Imogen's Sketchbook of '22

[JosephCow]

The more I practice the worse I get. 

Aw I'm sure that's not true! Although I find it believable that art school, and years of looking at various tutorials haven't yielded much. It's probably a matter of just not approaching it in the right way where you understand the basic principles, and with them do whatever it is that you desire. You mentioned looking up yknow like realistic cloth tutorials, and I think that's kind of the issue. A lot of tutorials do not focus on principles, but on specific tips and tricks to achieve various effects. Those aren't what you need, really. They do practically nothing for you in the long run.

So for the shirt, you are going to achieve it! Believe that you can. But I disagree with darktiste that realism is easy style to achieve. If it was, there would be no need for art schools, and books, and special training for realism, since anyone could easily just sit down and do it. This photo of a shirt is also very tricky, so don't feel bad if it isn't looking how you wanted. I also really think you might want to ignore stuff about using layers, and what brush to use etc. etc. because all of that is honestly just kind of a distraction.

When I'm looking to paint the shirt realistically, I'm going to cycle between 3 different skills or principles.

1. Drawing
  -Look at angles, proportions, shapes. It's kinda self explanatory that good drawing is needed for realism.

2. Relationship of colors/values
  - Finding abstract relationships unique to the picture. Like how bright is one plane of the shirt compared to the one next to it, compared to the background? Is the jump between one piece and the piece next to it greater, smaller, or is it perfect? Should the color be warmer or cooler? So this is thinking completely abstractly, not about whether it's a shirt or anything, just the impression of it two dimensionally.

3. Modeling form.
  - Completely opposite of the abstract value relationships, this is using the values that you got from that to describe the 3D effect of form. This is the perception that as the form bends towards or away from the light it gets brighter or darker, and gradating it accordingly so that it appears to have the same form as in the reference.

So if I look at the shirt and think about number 2. The shirt is much more similar in value all over. It doesn't get so dark in the shadows and creases. This is what makes it so hard and delicate to paint. It's not really so bright either. So I break it down into large masses and compare one to the one next to it. From this reasoning I get a general impression of the shirt, and some specific relationships that I think are right.

Then I can think about number 3, and use those values that I gathered from the ref to mold the shirt into specific forms. For example I notice the right collar, the shape and form is quite different. It comes to more of a point at the corner crease. And it rolls over in a curve into the chest, it isn't a sharp edge. So I can show that by make a subtle gradation at that edge. And every single inch of this fabric could be analyzed similarly, but I just start with the major things.

Number 1 gives me the shapes. Number 2 gives me the value relationships. Number 3 informs me specifically on where to place the values so it looks 3D.

For an area like the bottom left, that is probably the most complicated part of this. Very subtle changes in value which describe numerous complex forms. I recommend isolating that. It's much easier to see what needs to happen when you abstract it, isn't it? When I cut it out so it kind of looks like an interesting white leaf, suddenly it becomes a lot easier to slow down and paint that. In general you need to slow down and get a lot more specific. You're too general about it, like the left collar lapel thing looks muddy just because the shape of the shadow isn't closely observed.




The longer you spend the more specific and delicate your corrections should be until you've been as thorough as you want to. I don't usually study things like shirts super closely because I probably wouldn't put a super detailed shirt in one of my paintings. But as an exercise it's good. I like to find the strong relationships of value that make things look real in the sense that it strongly describes the light and form, but not necessarily model it completely. When you see loose paintings, which are realistic like those by Zorn or Sargents that's basically what they've done. The colors are perfect, and the drawing is perfect to the extent that it is there, but it isn't smoothly finished. Photorealism just keeps going until every detail has been depicted. Which is absolutely fine if that's what you desire.

So yeah, this definitely isn't a perfect, there's a lot more that could be done. I could sit and correct each part of what I painted for a very long time. But that's the goal, to get the structure well enough that finishing it is just a matter of time spent. Hopefully this gives you some idea of how I and other students of realism might approach the problem. There's really just those three principles or ideas that get applied. Within each there is a lot of nuance, but they are also all things you can specifically practice and get better at. It's important to note that I never mentioned anything about how to render fabric texture, or how many layers to use, or brushes. I really just don't think those types of things are important. I do know that this is kind of just adding another cook to the kitchen, and I realize I might just be confusing. But I still wanted to add my 2 cents. Others have also mentioned some good places to look for training like Dorien Iten. He's really good.