space, with the vertex at the intensity=0 point.
> >>>>> "Peter" == Peter T Daniels <gramma...@worldnet.att.net> writes:
> >> 1 Interesting. It's rare anymore that my spell-checker > >> legitimately flags a word whose spelling I was sure of. I > >> would have bet the farm it was *"expressable". Is there a > >> useful rule for <-able> vs. <-ible>?
> I've never seen "expressable".
> Peter> I actually did once see a rule of thumb for it, but was > Peter> never able to find it again. But I think you were supposed > Peter> to know the etymology of the root your were attaching it > Peter> to, so it could be of limited usefulness.
> Do you mean: "-able" <==> active, "-ible" <==> passive?
No; what is that supposed to mean? -- Peter T. Daniels gramma...@att.net
>>>>> "Javier" == Javier BF <uaxuc...@hotmail.com> writes:
Javier> I guess you meant "Red is not the opposite of green".
Javier> But red _is_ (perceptually) the opposite of green.
I don't perceive it that way.
Javier> If you place white and black side by side, you will Javier> perceive a perfect contrast between both colours, and the Javier> same perfect contrast you will get if you match red with Javier> green or blue with yellow.
Try to adjust your computer screen to display green text against a red background. With the perfect contrast, you should find the text easy and comfortable to read, right?
On 9-Jan-2004, LEE Sau Dan <dan...@informatik.uni-freiburg.de> wrote in message <m3oetc7tej....@mika.informatik.uni-freiburg.de>:
> >>>>> "Tak" == Tak To <ta...@alum.mit.edu.-> writes:
> Tak> LEE Sau Dan wrote: > >> And given that the blending done by our brain is not > >> degenerate, the colour space we perceive is 3-dimensional. > >> It's just a matter of changing the basis.
> Tak> The issue is that the actual signals are not like mutually > Tak> orthogonal coordinates: R+G, R-G, G-R, (R+G)-B, B-(R+G). > Tak> They do _not_ form a basis in the vector field sense.
> A basis need not be orthogonal (and hence need not be orthonormal). > The vectors in a basis only need to be linearly INdependent.
> Javier> But red _is_ (perceptually) the opposite of green.
> I don't perceive it that way.
Draw a pair of squares and place them side by side. Paint one in green and then alternate the colour of the other: chartreuse, turquoise, blue, yellow, purple, orange, red. Then see which hue appears to be furthest away from green, providing the best chromatic balance: Green and chartreuse? I don't think so. Green and turquoise? Nope. Green and yellow? Ditto. Green and blue? Look somewhere else. Green and purple? Warm. Green and orange? Warmer. Green and red? Disco.
> Javier> If you place white and black side by side, you will > Javier> perceive a perfect contrast between both colours, and the > Javier> same perfect contrast you will get if you match red with > Javier> green or blue with yellow.
> Try to adjust your computer screen to display green text against a red > background. With the perfect contrast, you should find the text easy > and comfortable to read, right?
Shapes and colours are processed in different cortical regions and the recognition of shapes is based on lightness, not on chromaticity. Reading is about recognizing shapes, not about recognizing colours; that's why it's based on lightness constrasts instead of chromatic contrasts. Thus, the best choice for reading is black vs. white. Since yellow is the hue with the highest value of inherent lightness, it makes a good substitute for white. A substitute for black may be blue (e.g. in blueprints). Blue vs. yellow make a decent pair too. But red vs. green is a bad choice for reading because the difference in inherent lightness between those two hues is small. But this is notably improved by using light red vs. dark green or light green vs. dark red, instead of pure red vs. pure green (and conversely, it is notably worsened by using light red vs. light green or dark red vs. dark green).
But if instead of reading, we are contrasting chromaticity, e.g. of adjacent areas like a blouse and a skirt or the colours of traffic lights, the most contrasting pairs of hues are red/green and blue/yellow.
>>>>> "Javier" == Javier BF <uaxuc...@hotmail.com> writes:
Javier> But red _is_ (perceptually) the opposite of green. >> I don't perceive it that way.
Javier> Draw a pair of squares and place them side by side. Paint Javier> one in green and then alternate the colour of the other: Javier> chartreuse, turquoise, blue, yellow, purple, orange, Javier> red. Then see which hue appears to be furthest away from Javier> green, providing the best chromatic balance: Green and Javier> chartreuse? I don't think so. Green and turquoise? Javier> Nope. Green and yellow? Ditto. Green and blue? Look Javier> somewhere else. Green and purple? Warm. Green and orange? Javier> Warmer. Green and red? Disco.
When I place blue and yellow, which are opposite in your theory, side by side, I don't feel this kind of contrast. I even feel that yellow blends with blue to give green.
>>>>> "Jim" == Jim Heckman <wnzrfeurpx...@lnubb.pbz.invalid> writes:
Tak> The issue is that the actual signals are not like mutually Tak> orthogonal coordinates: R+G, R-G, G-R, (R+G)-B, B-(R+G). Tak> They do _not_ form a basis in the vector field sense. >> A basis need not be orthogonal (and hence need not be >> orthonormal). The vectors in a basis only need to be linearly >> INdependent.
Jim> They also need to span the whole vector space.
This requirement is redundant. When you've got 'n' linearly independent vectors in an n-dimensional space, they MUST span the whole space.
> When I place blue and yellow, which are opposite in your theory, side > by side, I don't feel this kind of contrast. I even feel that yellow > blends with blue to give green.
I think you are taking into account other features of those hues aside from their pure chromaticity. Green has a degree of inherent lightness that lies between those of blue and yellow, and as for its 'temperature' it also lies between yellow and blue; so, in those two aspects, green is a colour "between yellow and blue". Also, green is the less salient of the pure hues, because it is the opposite of the most salient one, i.e. red. So, if one had to stay with only three chromatic colours instead of four as the most basic, those would be red (the warmest and most salient), yellow (the lightest) and blue (the coldest), while green could be obviated on the grounds of its relatively low saliency and its intermediate position between blue and yellow as for lightness and 'temperature'. But, still, it is not possible to obviate green completely as a basic colour because it is an independent basic hue with no blue-like nor yellow-like characteristics in this aspect; you cannot describe turquoise as a "yellowish blue" nor chartreuse as a "bluish yellow".
Blue and yellow, for their part, contrast notably in every aspect, while red and yellow and especially green and yellow do not contrast so much. So, if I had to choose the two pure hues that in their overall appearence appear most different from each other, the pair blue/yellow would take precedence, because apart from their contrast in hue, they contrast in 'temperature' and inherent lightness, while red/green and red/blue contrast lowly as for their lightness, red/yellow contrast lowly as for their 'temperature', and green/yellow and green/blue contrast lowly as for both their lightness and 'temperature'.
