Sunday, January 17, 2010

Mixing Color: One Approach

(This is based on a post I wrote over at ConceptArt.org)

There are many ways to think about mixing paint, and knowing more than one is useful. I might even go so far as to say it's necessary to have multiple mental approaches to mixing paint, and have the ability to switch between them as the situation calls for it. However, here's a good method to get started:
  1. Choose two colors that "bracket" the hue you are trying to hit- meaning, they are on either side of the hue on the hue wheel. Occasionally one color will hit the hue exactly, but chances are it will lie between two.
  2. Use white to raise both of these colors to the value you're trying to hit.
  3. Mix the two lightened mixtures together until they match the hue you are trying to hit more precisely.
  4. If the chroma (better word for saturation when painting) is too high, mix black and white together to the value of the mixture.
  5. Use this gray mixture to bring the chroma down to the chroma you're trying to hit.

The basic idea is- start with an approximation of the hue, then match value, then match hue more precisely, then match chroma.

Of course, there are complications to this method. The biggest one is that, as written, the hue will likely shift towards blue-purple as you mix if you're using Ivory Black and Titanium white, so you would need to learn to compensate for that. One way is to choose the initial colors slightly shifted away from blue. Another way is to use a more neutral black and white- for example, Cremnitz white and certain brands of Mars black will shift the hues less significantly. You can also use a pre-mixed neutral of black, white, and brown. Note than none of these are foolproof, but they will get you closer. In the end you still need to rely on your own judgment rather than leaving it to the system.

You also need to use some judgment about the chromas of the initial colors you choose- they need to be at least slightly higher in chroma than the color you're trying to hit.

This method can also be a bit slow when painting from the model. As you get used to it, there are lots of shortcuts you can take, or you can use a different method altogether. A few shortcuts:

  • Premix neutrals from black to white in equal steps, which saves time for steps 4 & 5 and also give you "targets" to match the colors to.
  • Also premix the colors to the same values, which saves time in steps 1 & 2.
  • Be less rigid about the order
  • Adjust chroma by choosing initial colors that are further away from each other on the hue wheel (this can be unpredictable, but with experience you can learn to control this).
  • Supplement this method with another method of mixing color.

It's not a magic bullet, but it's a good starting point. It's especially useful if you're used to thinking in HSB on the computer.

I highly recommend http://huevaluechroma.com, a site put together by David Briggs, who teaches out in Australia. It can read like rocket science, but it's good to read and then re-read as you work more. If you do that, more and more of the complicated stuff will make sense. There's a lot of misinformation about color out there, but huevaluechroma.com is all solid, and refreshingly free of bias.

Monday, January 4, 2010

The Value Scale

It can be very useful to assign numbers to even steps value, often referred to as a value scale. I learned to use an eleven-step value scale- 0-10, with zero being black and 10 being white. The number of steps doesn't really matter, nor does which end is white (some people have 0 as white). But, it helps to pick something and stick with it, since you can then build a library of values to expect for different subjects in different lighting situations.

For example, when I go to paint a model under an artificial fluorescent light, I can expect that a lighter skinned model might have something close to a 3 for the shadow value, going up to an 8. For a darker skinned model the shadow might be closer to 1 or 2, with the main lights around 6 or 7 and highlights approaching 8 or 9. These values might change depending on the exact setup. For example, a light source within the scene might force the light values on the figure to compress down further to leave room for the light source to appear relatively very light in comparison.

It's also useful for keeping values consistent and properly differentiating values. For example, under most circumstances the shadow area will not be as light as the areas in light- even most reflected lights can't reach as high a value. So once you've determined the darkest number in the light, you can make sure to always use a darker number for everything in shadow.

The concept is bit difficult to get used to in transparent mediums such as graphite or watercolor because it's difficult to know when you're at a 5 versus a 6. However, in oil paint it's easy to pre-mix colors in equal steps. You can eyeball it, or mix it to match a predetermined grayscale- either from a good inkjet printer, or from a standard grayscale such as the one determined by Munsell System.

Once you're used to the idea in oil, it can be translated to a transparent medium, although it generally is not as exact- often the darkest dark in a drawing changes slightly as a drawing progresses, and thus it's tough to determine what your 0 value is and how the others should relate. As such, there's a period at the beginning where the value scale is a bit nebulous. It can be useful to draw a quick scale on a small piece of the same paper your drawing is on and compare the values in your drawing to this scale to determine the value numbers. Or, you can keep things more fluid and let the scale evolve naturally. Either way, it's useful to decide which pencils are associated with each value- generally you want to leave the harder pencils for the lighter values, and add softer pencils to the darker values (though you can still use the harder pencils over the soft to smooth things out).

It's worth noting that you don't have to be deadly precise with this for it to be useful, and you don't always need to stick to whole numbers. In some paintings it becomes clear that a value must be between two steps, and in some lighting situations the darks or lights might be so compressed that you can't use the whole steps.

Friday, January 1, 2010

Notes: Modeling Factors

The concept of modeling factors builds on the idea of the poster. Within an individual form, the artist decides on a certain number of steps to use in the poster, and these steps correspond to the progression of values one sees in chiaroscuro. Again, the exact number of steps depends on the artist on the technique being used. For example, in more naturalistic techniques the artist might use as many as twelve modeling factors, but in a more impressionist technique five or six modeling factors might be sufficient.

These five modeling factors are quite efficient at representing forms, and are a good base for realistic paintings:
  • Shadow (or Tone)- This area is all parts of the form that are parallel to the direction of the light source or further. It also includes cast shadows.
  • Light - The area of the form generally facing the light.
  • Halftone - Between the tone and light. While the light and shadow together establish the overall mass of the object, the halftone is critical for defining the more precise shape of the object.
  • Highlight - A specular reflection of the light source. This differs from the light in that it usually will not appear where the form is facing the light, but rather in the spot where the light source would show if the form were a mirror.
  • Depth (or Accent) - These are areas where very little light is reaching. Particularly when the light is from above, there will be small pockets where very little light reaches. Generally they are found in the shadow areas, but do also occur in the light where two forms or subforms overlap each other dramatically. For example, underneath the edge of a cloth laying on a table you will usually find a dark line- this is a depth. They're also often found where the feet touch the ground, where necklines and cuffs of clothing overlap the figure, and in crevices like the armpit and the crotch. Depths are quite valuable for showing one form sitting on top of another.
As few as two modeling factors can be used and with intelligence still be quite effective- and very beautiful in the right hands. Note that for certain surface materials some of these factors might not appear, or you might choose to leave them out. For example, matte materials may or may not exhibit a highlight. On fuzzy objects such as a tennis ball there will be no highlight. On very shiny or reflective objects some or all of the modeling factors may obscured by reflections. You might choose to leave out a highlight in situations where the technique you are using doesn't allow you to provide enough information to clearly identify the mark as a highlight.

To create a very convincing representation of form and surface more modeling factors might be needed. In such a case, it can be useful to divide the light, halftone, and shadow areas into smaller and more specific modeling factors. For example:
  • Highlight
  • Light Light
  • Middle Light
  • Dark Light
  • Light Halftone
  • Middle Halftone
  • Dark Halftone
  • Terminator (the division between the halftone and shadow)
  • Dark Dark
  • Medium Dark
  • Light Dark
  • Depth