Color theory, revisited

PART 1/4

This post is a continuation of and response to many previous posts on color, generally, and color theory specifically:

• http://www.styleforum.net/t/89372/color-and-complexion
• http://www.styleforum.net/t/228299/pocket-square-review-an-ongoing-series
• http://www.styleforum.net/t/36332/color-theory-wtf-is-it

It's motivated by a visual art student's dissatisfaction with existing theories and a color-enthusiast's obstinacy in rejecting skepticism about the very possibility of color theory in the context of classic menswear.

Before we begin, though, let me deflate what's about to come by saying that the most important thing to know about colors, when it comes to menswear, is which and which combination to avoid. Yet because this is such common knowledge, I'm actually not going to talk very much about it. What I'm going to focus on instead are principles for good color coordination. This is the stuff that Alan Flusser or Carole Jackson, among others (Cognac, Kentner, Caygill, etc.), attempt to theorize but have done so problematically. Perhaps in an attempt to sell the audience, these authors never explicitly mention that the nut they're (and I'm) trying to crack is really of marginal significance to the evaluation of an outfit – esp. by anal SFers. With the exception of the aforementioned provocatively bad colors, there's really not much of a male culture of judging one's wardrobe on the basis of palette. And, without the culture, it's easy to focus attention on other elements of an outfit, which then steal one's attention entirely.

Yet, this being said, almost all the outfits that I've found memorable feature what I think to be particularly great colors. And, it must be emphatically stated, what follows reflects my (considered) sensibilities and mine alone – it's not the color theory but my color theory. It (i.e. my sensibility) is historically informed to the extent that readily accessible historical resources enable, but, beyond this, I've no evidence for its universality (though I've my suspicions). In fact, it's precisely in wanting some evidence that I'm sharing it on SF. I hope to go further than previous commentators by clarifying how I judge a palette so that anyone who disagrees with me will do so with a better understanding of his own aesthetic sensibilities.

Let's begin with some definitions. The following three concepts are, in my experience, all that one needs to consider to master color in menswear. Concepts like contrast and season popularized by Flusser and Jackson, respectively, can be analyzed in terms of these three concepts. And these three concepts constitute an arguably preferable toolkit when thinking about color as one of the biggest problems with previous alternatives, as we will see, is their tendency to inspire overgeneralization.

Hue is that property of a color that places somewhere on the wheel below. Some people call hues the pure colors, referring to the fact that a hue (e.g. orange) is neither tinted (e.g. light-orange) nor shaded (e.g. brown). There's some definitional confusion in such a reference, but it conveys the concept usefully enough.



Value is determined by adding black and white to a hue. When we add black to a color, we're increasing its value (or shading it) and when we add white we're decreasing its value (or tinting it). Brown or tan, as just mentioned, is a shade of orange and light-orange is a tint of orange.



Saturation is, very roughly, how much gray is in a particular color. Adding gray to a color tones that color, creating a tone of that color.



There is a host of additional resources for more technical details about these concepts (wikipedia, I've found, is quite useful for additional information). But this rough understanding is enough for the purpose at hand.

A few additional points must be taken into consideration at the outset. In the following posts, I will be using many comparison photos to illustrate my arguments. Preferences for different palettes are best est. by one-to-one comparisons. Our aesthetic judgments tend to be bent by thoughts introduced at the moment of judgment. To verify my arguments against your own taste, then, select which photo you prefer before reading the text. For the sake of time, I'm using several shortcuts on my photo editing software that leave sometimes unappealing residue. You'll have to take this into consideration as you evaluate my arguments. Finally, following the precedent, I will be focusing on the staples of classic menswear: jacket, tie, shirt and pocket square. With some creative imagination, of course. my arguments can be more generally applied.

PART 2/4

Let's start with coordinating colors in an outfit without considering complexion and hair color. As would anyone applying my theory in front of the mirror, I will individually consider each of the three concepts just defined, beginning with hue.

According to standard theory of hue coordination, contrasting hues – hues opposite on another on the wheel above (e.g. navy and brown) – triads – hues 90 degrees from each other on the wheel – and analogue hues – hues adjacent to one another (e.g. green and blue) – are good combinations.



While sometimes useful, this must be accompanied by some additional considerations. First, when taken together, the standard theory actually leaves us with very few bad combinations – fewer, in any case, than I think there are. In particular, I've only gravitated toward one combination – red and blue – in line with the triad principle with the same intensity that I do toward contrasting and analogous hues. In addition, this theory, which is as far as some commentators will go, provides us with no help when it comes to grays and black (both colors but not hues). There is simply no general principle for great combinations involving these two colors – tan and black, gray and chocolate, etc. – and they must instead be individually memorized.

Second, there are certain combinations of hues that evoke particular connotations – of a very subjective sort – that also affect what one considers a good color combination. For me, these are the colors of sport teams (e.g. blue and orange (Syracuse), purple and yellow (Lakers)) and politicians (e.g. navy and red). Even if they appeal to my more immediate sensibilities, I can't help but make associations that render these combinations trite and unappealing.



Finally, there are really only two classic hues for jackets – blue and orange (tan, camel, brown are tints and shades of orange) – and respecting the classics is an assumed value in this post (and this forum). The colors of the tie, square and shirt must to some extent correspond with that of the jacket. This means that, with the exception of the fabulously clothes obsessed, men can get by by memorizing a list of tried-and-true combinations without really taking generalizing principles to heart.



Having determined a set of hues that work well with one another, one must next consider value. As shirts are usually light and a pocket square too small a canvass, value considerations apply most urgently to the tie-jacket combination. When a significant incongruence occurs here the lower-value item seems to recede to the background, an unflattering effect.

In the following images, I only adjusted the images by adding white or black to the original photo (which I'll leave unidentified). One of the two pictures, in each example, show the tie and jacket on different "perceptive planes," with a less than ideal result.







Saturation, altered by adding or taking out gray from a color, is closely related to value. It's not a surprise, then, that the same basic concern applies when harmonizing colors in terms of saturation.



The same can be said about the saturated palette in the tie on the right, dulled to a more congruous color on the left.



For the reason mentioned above, more latitude applies when considering pocket squares as they occupy a small visual space. However, I still tend to shy from strong contrasts.



Harmony in value and saturation, of course, is a principle often flouted. It's common nowadays to see a provocatively high-saturated color on accessories.



Even excluding complexion and hair, though, my intuitive judgments reflect a prioritization of coherence and harmony. Consequently – and when it comes to ties, jackets and squares in particular – I'm not a big fan of the trend.

Great thread. I am a fan of color theory and I wear the blue/brown/yellow and blue/green combos frequently. However, I have noticed the value (tint and shade) is also crucial but I don't have a unifing theory that ties it all in.

Looking forward to the rest.

PART 3/4

Let's now bring hair color and complexion into the discussion.

The addition, I submit, doesn't really change the fundamental principles identified above, and it's when we apply those principles to consider complexion that problems with alternative theories emerge. Here, I'm referring to Allen Flusser's theory of contrast. According to Flusser, men with high-contrast features should wear high-contrast outfits and men with low-contrast features should wear low-contrast outfits so that the outfit doesn't draw attention away from their faces. These claims are vague and incomplete.

It's true that contrast – by which I mean (Flusser never explicitly clarifies) contrast in the value of adjacent items – draws the eye. If a high-value item is placed adjacent to a low value item, creating contrast, the eye will be drawn to two items more than if they were closer in value. But, while Flusser relies almost exclusively on the concept of contrast, there are a number of other phenomena that also draw the eye:

loud patterns



highly saturated colors



In addition, returning to the point made at the very beginning of the thread, other than in a comparative situation, it matters little whether or not attention is drawn to the face. By all accounts, for instance, low contrast Redford looks great in a classic tux – black and white, as high contrast a palette as they come.



Conversely, high-contrast East Asians do just fine in light spring gear.



And Flusser, with a mane that contrasts little against his skin, often doesn't take his own advice.



With no comparison readily in mind, people simply don't judge one's appearance by contrast. Still, when we consider an one-to-one comparison, high-contrast men do indeed look better in high-contrast outfits. After all, the picture on the right no doubt looks better than the one on the left.



Also, I give you (badly edited) brunet Redford.



I argue, though, that the story doesn't stop at contrast. Let's introduce a different comparison, this one between two low-contrast outfits, albeit two outfits that are antipodal on the value scale.



Isn't the improvement from left to right nearly as stark? Of course, the dark shirt and tie in this photo are suffocating, but this, I'd argue, is not so much because of the relationship of the outfit with the wearer as that of the tie and, in particular, the shirt with the jacket. Especially when worn with a jacket and tie, a dark shirt rarely looks good. Consider these two high-contrast outfits. Certainly, one is preferable to the other.



Or, consider this example of an African American gentleman who Flusser cites as someone whose complexion requires a low-contrast palette, provided for us in the form of a yellow jacket and beige shirt. I submit that he looks even better in this higher contrast alternative on the right.



Much more important than contrast, then, is the fact that the darker jacket in both examples provides a large piece of color that plays off the wearer's skin or hair. To a large degree, the contrast coordination principle simply follows from this fact and the principles we've examined above, principles about the internal coherence of an outfit with no regard whatsoever to complexion and hair color. Consider how great Redford looks here in a jacket that corresponds with his hair. If either of these pictures were shown on its own, it would solicit plenty of compliment. It just so happens, though, that the light jacket looks better with the light tie – regardless of Redford's hair and skin colors.



Contrast matters, but it's far from sufficient for a consistently useful theory of color coordination. For a more effective set of principles, we must revisit hue, value and saturation.

PART 4/4

Just as the hues in one's outfit must compliment each other, they must also compliment one's complexion and hair. Our asian gent has more yellow in his skin than the sky-blue shirt and square require but nowhere enough green to compliment that royal purple tie. With the tie in an analogue of his skin color, the shirt/square in his skin color's direct contrast, and his suit a good old dark navy (that still reflects his dark hair nicely), the image on the right is much more harmonious in hue. (Note all the principles for matching pieces of an outfit to each other suggested above are still in effect here.)



It's close, but I'd say that Mr. LeBeouf, with his orange-red hued hair looks better in the blue-green jacket on the left than the blue-purple one on the right.



For our commander in chief, very minor tweaks in hue, removing the green entirely from the jacket and adding a bit more yellow to the tie, can have significant positive effect. (Focus, here, on the person as opposed to the background, which is so verdant that it makes an otherwise unattractive moss green tolerable.)



As with the tie and jacket discussion above, a consistency in value is required in order to keep the wearer's face and his outfit in the same plane. The photo on the left, as noted by the SFer who posted it, plays to the value of each president's complexion and hair well. Notice how when I give each man's outfit the other's palette, the balance is significantly disrupted: Obama's head seems, in contrast to his suit, to creep into the foreground while Bush's head recedes to the back.



For those lucky enough to still have hair, an obvious question at this point in the discussion is whether the outfit should coordinate with one's hair color or one's complexion. With hues, the examples above hopefully show, one must consider both. But the same principle cannot be applied to value considerations for a clear reason. Consider the colors on this young man below – dark brown hair and light tan/peach skin, complete contrasts in value. For me, it seems pretty obvious that the priority should be given to the hair color. In a darker jacket that complements his hair in hue, his face is nicely framed by the jacket and tie in the photo on the right whereas his head floats incongruously above the colors on the left.



However, for someone with the inverse contrast (i.e. high value complexion and low value hair), the outfit must correspond with the face. The justification, here, is essentially Mr. Flusser's principle. In order to avoid the so-called clothes-wearing-the-man effect, we must direct attention to – or, in any case, never direct attention away from – the face.



These claims drag me into controversial territory. Let me reiterate my earlier point, then, that color is only one part – an often marginal part – of one's outfit. It's not that, for instance, someone with dark hair cannot wear light jackets, but when I'm presented with a comparison, I'm happy to argue that, generally, he would look better in a darker jacket. Here's one more illustration.



I conclude with saturation. an outfit that's significantly more saturated than one's complexion brings the outfit to the foreground and distracts from his face – the definition of the clothes wearing the man. This is another reason, I would guess, that classic menswear rarely features very saturated colors.





As two saturated colors each make the other appear more vivid, an outfit that's insufficiently saturated, in comparison to the face, recedes into the background and insufficiently complements one's hair and/or complexions.





With saturation, the answer to the face-or-hair question echoes that to the question in the context of hue – that is, both hair and face must be considered together. Consider the following two sets of image, which show two complexions and outfits of different saturation (hair color remains the same).



The top-left photo appears to me the most attractive. The more saturated outfit brings out the President's hair and skin colors better than that on the top-right. However, what if he had the complexion shown in the bottom images? While the bottom-left image nicely brings out the olive color of his hair, it makes his face look ashen in contrast. The bottom-right, conversely, corresponds nicely with his face, but squanders the color of his hair. When there is significant contrast in saturation between the hair and skin colors, then, a balancing act is required.






The theory I've provided here leaves many problems unaddressed. It disregards luminosity, how different fabrics reflect light differently, what constitutes tasteful as opposed to kitsch matching (i.e. "too matchy-matchy"). Nevertheless, I've found it much more useful than alternative color theories. Keeping hue, value and saturation in mind, one will be much more nimble and discerning when putting together new outfits.

Thoughts?

I think lighting is something people do not consider enough in this type of discussion. Often the unspoken assumption is that the person in seen in ideal lighting. For example, I think the high contrast look of the tuxedo is best in relatively dim light. A lot of the subtlety of coloring, that is best in daylight, is lost at night.

poorsod said:

I think lighting is something people do not consider enough in this type of discussion. Often the unspoken assumption is that the person in seen in ideal lighting. For example, I think the high contrast look of the tuxedo is best in relatively dim light. A lot of the subtlety of coloring, that is best in daylight, is lost at night.

Click to expand...

Absolutely. As I mentioned, my posts do not consider how different colors are produced by different fabrics under the same lighting condition. It's simply difficult to find photos of the same fabric in clearly different lighting conditions to do so on SF. A dark tux, though, can be extremely tinted under direct light. I've seen photos of lapels that turn silver from camera flashes.

Have you read vox's coherent combinations thread?

http://web.archive.org/web/20120711...oughts-on-coherent-combinations-for-beginners

He makes the argument for matching the formality of each item of clothing, but I think a lot of it can also be explained by matching saturation or tone of each piece. For example, the "city" dress mix items which are more highly saturated than "country" dress. The city suit with worsted suit and silk tie are best with clear colors and little texture. The country tweed has more muted colors and matches with other items with more muted colors too.

Yes! I enjoyed that thread very much. I think color coordination and city-country (formality) coordination are only tangentially related – precisely in the way you describe. If I hesitate to put saturated colors on a worsted wool and silk, I'd fight dentures and cuticles against putting it on corduroys and tweeds. Interestingly, though, this may mean that one must choose between coordinating colors and coordinating formality, much like the tradeoff between season and color coordinations: dark-complexioned men simply look worse in the summer pastels than light-complexioned guys.

He and I also begin from the same philosophical premise of achieving harmony and coherence. Obviously, I think he was a bit premature to close the book on color.

The "colour wheel" is a deeply flawed concept.

Objects do not have colours. Only light has colour. Colour is not an inherent property of a surface (e.g. a fabric). It is merely an interpretation, by the brain, of the light reflected from that surface into the observer's retina.

The retina sends three distinct sets of colour signals to the visual cortex, corresponding to the three types of "cone" (colour-sensing cell) found in the retina. Each cone will respond maximally to a particular wavelength (i.e. spectral colour) of light - red, blue or yellow-green - with a rapid drop-off in sensitivity to wavelengths on either side. The wavelength of light at which a cone is maximally stimulated is determined by the nature of the light-triggered chemo-electric reaction taking place within that cone. There are three different possible chemical reactions, of which each cone is capable of just one.








Colour information about a surface therefore consists of three completely independent sets of data. We therefore need three spatial dimensions to represent this data within a manifold.




The "colour wheel", however, is a two-dimensional representation. It is therefore wrong. You can easily demonstrate that it is "wrong" by considering the spectrum of visible light:



The "colour wheel" appears to take this strip and join it up, end-to-end, in a loop. This is clearly bogus.

Any theories about colour matching in menswear must necessarily be based on the human brain's interpretation of colour. The "colour wheel" is a very poor tool on which to base any such theory, as it bears little or no relationship to what actually happens within the human retina and visual cortex.

Coxsackie said:

Any theories about colour matching in menswear must necessarily be based on the human brain's interpretation of colour.

Click to expand...

I appreciate the argument but would object that it's both vague and unjustified. I think, but cannot be sure based on your post, that you mean by "interpretation" the neurophysical mechanisms of color-perception. I'm very unclear as to what, in your mind, a "[theory]... based on the human brain's interpretation of colour" would look like.

What a theory about color in menswear must be based on is whatever achieves the most aesthetically pleasing palette. The color wheel – even concepts like hue, saturation and value – constitute a very rudimentary lens through which to think about color. But for the purposes of putting together a wardrobe, I find them sufficient. The evidence is in the images provided above. However, I and, I'm sure many SFers, would be very interested if you can provide a more elaborate alternative.

You've outlined a "theory of colour in menswear", or at least, a methodology. It's based in part on the principle of the colour wheel. I'm pointing out that the colour wheel is a flawed concept. I think my argument in support of that is fairly easy to understand. The rest of your methodology is fine, but I believe your overall conception would be improved if it relied less on this flawed principle.

Here's another illustration: according to science, the manifold containing all colours perceivable by the human retina and visual cortex is at least 3-dimensional. The colour wheel, however, is 2-dimensional. Therefore, not all perceivable colours can ever be represented on any one colour wheel. What if we have an item of clothing whose precise colour is not on the wheel? How do we use the wheel in this situation?

(In fact, the chance that any one perceivable colour exists on any one colour wheel is extremely low, as the number of possible discrete points within a 3-dimensional colour manifold will be enormously greater than the number on a 2-dimensional wheel.)

Furthermore, since the functionality of the wheel relies on choosing "complementary" colours as being those on opposite sides of the wheel, we should ask ourselves what it means, in physiological terms, for two colours to be "opposite each other on the wheel". How does this correlate, if at all, to the actual firing rates of the three different cone types when presented with those two different colour types? I would posit that there is no correlation, i.e., no function which maps one to the other - except the function of the colour wheel itself. Which makes the colour wheel a bit like the old IQ tests from the 50s - the only thing they measure is the ability to perform an IQ test.

Well, black and white and everything in between are not on the colour wheel, as OP already noted, so how to deal with that in a menswear colour theory? E.g., lighter shades of grey go with other colours than darker shades. What's the theory behind that?
Anyway, interesting subject and some interesting observations as well, so thanks for posting, gents!

Coxsackie said:

You've outlined a "theory of colour in menswear", or at least, a methodology. It's based in part on the principle of the colour wheel. I'm pointing out that the colour wheel is a flawed concept. I think my argument in support of that is fairly easy to understand. The rest of your methodology is fine, but I believe your overall conception would be improved if it relied less on this flawed principle.

Here's another illustration: according to science, the manifold containing all colours perceivable by the human retina and visual cortex is at least 3-dimensional. The colour wheel, however, is 2-dimensional. Therefore, not all perceivable colours can ever be represented on any one colour wheel. What if we have an item of clothing whose precise colour is not on the wheel? How do we use the wheel in this situation?

(In fact, the chance that any one perceivable colour exists on any one colour wheel is extremely low, as the number of possible discrete points within a 3-dimensional colour manifold will be enormously greater than the number on a 2-dimensional wheel.)

Furthermore, since the functionality of the wheel relies on choosing "complementary" colours as being those on opposite sides of the wheel, we should ask ourselves what it means, in physiological terms, for two colours to be "opposite each other on the wheel". How does this correlate, if at all, to the actual firing rates of the three different cone types when presented with those two different colour types? I would posit that there is no correlation, i.e., no function which maps one to the other - except the function of the colour wheel itself. Which makes the colour wheel a bit like the old IQ tests from the 50s - the only thing they measure is the ability to perform an IQ test.

Click to expand...

I hear you, but am still unclear exactly what your critique is. Why don't you help me clarify the theory that you've in mind in contrast to which my arguments are problematic?

You say, "I'm pointing out that the colour wheel is a flawed concept." What do you think is an unproblematic conceptual alternative that achieves what you think the color wheel attempts to? Can you then provide some concrete examples of good color combinations (the objective of my posts here) on the basis of this alternative? If they are combinations inconsistent with what I've written here and are aesthetically pleasing, then you would no doubt be right and have provided – to me in the very least – invaluable insight.

It may just be that @Coxsackie is a tiny bit nettled by the rather condescending, 'know-it-all' tone of your posts. Just sayin'....