Рет қаралды 617
After listening to sometime machinist and now student of painting, Kevin, Paul discusses with him aspects of and times in the painting process when painters can be less than precise in pursuing the look of nature.
In Response to
Kevin S
QUESTION: I've been chewing on a thought and finally feel like I have the words to describe it:
There are a couple of points of context which I want to make. The first is that I was thinking about how when we work, we discretize values into about 5 or so levels. Though in nature, at least in theory, it would seem that the range of values is more of a gradient with infinitesimally small steps. The reason we discretize the full value range into only a few levels is most likely because our eyes can only perceive a certain amount of difference between values. That is to say that we might break up what we see into 5 values instead of breaking it up into 50 values because there is a perceptible difference between the 4th and 5th value in the former case, whereas there is no perceptible difference between the 49th and 50th values in the latter case. (see attached image for depiction of a value range with varying step size)
This got me thinking that there are certain allowances in the painting process and it took me to the idea of machining tolerances. There is this idea in machining that one cannot machine an exact dimension with a high degree of repeatability, and so there needs to be a tolerance factored into the process which fits the function of the part. A typical tolerance for a machined part is 0.005 inches, which means that if a customer requests a machined pin at 1in diameter, and I machine it to 1.003 inches in diameter, I can still give them the part and it will serve the function. However, let's say they need the part to press fit into a hole, in which case they might give me a tolerance of 0.001 inches, so that if I machine it to 1.003 inches, it will not work for the function and I will have to scrap the part. Larger tolerances are easier to hit and require less setup, so they cost less. In other words, tolerances are designed to optimize fitting parts together and to reduce cost.
One might think of tolerances as 'wiggle rooms'.
I wonder if the painting process is analogous such that there are tolerances designed around fitting the parts together to serve the function and saving time/ saving from noodling.
For instance, it seems to me that we allow some tolerance with respect to drawing shape in our first pass, since we are principally concerned with light effect, which is to say that we try to hit the effect and get around the canvas without concerning ourselves that it isn't exactly the right shape. Otherwise we would spend too much time in one area getting the shape exactly right, even though we might have to move it later. But we will get the right shape eventually, in later passes.
I also wonder if what distinguishes master painters is what tolerances they allow in their process. Monet might be said to allow larger tolerances in drawing than Decamp, for instance. Sargent might be said to allow tolerances in color for the sake of tone or for efficiency, or something.
Ultimately, though, whatever tolerances a painter allows, all of the parts must fit together in the end. Each element, according to my analogy, might have a tolerance associated with it (drawing, value, color, etc.), but all of the tolerances must mesh in such a way that they do not destroy the truthfulness of the impression.
And so perhaps the large question here is, when and where do we allow tolerances in the painting process, and how much tolerance can we allow without affecting the truthfulness of the final, general impression?
Kevin S