What a great topic Thomas—thank you for putting this forward. Color “matching” can be an incredibly daunting task for a visual artist. There are so many variables to consider that it can truly seem to be a Sisyphean task. Even the term “precise color” can seem almost meaningless quite a few contexts involving visual representation. However, there are steps we can take to find a path to what we might see as effective color matching. Let’s look at some of your problems as you have laid them out:
1. Because of adjacent, surrounding colours, the colour looks different in isolation than it does on the canvas in proximity to some other colour.
Absolutely. Color is not a property of the environment but is rather a visual experience arising from the distribution of light intensity across the visible spectrum of electromagnetic energy. It is a product of a biological vision system interacting with the spectral composition of visible light that is emitted, transmitted, or reflected by the environment. As such, speaking of color in terms like “precise” do not make much sense as the biological systems responsible for the experience do not operate via accurate or precise measurement. Our perceptions have evolved for fitness—not veridicality.
For example, let’s take a look at what type of stimuli might initiate a cascade of neural activity that results in a perception of yellow. Exposure to a wavelength that is about 570-590 may initiate such a cascade. However, due to the nature of our visual system, exposure to a wavelengths at about 495–570 nm and 625–700nm (+/-)—what we might perceive as red and green light (minus blue), can do the same thing (think of your average RGB computer screen.) BUT—we can ALSO find an experience of yellow without ANY of that due to perceptual adaptation (color constancy). One of the contributing mechanisms to color constancy is what is known as chromatic adaptation. Chromatic adaptation influences the perception of a particular color within the context of its surroundings .Therefore we don’t need any of the aforementioned stimuli to generate an experience of yellow—just the right context. You can see this for yourself in the sunflower example in the above image. This latter mechanism is doing much of the heavy lifting where your palette mixtures, and applications, and reference material seem to differ significantly even after careful mixing. The contexts are all likely vastly different.
Now while this may seem to make your job even MORE Sisyphean–it really shouldn’t. While these mechanisms are not fully understood by any means—we know enough to make some pretty good (useful) predictions . In my process, it’s not a matter find perfect color matches. Rather, it’s a matter of building a context that results in a percept that I intend based in available materials. This sometimes means not achieving that “precise” color match.
For a little more on this I have this post here: From Palette to Canvas - Color in "Context"
In my experience, with my goals, the most effective starting point has almost always been the establishment of anchors that “calibrate” my palette to the observed subject. Anchors are often the darkest darks, lightest lights, or highest chroma paints that I have access to. They are often NOT mixtures, but something straight from the tube. I apply these marks first. I can then begin to add marks that appear to relate to the anchors in a manner that is reasonably similar to the relationships perceived during the observation of my subject (often with techniques like dynamic squinting). In a sense, I am solving for “x” from a number of given values/anchors (shown here as k and e). I then continue to add subsequent marks–building outwards from the anchors–to establish a growing context that I can feel reasonably confident in.
In the subject grid above the palette here, you can see two anchors that I would be likely to apply to my canvas first (e and k) in an attempt to solve for “x” (and beyond.) Can you quickly spot two more that could be applied to increase confidence in “x”?
2. The paint’s colour feels as if it changes when it is in a thick glob on the palette to when it is thinly spread on the canvas.
Again, absolutely. A dense blob of color may seem quite different in appearance from the same color spread out onto a surface (especially a bright reflective one.) Contending with this comes from experience. I would also recommend our Language of Painting Color Chart. The chart will help you to get into the ballpark for certain colors with your available gamut. Some have even subdivided the chart’s cells so that each cell contains a single-layered section and a multiple-layer opaque section. The disparity in appearance between such sections will be quite significant for translucent/transparent colors.
Related link: https://www.smartermarx.com/t/lop-basic-color-chart/228
While we are on the topic of opacity and appearance, you might find this video from Winsor&Newton informative:
3. Because I paint several layers one on top of the other, and in particular because I paint colour on top of a black and white underpainting, the colour I apply to the canvas is always going to be modified by what is beneath it anyway.
Definitely. Again, see that above video from Winsor&Newton. Continued experience will get you to a place where you can make better predictions (including predictions re: opacity/appearance.) You just have to keep practicing (deliberately).
As to your comment: “ Now as much as I want to improve my colour matching skills, there is another school of thought within me. And that is that exact colour matchings are not particularly relevant….From these optical illusions, can’t we extract a moral of the story along the lines that we needn’t care too much about precise colour details?”
Now generally speaking, I do not aim for “exact color matches.” I try to get into the ballpark based on my available materials (see anchors above) and do my best to maintain similar relationships to those observed in my subject (allowing some room for creative license.)
As to this: “ Another aspect to this issue is the following. Last year I did a painting of a teddy in a sweater. The sweater was brown but in the painting I changed it to blue because I had a nice blue that I wanted to incorporate. Now the question is, is there anything wrong with this?
Here’s the thing. How would anyone know the difference? Sweaters come in all colours. So why would anyone care if the colour had been modified from reference to finished work?”
Not at all. There is no reason that you cannot change the color of an element in a work as long as the the context is adapted to the change. For example, if I change one of the orange sphere on the right to blue—the context no longer in accordance with it. I would have to modify the surrounding context to marry the subject and its environment.
One more thing you might like to look into is the Color Matching software from Zsolt. You can use the software to make some pretty goof predictions about colors represented in a number of ways. Check it out here: RESOURCE:Zsolt Color Mixing Tools
Ok that’s enough rambling from me. Hope some of this helps Thomas!