Sunday, June 4, 2017

Towards Successful Painting






Since this is by far the most successful painting I've ever done, I would like to record my experience about it and share with those who are interested in learning the process.

I normally divide my entire process into three stages: 1) the initial block-in, 2) features and details, and 3) final polishing. I suppose you must have heard about them hundreds of times from different instructors/books. Well, here we go again - I'm going to explain how I interpret these stages.


1) the initial stage is usually done rather quickly, within 30 mins to an hour, which means in 1 or 2 sittings if you have a live model. The result from this stage looks very simple, but contains the MOST important decisions/designs made for the entire painting. At this stage, do paint BIG colored shapes. Make CLEAR distinction between any two adjacent shapes. If you are unable to make such separation then they should be painted as one. Each shape is preferred to have straight boundary edges rather than curved ones. Then assign a color to a shape. You only need several colors in total to fill in all blocks. These colors can be ordered by their temperatures. It is important to retain the color order between any pair of blocks on the canvas, as such relationship is observed from the model.

For example, in my first stage painting, I only had 7 colors, ordered by temperature from cool to warm, they are: bright shirt, background,  green shirt, skin tone, hair, fur shawl, brown jacket.  Some people may disagree on my color order of green shirt and skin tone, for example. It is ok. I think it is a subjective thing. Just be consistent with your own judgement throughout the entire course of painting once your decision is made.

Keyword: "big colored shapes with clear distinction."
Q&A
To put down the blocks as accurately as possible, you need to compare each with other shapes (especially neighboring ones). One way to achieve this is to compare the angle of the block edges. If you have curved edges it is difficult to compare the angle and therefore more likely to make mistake. In another sense, a curved edge can be considered as several connected smaller segments. You may choose to compare two long straight lines for the relative angle, or to compare among a number of smaller pieces if you use curve. It seems a no brainer to do the former. Additionally, long straight lines helps while verifying what you have put down too. It’s easier to spot something wrong by straight lines than curves.
I usually do an initial assessment of all major colors in the scene first. I then know the range of the color I need, whether I need to add more paint to my default (limited) palette (see “Limited palette”). Then I roughly sort these identified colors by their color temperatures.  Then I evaluate whether I’m able to keep the two most similar colors distinguishable throughout the entire painting. For example, the two colors on the skin and the bright shirt in the initial stage paint are fairly close. If I’m not confident that I’m skilled enough to maintain the difference between these two colors till finish, I would make them more different by increasing their  temperature/value difference. That means I might have to change the rest colors as well in order to avoid  new situations like this from happening. As a consequence, a brown jacket will be painted orange, or a pair of green pants will be painted blue. It is ok. I think it’s better to be off from observation than not being able to describe what you want to represent. Additionally, even if the above situation didn’t happen, you may still stretch the chroma to full extent for your artistic taste (see “tonal painter vs color painter”).  Once I decide the color assignment, I go ahead start painting from the darker ones to the lighter ones. (see “Why paint the darks first?”) Final words, I’ve seen brilliant painters literally started their paintings from any point on their canvas, and found their way to spread out the whole canvas till finish. Total master piece.  I guess until that day, let us all stick with the student version of the starting.
You must have heard/read that to paint the darks first. This is the way I do it. Many painters too. But I only agree with this rule to some extent. I’m not sure if painters that have painted this way truly understand why. To me, the goal is not to paint the darks, but to clearly separate darks and lights, yin and yang, during the design stage. The resulted positive and negative space helps double check the design, and the accuracy of placement of shapes. Therefore, if I work on a black canvas, I see no reason not to start with the light first, so I can do the separation quickly. Having said that, there is another reason one would start by painting darks first, due to material reasons. If you paint light first, it is difficult to make it dark again by painting it over it. This is because the pigments in white paint will always reflect lots of light even being mixed with black paint. The white paint also reduces other paints’ saturation, for the same reason. But if you paint dark first, then you can paint light over it to make it lighter with no problem. For this reason, in particular for oil painting,, it is almost always desirable to start with the darks.
This is particularly important but difficult to do for colors similar in temperature. For red vs blue it’s easy, but how about a light yellow vs olive green? Warm gray vs pink? Well, for those I myself have problem to tell. In this case, I try to compare two colors similar in value, rather than two whose values are very different. If I can’t find other similar values, I will have to make a guess. Just put down what I think the best hoping it will turn out good. Another difficulty is to compare two colors both with low value (like in the dark). My eyes can’t see clearly the chroma. Perhaps more experienced painters can. In this case, I would have to make a guess again. But this time I can relate to the color of the light source. If the light source is warm, then the shadow is more like to be cool. Then I will make a bet by blue paint + the observed true surface color of object (e.g yellow for lemon).
The simple answer is to add any warmer paint to a paint to make it warmer, and any cooler paint to make it cooler. Sorry, I haven't been very helpful. Let me try to answer that again. For me, as a lazy person, I just add yellow/red paint to make another paint warmer. I add white/blue paint to make another paint cooler. (Note here there are more than one choices to change the temperature. Which one to pick depends on the natural color observed from the model. Chances are you need to pick both with certain proportion) This is because I normally only have white, yellow, red and blue paints on my palette (see "limited palette"). Be mindful you still need to adjust value after mixing, which is likely to change the temperature again. Once you're more experienced, you will get the color you want with minimal mixing.
Here the color means both value and temperature. Both need to be retained. For example, if shape A is darker than shape B based on observation from the model, shape A should be darker than shape B on the canvas as well. The same applies to color temperature. Luckily, you have some freedom to tune the amount of that difference. For example, you can exaggerate the temperature difference between A and B, or suppress it, but NOT switch the order. When the temperature relationship between A and B is switched, that area may look “dirty” or unnatural. It is even more damaging to switch value relationship, as this will break the perceptional local 3D forms and create distorted surfaces. Our eyes are very sensitive to the value relationship, not only by the order, but also the amount of difference. I would recommend stick with the true value relationship as being observed from the model for painting human face, unless you really know what you are doing. There are two exceptions. First of all, when you want to change the surface color of an object, such as paint a red dress as if it is blue. In this case it is ok to alter the color relationship of the red dress with respect to other blocks. However, within the dress itself, the internal relationship between light and shadows should still remain. Secondly, it is ok to break this rule somewhat within small local regions, so that when you look at the painting from far away, you can only see the average value/color at that spot. It will look like this painting has lots of noise. Many impressionist painters use this approach. (Also see “How do I know if the painting looks good?”)
Whether a painting looks good is a very personal thing. But I think a painting could look wrong. There are a million ways to look wrong for a painting and therefore beyond summarizing. Having said that, I may suggest a way to check the painting. Whether or not your painting looks good cannot be told from standing 1 feet away, but from a distance. As the rule of thumb, for normal sized painting you should step back at least 2 meters (6 feet).
2) The second stage is when I add features and details. It can be so fun to do and I sometimes get carried away. Don't. In this stage, try to avoid "inventing new colors". Rather, use variation of the color of a shape when the feature you are working on is inside that shape.

For example, in the face area, I varied the colors from a solid generic skin tone to pinker, browner, or whiter versions. (Well, I think I really shouldn't list specific color names. In correct "terms", I varied the color to cooler or warmer, brighter or darker.)

You may find your drawing skill and understanding of objects structure starts to kick in at this stage. Some times if you feel "something doesn't look right", that's probably because of the drawing. Additionally, wrong value will make the 3 dimensional form distorted, also creating this "weird" looking feeling. So constantly compare the value of a stroke you put down to the existing ones.

If you find the average value/color of your initial shape starts to drift away from its original location after you add all features, the chances are you could be in trouble soon, unless it's intended. For instance, in my initial painting the bright shirt is only slightly cooler than the skin and similar in value. In later stage, I realized these two similar colors lead to confusion that the shirt is also perceived as bare skin. To fix that I made the bright shirt cooler and lighter, as shown in the 3rd painting.

Keyword: "vary initial shape colors with constant evaluation of value and temperature."
Sub-session: How to paint the features(under canonical light setting, see “definition”)
To paint a believable human face, the facial features need to be first, obey perspective, second obey anatomy. Unless you have master drawing faces, it is a good practice to lay in the landmarks, as forms of shadow blocks on the face: 1) eye: a shadow block with upper boundary aligned to upper eye socket. Lower boundary goes down to at least the lower eye lid.  Be mindful of the keystone shape between two eyes. 2) nose: a shadow block with upper boundary aligned to the tip of the nose. Lower boundary varies. 3) mouth: a shadow block with upper boundary aligned to crest of the lower lip. Carefully determine the location of block, as it defines the length of the head. Upper lip can be ignored for now. Check points: 1) double check if (1)-(3)  sit right with respect to the center line of the face (you may draw a imaginary line). 2) double check if the boundary at the top of the eye socket, bottom of the nose, and then chin are parallel (or in correct perspective). Sometimes for female models with pointy chins, it’s easy to get it wrong. In that case, give her a squared chin for now so you can check the perspective.

3) I call my third stage the "polish" stage. The goal here is to correct previous mistakes, re-enforce some marks, vary edge quality, and add some spice. I would do highlight very moderately, usually only around 4~8 spots. I find too many highlights are very disturbing. I always leave it out when in doubt. Based on my experience this stage is where the "overly-painted" happens. As a consequence the result from this stage is less visually pleasing than that from the end of last stage. I've heard it could happen to EVERY painter.

Keyword: "When in doubt, don't paint it."
Q&A Definition:
Value, in this context, means luminance or light intensity reflected from the paint. Color means the chroma, in this context. In broader terms, color should include both value and chroma.
This is a very subjective feeling of a color, from millions of years of human evolution. Warm colors usually are associated with fire/sun, which, as a source of energy, appear yellow, orange to red. On the other hand, cool colors are associated with winter scenes like snow/ice, therefore appear white to blue. Based on modern science the warmness/coolness can be represented by a color temperature by the blackbody radiation. In contras arty to the psychological acceptance, a warm light is emitted by a blackbody radiator with low temperature, while a cool light is emitted with high temperature.(Check Wikipedia). In terms of color, the light from warm to cool can be named from red, orange, yellow, gray, blue, violet. Note that the color green is not included in this set. For those absent colors they are more neutral and can be manipulated to warmer/cooler versions easily. Also note that even those gray feels like a neutral color, it is not. More often it gives a cool feeling. On the other hand, the black usually gives warm feeling. Sunlight at noon at 6000 Kelvin degree is generally considered the “whitest” light and is cool.
Another important concept while understanding this concept is that everything is relative. Use the noon sunlight as example, if you have a studio light that has the same 6000K temperature you get a cool light. But when go outdoor, the blue northern skylight is much cooler so that the sunlight is actually appear warm.
A tonal painter focuses on representing the scene by dark/light relationships. Color is secondary to them. A color painter focuses on recording the light reflected from the surfaces, and tries to describe the light by its color. However, when the surface reflects very little light (such as dark objects or in shadow), it loses its chroma and appear gray or black. To cope with that, the color painters compress the dynamic range of the paintings a lot, by raising the darkest dark to midrange value, such that the color paint can retain its best vibrant. Because of this reduced dark/light contrast, they have to exaggerate the chroma somehow to make the paintings interesting again. Similarly, bright paints also loses chromaticity and are avoided, even though the loss isn’t as severe. However, keep in mind that the color painters still have to make the value relationship correct in order for the chroma to work for their painting. A correct value relationship between two surfaces means that if A has less value than B from the scene, then it should be so on the canvas. It is perfectly ok to change how much A’s value is less than B but not to switch this relationship. Usually a tonal painter uses full dynamic range of the value. A color painter uses half (Wikipedia Monet). Great painters can paint both. It’s their preference to concentrate on value contrast or color contrast.
In studio scene, where the lamp light source is 5500-6000K. There can be a secondary light source but it does not compete with the dominant light source.




Monday, October 17, 2016

The Joy of Drawing





Wow! Can't believe it has been 3 years since the last time I took a figure drawing class (with Kevin Chen)! This time I'm learning figure drawing from the legendary Mark Westermoe~

Interestingly, after a 3 year break, the way I draw now is different from before. Actually, the way I see the model becomes different too. With regular training in painting in the last 2 years, I've got a new habit to draw/paint with shadow as my priority, rather than lines that describe human anatomy previously. After all, there's seldom lines in paintings. Often, I even block-in mass first before any line laid in. Particularly in short pose sessions when there's time constraint, I find myself unconsciously use this approach a lot.  Above are some drawings I did during the first 3 weeks in the class. Just for the record 10 mins, 5 mins, 1 hour.

Sunday, September 28, 2014

Sketchbook 2013-2014

Just finished another sketchbook from model sessions at artcenter. Before i-will-never-look-at-again as previous finished sketchbooks, this time I decided to take photos of some pieces and compare them side by side in timing order to see the progress, if any....



Thursday, June 19, 2014

Sketching for Environment SU14 Week 03



Pasadena City Hall, Spanish Colonial Revival Style architecture,1924.

Sketching for Environment SU14 Week 02


William G. Kerckhoff Laboratories of the Biological Sciences, Beaux Arts Mission Style, Caltech, 1917

Wednesday, April 24, 2013

Watercolors

Water color was so much fun. It was frustrating at the beginning but when you got a grasp of it, it can be so creative.

Some homework,

 


This was the final project. Unfortunately, it's somehow unfinished and I was not in mood to complete it any longer...