Modern colour theory
I've just had a bunch of visitors coming to my site The Dimensions of Colour from Muzz's recent thread here on art terminology. So first off, hello and thanks for the interest! And second, I'm here if anyone wants to discuss anything to do with colour, or has any questions relating to my site, which some people apparently find quite dense. :poly136:
If you haven't checked out my site before please don't miss the section near the end on principles of colour and light for painters, which touches on colour relationships in light and shade, effects of coloured illumination, multiple light sources, and atmosphere, and the effect of distance and inclination to the light source. Most of these sections were written in 2007 and are overdue for an update, but they should still be OK as an introduction for anyone who needs one.
Finally, I gave this presentation in July for the Colour Society of Australia explaining my own take on the issues raised in Muzz's thread:
Traditional and Modern Colour Theory
[ame="
https://www.youtube.com/watch?v=4enFjTGVTnc"]https://www.youtube.com/watch?v=4enFjTGVTnc[/ame]
If you haven't checked out my site before please don't miss the section near the end on principles of colour and light for painters, which touches on colour relationships in light and shade, effects of coloured illumination, multiple light sources, and atmosphere, and the effect of distance and inclination to the light source. Most of these sections were written in 2007 and are overdue for an update, but they should still be OK as an introduction for anyone who needs one.
Finally, I gave this presentation in July for the Colour Society of Australia explaining my own take on the issues raised in Muzz's thread:
Traditional and Modern Colour Theory
[ame="
https://www.youtube.com/watch?v=4enFjTGVTnc"]https://www.youtube.com/watch?v=4enFjTGVTnc[/ame]
Replies
Hey briggsy! Glad you decided to drop by here!
I just went through the webinar, which was really good, so thanks for putting that online
So i only just recently wrapped my head around some of the final pieces of the puzzle of modern colour theory as far as i can tell, with the main confusing thing being opponency in the brain. HueValueChroma.com is probably the only website i can recommend for complex theory for colour to people, but like i was saying in the other thread about how dense it is.
I don't consider myself a person that has a hard time learning things, but honestly the way i had to work through the documentation was constant cross referencing with self experiments, painting, and reading tech documents on how 3d renderers work. Also, although most illustrators in my circle know of it as a resource, most tend to treat the site with a lot of respect but keep a barge poles worth of distance. Only one other artist i know has made his way through it, and it does show in his very strong colour work.
I don't think that simplifying hvc.com is the answer, but what are your thoughts on producing a much easier to understand subset of information, which strips out the history and a lot of the superfluous information?
I understand your argument that traditional colour theory has persisted because of that's just the information that got popular and entered the public consciousness, but i'm not so sure that's it. I feel that a lot of it has to do with the fact that it is so much simpler to teach to students and children, and modern colour theory doesn't seem to have a textbook that even attempted to simplify it down to a level that beginners can approach. The standard approach right now seems to be to learn traditional colour theory until you reach the end of what it can tell you and you consequently need to unlearn a lot of what you knew before.
I don't think it helps either that all the websites that present this information look dated. (handprint, huevaluechroma and midimagic), and compared to the websites that cover traditional colour theory, they do not look like the most up to date and professional places to learn colour theory from. (I would actually love to help organize a more modern webdesign)
I guess i just want to hear your thoughts on all of this, and your thoughts on how we can work towards solutions... because something is seriously broken about colour teaching as it is now.
I've bookmarked your website in the past but what I've really followed is your Color thread you made on Conceptart.org
http://www.conceptart.org/forums/showthread.php/112049-The-Dimensions-of-Colour-a-colour-theory-discussion-thread/page25
I highly recommend others to see the thread if you haven't already. The color studies posted are amazingly done and photorealistic.
I'd certainly write an introductory textbook or tutorial for beginners in a quite different way, and I hope to do just that before too long, but a month ago I was asked to write a chapter on colour spaces for the forthcoming Routledge Handbook of Colour for Philosophers, so I'm afraid that's me for the rest of the year!
As I said in the Youtube video, as far as improving the dismal state of colour education goes, our best hope is student demand, especially in the very concrete form of student feedback surveys penalizing, or threatening to penalize, lecturers who persist in obsolete teaching.
I wish I'd been able to get educated on this years ago.
As I said i don't think modifying the content on HVC is what I'm asking for, it's an awesome resource and the level of depth is necessary for the website touting to be colour theory made hard.
I was just trying to get your take on how we should be better writing the intro to colour courses.
That is only a solution within teaching institutions though? The majority of illustrators are supplementing their education with online resources, and i don't know where to point people to as a solid resource from them to start with, if they aren't ready for HVC, or they are easily discouraged from heavily technical documents.
Do you have any favorite intro to colour online resources that you feel are doing it correctly? Though I am very keen to see you write an introduction course!!!
Muzz, as an overall introduction for beginners I always recommend James Gurney's Color and Light, but specifically for digital colour I don't know. I've seen some courses advertised commercially that I'd expect should be good, but without seeing the content I can't be sure, because I've seen a couple of Gnomon DVDs on colour that I thought were mediocre to terrible. All I can suggest for now is that you tell anyone who has trouble with HVC to work through it slowly and step-by-step, like they would a site on perspective or anatomy, or find an actual teacher who can take them through it. I hear this course is pretty good, for example.
Maybe a dedicated tutorial on how to use blending modes to do some neat colour maths or something.(Next to an article about semiotics and your drawings)
I was wondering, would you care to comment briefly on the following two tutorials which are reasonably widespread in the community, and if you feel they hold any major misconceptions?
http://androidarts.com/art_tut.htm and http://www.itchy-animation.co.uk/light.htm
I find your article to be incredibly more in depth, but frankly very intense and complex, whereas I feel these two are a bit more easily digestible. For example, what is your opinion on generally increased saturation in the terminator/core shadow?
Thank you!
Otherwise afaik there is nothing in the maths that would cause it.
This tutorial makes many good points and wise comments, and I'd have no problem recommending it provided people make use of the following footnotes:
Terminology
"Saturation - How much color there is. Grays aren't saturated. Neon colors are very saturated. In the 8 bit RGB color model (0-255 per slider) you can measure saturation by looking at the difference between the sliders (min-max). Dark and bright colors look less saturated because it's not possible to bring the sliders far apart. For example, 255, 0, 0 is very red. 255, 220, 220, while full red, looks less saturated because the delta is just 255-220 = 35 and not 255 like in the former case."
This section seems to be partly talking about chroma instead of saturation, e.g. "dark ... colours look less saturated" and in a later section "shadows are often less saturated". I would replace it with this:
Saturation and chroma - Saturation in HSV (or HSB) is a measure of the purity or non-whiteness of the colour of the light making up a digital colour. For example, R255 G0 B0 and R100 G0 B0 are both fully saturated (S=100) because they are made of pure red light with no desaturating green or blue. Chroma is the visual strength of a colour, or how different it is from a grey of the same value. High chroma depends on a combination of saturation and relative brightness, so although R255 G0 B0 and R100 G0 B0 are both fully saturated, the former has much higher chroma. See figure 9.9 here:
http://www.huevaluechroma.com/093.php
Other things being equal, we paint shadows on an object with image colours of lower chroma but the same saturation as the lights, i.e. shading series tend to follow lines of uniform saturation.
For more information see:
http://www.huevaluechroma.com/012.php#chroma
http://www.huevaluechroma.com/012.php#saturation
http://www.huevaluechroma.com/101.php#shadingseries
"HSV - Three dimensions of light, Hue, Saturation and Value".
If this refers to HSV colour space, please note that "value" here is not value in its usual sense, but refers to brightness relative to the maximum possible for a given hue and saturation. See Fig. 1.2.4 here:
http://www.huevaluechroma.com/012.php#brightness
Light stuff
"A red apple reflects mostly red wavelengths, the rest are absorbed and turned into heat or something. That's why black stuff get so hot in the sun. Anyways, the reflected light bounce away differently depending on the surface. If the surface is bumpy it will bounce away sort of randomly, like a tennis ball that hits rocky terrain. If the surface is smooth it will bounce away in a predictable path."
This is an often repeated mistake. Like most things, an apple reflects light both randomly and predictably at the same time, through two distinct processes. The randomly directed or diffuse reflection, which is the part that is affected by selective absorption and so looks red, has nothing to do with surface roughness; it is considered to result from subsurface scattering - that is, light enters a surface layer, is scattered multiple times, and then exits in a random direction. The reflection that occurs at a predictable angle is called the specular or interface reflection. Unlike the diffuse reflection it usually is NOT coloured by the object but retains the colour of the light source, and is considered to result from light that bounces at the surface. Surface roughness can make this interface reflection fuzzy, but it is still a different thing to the main diffuse reflection (i.e. it is still light-source coloured). See this page for a fuller explanation:
http://www.huevaluechroma.com/021.php
Overall a good basic tutorial, although Richard has some tough competition in this material in James Gurney. Of course I have a few quibbles:
Part 1: The Basics
"Finally the light that is reflected between the card and the ball is also predominantly blue (even though the card and ball are white) since it is blue skylight that is being reflected by the white objects."
In this passage and in the diagram Richard seems to be completely neglecting the light reflected from the sunlit parts of the card. The diagram shows a conventional pattern of light and shade with a continuous "core shadow" that would not be seen on a sphere on a reflecting tabletop. Compare Fig. 2.1 here:
http://www.huevaluechroma.com/021.php
But the main problem is that although he shows a fuzzy specular reflection in most of his diagrams, there is no mention anywhere here of the distinction between specular (or interface) and diffuse (or body) reflection, which is absolutely fundamental to visual appearance. So I'd suggest people supplement this with a look at this page (preferably plus James Gurney for more on different lighting environments):
http://www.huevaluechroma.com/021.php
Hard time struggle with it.
Whoa, that makes so much sense. Of course light needs to enter material in order to be filtered...
Another thing that should be common knowledge but isn't.
Conversation went as follows:
Me - Quick question on shadows for all my artist-friends out there..
Since taking the painting-class with Dice Tsutsumi and Robert Kondo earlier this year, in my paintings, I've been blocking in shadows using a hue/saturation adjustment layer with brightness set to about -60/-70, and an increased saturation of about 20-30 (which is what they do in general). Thing is, I think this looks good, and gives me a good starting-point for building the colors in the shadow, gradually taking things like ambient light, bounce-light, sub-surface-scattering and occlusion into account (each being a new adjustment-layer), and I feel it echoes what I've observed in still-life studies as well.
What I don't understand is why that which is in shadow is generally perceived as more saturated. Can a general rule for this be formed, or should the color of shadow always be considered in relationship to all other colors and light in any given picture? I do feel the need for a general starting-point though (when painting from imagination).
A friend of mine suggested that we perceive shadows as more saturated, because the light-source tints the planes which it hits, thereby increasing it's value and successively reduces it's chroma (unless it's yellow). By contrast that which is in shadow comes off as more saturated. This is the best explanation I've gotten so far, and it's my current hypothesis.
I'm not quite content with "it looks good", I'm trying to gain a deeper understanding. Does anyone know how this works, and why? Are shadows in general more saturated, less saturated, or have the same saturation as that which is in light? How do you guys block in shadows (from imagination)? Cheers!
Some answers I got which I found useful: I'm not saying all of these are accurate or true, but provides a selection of the plethora of answers.
Allison P - I think it depends largely on the situation. In a neutral environment (matte surfaces, single white light source), I was always under the impression the shadows would be perceived as less saturated, but there are most certainly situations where they can appear more saturated. As far as the "it looks better" explanation goes, I think more saturated shadows give a deeper, richer visual experience though in many cases might not actually be "accurate." Not sure if that answers your question but that's my thoughts on the matter - at the end of the day, the colours we use and the things we choose to empathize or deemphasize are I the interest of designing a better perception of a scene. If more saturated shadows serve that design goal, then that method is undertaken and called a style. Artists don't replicate reality but rather get inspired by it and make it better. The issue of saturated shadow colours I think is an example of the artist taking liberties to communicate something that perhaps a real lighting scenario could not.
Simon M - Saturation occurs in shadows when there is colour bleed from bounced light. It's often pushed in animation because it looks appealing.
Tristan E - This is part of the problem of color terminology. "Saturation" and "chroma" are often used interchangeably, but they're actually slightly different concepts. Shadows can be more saturated and less chromatic at the same time.
Yes, so under uniform, colorless lighting conditions saturation in shadows would stay the same, but chroma would drop. However, if there's any ambient or bounced light (which there always is), it's going to effect things. For instance, when a color reflects back into itself in the shadows, as we often see in the figure or drapery, it's going to increase saturation.
Devin P - I think it's important to have the knowledge of why things work [science behind color and light] but only enough so that you can break the things that happen in real life situations intentionally to achieve certain effects. At a point we have to decide how to balance the Science vs. Psychology of light/color.
Daniel P B - As far as occlusion shadows are perceived warmer I think I can give an explanation.
Occlusion shadows occur where light cannot easily escape from, and photons keep bouncing in these places losing energy and very few escaping to our eye.
The energy of light is determined by its wave length and blue colours have higher frequency that reddish ones. Below that, we dont perceive light as we reach the infra red zone.
So, my guess, no matter what colored light will reach occlusion zones it will come out weaker than it entered, thus with a longer wave lenght, thus tending to red or warmer colours.
Albert U - Usually I am desaturating colors in the shadows, and if the main lightsource is warm, shadows should be cold and if the main lightsource is cold, shadows are warm. But If there is a strong bounce light in the shadows, I do saturate it a bit more than regular shadows.And I am not using any different layers, just painting. But I have to learn a lot so no idea if it helps.
Sam N - The answer, as I understand it anyway, is way more complicated and less exciting than you'd think. And I don't want to write a three-page essay, so I'll leave it as this: there isn't a rule that says shadows will be more or less saturated than key lit areas. Saturation of the shadows is a function of the light color and the surface colors, and there are no absolutes. So skin tones under a cool key will appear less saturated than shadows that are lit neutrally, while skin tones under a warm key with cool fill will appear more saturated in the light and less saturated in the shadows. I know, I'm boring myself and not helping anyone.
Me -Would you say there's merit to the claim that a neutral white light (like the sun for instance) generally will tint the planes being hit, thus reducing their chroma? And supposing that, that by sheer contrast, we tend to percieve shadowy areas slightly more saturated? Is that too much of a generalization and simplification to be useful?
I recall reading in Gurney's book, Color and Light, that we perceive areas of dark contact-shadows/occlusion as warmer, but he never gave an explanation for this. As far as I've understood it seems to be a consequence of partly sub-surface-scattering, and of our eyes being more sensitive to warm-wavelengths of light, exaggerating the difference in a scene which might be lit be a cool ambient light-source?
Sam N - Contact shadows, or "ambient occlusion" are usually slightly warmer than the surrounding light, yes. I've heard a number of theories on this but nothing that convinces me, yet. But the fact that it can be observed means you should probably paint it that way even if you're not sure why.
Me - I feel as if any sort of notion I had of a general rule regarding more or less saturated shadows, in terms of physics and science, has been crushed.
Sorry for the wall of text
Ps. To be clear, I feel you've already answered some of these statements in previous posts, I am including some of these merely for context.
So in the simplest situation uniform saturation series definitely work, but as Tristan says as soon as you introduce reflected or other light sources this basic relationship will be modified.
The figure is from this page, which explains the theory of why uniform saturation series work as shading series:
http://www.huevaluechroma.com/101.php
About the confusion between Saturation and Chroma : this misunderstanding is indeed very widespread amongst digital artists, and I think that in order for people to be better educated on this topic it is necessary to tackle it on two fronts.
1 Demonstrating the concept using means that the audience can easily relate to.
For instance, on top of explaining it with its accurate definition like you did above, I think you could hammer the point home by simply dropping these colors in Photoshop and actually dragging up the Saturation slider to show that, indeed, all these swatches are already at full saturation. Like here :
In general I feel like a lot of the excellent diagrams on HueValueChroma could benefit from being illustrated in this manner (and this is true of the video webinar as well). No need for complex motion graphics - mere screen captures from the Photoshop UI with GIFcam can go a long way.
2 Acknowledging where the confusion comes from in the first place.
"Saturation" being inappropriately used to describe Chroma probably comes from the fact that many digital artists believe that a dark red (effectively already at 100% saturation, in the proper sense of the word) can be "further saturated" by simply "putting more red into it" as the RGB sliders allow (that certainly was my case for the longest time).
What is sad is that something that probably originated as a mere colloquial slip then goes on to have practical consequences (ie artists not understanding what saturation actually means). In an era where improper use of language can quickly become common ("litterally" used for "figuratively", "could care less" used for "couldn't care less", and the list goes on), I think it is extremely important for educators to identify the cases when misconceptions come from the improper use of terms, because these are very hard to get rid of once they catch on. The pessimist in me tends to believe that for saturation vs chroma it is already too late ...
Lastly, there might be another linguistic issue at play here : the fact that there is no verb in the english language for "increasing the chroma of a color". Therefore most of us ended up inappropriately using "to saturate" as a lazy substitute for it.
I hope this makes sense !
It gets even more confusing when you realise a lot of artists are used to cylindrical systems but HSV/HSI/HCY' and LAB are closer to double-conical. So you get questions as to why, if doing a curve adjustment in LAB and you turn down the lightness so that everything should be black, the chromatic colors don't become black... For a double conical system, black has no chroma, in the examples just discussed, black does have a chroma, it just looks the same.
It is silly to bake down things like shadow colour to simple rules as it analyzes the problem in isolation of what is really happening in the maths.
If you were to bake it to a rule, similar ambients to the local colour will boost shadow saturation while non similar will drop it, but this crude rule would totally ignore the natural hue shifts caused by the maths in the channels, and the correct change in value.
I know it seems impossible but if you sit down and learn how channels of light interact, it becomes quite elegant and easy to use.
Of course this all makes sense once one understands the notion of chroma and saturation in their modern sense, but I can see how one could confuse one for another since when starting with a slightly impure shade of red, cranking up the saturation slider not only boosts the saturation up to 100 but *also* increases chroma.
Therefore I can see how artists could call this effect "saturating a color", since pure reds (xxx 000 000) happen quite rarely and therefore, cranking up the photoshop saturation sliders almost always affects chroma as well.
Anyways, just rambling, sorry for the derail
Wolthera : Agreed - and one top of that one could also suspect that many digital artists are not familiar with Itten either, and therefore are just inferring these terms from what they read (but not necessarily understand) in the Photoshop UI !
I can't wait for HDR screens, which will probably end up using scrgb, for backwards compatibility reasons. https://en.wikipedia.org/wiki/ScRGB
The confusion between saturation and chroma goes way back, in fact it was the art teacher Arthur Pope who pointed out in the 1920's that the word "saturation" was being used by different scientists of the time for these two distinct concepts. So while Itten's use of the same term for both is one source of the confusion, it isn't the only one. Unfortunately Photoshop itself is another, in that the word "saturation" is used in different places for four different dimensions. I've just spelled this out a bit more clearly in my section introducing saturation:
http://www.huevaluechroma.com/012.php#saturation
In the "Hue/Saturation" adjustment panel illustrated by pior, "Saturation" refers to S in the HLS model, which is different to S in HSB. Increasing HLS "saturation" moves a colour outwards in the colour space along a line of uniform "L". This has the effect of increasing HSB brightness as well as HSB saturation (!).
http://www.huevaluechroma.com/012.php#other
Wolthera, that's very curious about the result of reducing L in Lab to zero. My guess is that it reflects the relatively simple way Photoshop converts to Lab coordinates compared to the actual CIEL*a*b* formula. The colours seem to continue following the white lines in my illustration C on the previous page until they hit the zero L plane.
Muzz, you might want to rethink the location of those highlights on your spheres - see the discussion of Fig. 2.3 on this page:
http://www.huevaluechroma.com/021.php
http://www.dorian-iten.com/light-and-form/
That's the stuff I like to hear.
I have actually talked about it with several color management specialists, and it's not due to the way the space is calculated but rather that curve-adjustment type-filters just don't know how to treat lab right due it's biconical nature.