Underpainting example
 

This is a finished underpainting of my dog, "John Singer Sargent Wells" aka "Sarge." He is a mixed breed with the face of a black lab and the body of a hound (or a Rhodesian water spaniel?).

The painting is 10" x 8" and is oil on canvas. I used Genesis (non-toxic heat-set oils) and the process and result is EXACTLY as if I used traditional oils. Because, I did not have long "drying times" this project was completed in less than a day.

Getting Sarge to lie still while photographing him was the hard part...it took a lot of cheese cubes to convince him to cooperate.

Anyhow, this is generally how I do an underpainting. I really think of it as drawing with a brush. Note the very narrow range of values here. The colors I used are raw umber + titanium white.

This is the paint I used numbered from the darkest value (1) to the lightest value (5). The only reason that the #5 value jumps to become so light is because the cloth is white. Without a "white" object in the painting, I would have used a much darker value than this.

Again, the mixture is raw umber + titanium white mixed to the five values shown. Oftentimes I pre-mix my values so that I am not tempted to change them. The piles of paint are displayed below on a white paper palette.

The background value (imprimatura) here is #3 - roughly the value of a brown paperbag.

My loose drawing delineates the general light and shadow on the pooch after the imprimatura is dry.

I add thick paint for the shadow (value #2) and pile on more light (value #3).

For the sake of interesting patterns I make it a point to connect my shadows with shadows and lights with lights wherever I can. This will greatly strengthen a composition and can sometimes be rather "abstract" at this stage.

Note that the light #3 on the dog and the background #3 are the same.

The following is very thick paint - unblended. After I pile the paint on the canvas, I put the palette down and begin to work on the details of where and how the general area of light meets and blends with the general area of shadow.

By the way, when I say "thick" paint, I mean THICK! Oftentimes I pile it on with my palette knife.

When the general areas of light (#3) are blended with the general areas of shadow (#2), the halftone is created.

The halftone is where light and shadow meet. If they meet quickly (as on an edge), the halftone is narrow. If they meet slowly (as on a gentle curve), the halftone is wider.

I add more light (#4) which will be the highest light on the dog.

The highlights (#4) are added thickly within the area of general light (#3).

Note that absolutely NO highlight (#4) will ever creep into any halftone or shadow...

I blend the highlights (#2) into the general light areas of the dog. Then I add the darkest shadows (#1) to the dog.

I draw the cloth. Value #4 will be the general light on the cloth, and value #3 will be the general shadow. (Remember that value #3 was also the bacground).

Dark shadows(value #1) are added to dog and the cast shadow under his chin.

The highest value #5 is used in ball and cloth...I could have gone much further here (with only these two values) by allowing the cloth to dry and then adding an additional thick layer of paint on top of the cloth.

The stripe detail on the cloth is added only after the light is built underneath.

Note that in dog, the highest value #5 ONLY appears in the lower lid of eye on left.

The highest value #5 is used in ball and cloth.

Whew, I'm pooped. It is tough to post on a busy internet night with a 56K dial-up modem. Hope this all makes sense to somebody who struggles with underpainting.

Here is my original photo reference (using a single source of light of course).

You can see that, except for Sarge's features, I pretty much ignored the "reality" suggested in the photo.

Hi, Karin.

I'm replying to your old (May 2002) demostration of an underpainting. I must say that you made a beautiful monochromatic painting of the dog. I was following very closely the description of your procedure because I'm interested in learning your technique. However, I got confused when you said that value #4 for the background was the same value #3 for the light on the dog. Why did you duplicate value #3 and #4 instead of mixing only one value for either the light on the dog or the background?

Then you said that you added value #4 as a highlight in the general light area of the dog which is value #3, however values #3 and #4 are the same.

Then you continued saying that the highlight value #2 was spreaded thickly on the general light area of the dog, but value #2 is supposed to be a dark for the shadow.

Fnally, you said that value #4 was used for the general light on the cloth and value #3 as the general shadow, how can this be if both values are the same?

I'm sure you might have gotten involuntarily confused with the numbers, I'm trying to get it right because I want to use it in one of my paintings.

Tito's comment Whew. I tried so hard to simplify this that I seemed to make everything more complicated. Sorry about that. :)

Let's try this again. I mixed a narrow range of five values from the lightest (called #5) to the darkest (called #1). I have repeated this palette below.

What I tried to say was that the light area (color #3) on the dog and the general background (color #3) are the same at this stage, i.e., I dipped my brush in the same pile of (#3) paint.

Here is that particular stage that I was referring to.

It should be pretty clear that the light on the dog and the general background are the same color and value. There is a darkish area at the top of the canvas that I presume is a shadow cast by the "gripper" that holds my canvas on the easel.

I began by painting the canvas all one color (value #3). Covering the stark white of a canvas with a darker color is called an imprimatura. When this imprimatura was dry, I lightly sketched in the dog with a brush dipped in a darker paint, and then I began to differentiate the light and shadow on the dog.

Perhaps some of your confusion may be that you discern a slight "color or value" difference in the photo. The wet paint on the lighted areas of the dog photographed a little differently from the dry areas in the background, but they are not different.

I hope that this clarifys the process a bit for you. This was originally meant to be simple.

The "fat over lean" principle
 

"Thin" paint does not mean "lean" paint. When I did the outline drawing ("guidelines") what you see is thin paint....other than this, ALL the paint applied to this canvas was straight from the tube, thick (and lean). I paint so thickly in my bottom layers that I nearly always tend to cover and smooth out the canvas weave. For some mistaken reason or other, most people assume the opposite is true about my work.

If I had one piece of advice for a beginner, it would be (after the initial drawing) to PAINT YOUR BOTTOM LAYERS WITH VERRRRRRY THICK PAINT!

Here is an explanation of the "fat over lean" rule in indirect painting and I hope it clarifys:

In indirect painting ((Leonardo style - painting from dark to light) the artist builds up three general layers of paint and medium.

Bottom layer or layers: A color is used to block in the painting. The paint in this layer is often thinned with turpentine and a small amount of medium.

Middle layer or layers: This is where opaque colors are introduced. The medium can be thinned with 10% to 20% turpentine.

Top layer or layers: A glaze layer that modifies the opaque colors and makes the surface very rich. No thinners are added to your medium in this layer.

Paintings built in this manner follow the "fat over lean" rule. ALL PAINTING MEDIUMS ARE CONSIDERED FAT so thin your medium less as you work from the bottom layers to the top ones.

Oil paints that look shiny are FAT to begin with. And oil colors that look more matte are LEAN.

Some other painters such as Rubens who paint light to dark (and as shown in this demo) often apply thick opaque paint as the initial layer and add transparent glazes to modify that underpainting. Naturally the "fat over lean" principle applies here too.