Find The Design

A few months ago I was asked to do a talk at a local concept art workshop. My last talk, "Thumb War", dealt with the iterative process one goes though at the beginning of a project, using little pictures to communicate big ideas. I felt my new talk should pick up where "Thumb War" left off, discussing what happens AFTER a thumbnail is selected -- the fleshing out part. More than a year had passed, so I also felt I should incorporate new things I'd learned, new philosophies I'd embraced.

What follows are the notes and crude visual aids I created to give this talk, which I titled "Find The Design." I hope you like it!

---



Every company I've worked for has had a buzz word -- some catch phrase to describe the things they wanted done, and how to go about doing them. At Raven it was "Bluesky up some ideas." i.e "The sky is the limit, so go nuts." At Vigil it was "Make it diesel." Everything needed to be crackling with energy; nothing could be weak or second-rate. At Blizzard I constantly heard the phrase "Let's chase that." Everything was chasing, chasing, chasing.



This verb "chase" never sat well with me, for two reasons:
1) It implies that you never catch the thing you're chasing.  Exhausting!
2) The chase creates a sort of tunnel-vision, setting expectations which can result in disappointment.

Imagine two single people looking to be in relationships. Person A formulates a very specific notion of their ideal mate, and pursues them doggedly. "The partner of my dreams is 5'6", brunette, into horror movies and wears black. I will accept no substitute." Person B's approach is to simply go about their life, doing the things they'd do naturally, and, in the process, FIND their mate. Person A may get their quarry, but since their preconceptions were so specific and so idyllic, they may feel let down. Further, they probably glossed over other, potentially superior mates that weren't even in the running.

Better to take the route of person B, to be finding rather than chasing, to be alert to new possibilities as they arise, and to be pleasantly surprised by life's random offerings. Malleable. Organic. This is how I propose we design!



In school we're all taught the three 'R's : Reading, wRiting and aRithmetic. It's catchy, but come on -- two of those those don't even start with the letter 'R!" I present to you now "The 3 R's of Concept Art", as I believe all its problems can be lumped into one of the following categories : Research, Ratios and Rendering.

Research



Chris Metzen, vice president of creative development at Blizzard Entertainment, said something very poignant to me in a concept review meeting. I glibly remarked that the ideas I was presenting were unoriginal, to which he replied, "There's no such thing as originality, just good spin!"

Human beings are basically monkeys. We see our neighbor do something well, and we try our best to copy it. An artist sees a majestic sunset, or a nude, or a fighter jet and is suddenly inspired to make their own version. Monkey see, monkey do.

Once we've dispelled the ludicrous notion that there may, in fact, be something new under the sun, we can begin to truly appreciate and synthesize ("spin") that which already exists, lessening our mental load considerably.

I like to use the example of Tachikomas, the little spider mechs from "Ghost In The Shell." Their design is basically a "spun" version of a jumping spider. Rather than hunting down another artist's rendition of a mech and using that as reference, Masamune Shirow used the main attributes of something SIMILAR. The result is familiar, but unique.

Consider what's needed. Make observations, imitating the things you feel are important. Do this enough, and you'll begin to retain information, creating a visual library you can go back to when Google Image Search is unavailable.

Ratios



My friend and brilliant artist Toph Gorham introduced me to the concept of "Phi", short for "The Mean of Phidias" aka "The Golden Ratio." From his backpack, he produced a wooden measuring tool called a "Golden Section Gauge" that could be held up to things, close or at a distance, to see their divine proportions (or lack thereof). When I asked him if such a tool was necessary in designing, he replied, "No. Good designers think this way instinctively."

Most artists are not mathematicians. What we should take from the idea of Phi, or any of its many other complicated guises, is that things tend to look better when not divided evenly in half. When there's more mass on one side and less on the other, when the front doesn't look like the back, when there's more light than dark or vice versa...these harmonious ratios invariably produce a more natural, balanced look.

"Avoid half" has become a bit of a mantra for me, in art and in life. Think of all the instances where people prefer non-halfway things...

We like our drinks ice cold or piping hot. No one craves a luke-warm beverage. (Avoid half!)

We tend to think less of someone who's "on the fence" on an issue. Their indecisiveness stalls everything. (Avoid half!)

Neutral countries have neither alliances nor enemies. As soon as someone falls to one side of neutrality or the other, things begin to get interesting. (Avoid half!)

A "middle of the road" personality is neither coming nor going. It's hard to tell what they're all about. (Avoid half!)

We describe shoddy efforts as "half-assed." (Avoid half!)

Can you think of others?

Where this applies in art :

-We see in nature, in architecture, and in machines countless instances where things are not exactly mirrored. Its top is different from its bottom, its left is different from its right, its front is different from its back, etc. Carlo Arellano, concept artist and instructor, once pointed out that most things have a "business end." The mouth on a shark, the canopy of a tree, the lens on a camera, the blade of an axe, the entryway of a courthouse -- there's a hierarchy. "This thing is all about THIS, and everything else is secondary." Look at the things around you. Observe where their major "cuts" occur, and marvel at how seldom they fall at exact midpoints. This natural drift creates a sense of movement even in static objects. Use it!

-It's nearly impossible to get a directly frontal, side or rear view of anything.* To do so requires that we consciously align our head with the object, and close one eye to remove our stereoscopic vision. We see most everything from some angled, non-halfway vantage point. (*This is why orthographic concept sheets ALWAYS look unnatural.) Life repeatedly gives us a more generous view of one side of a thing than the other. Pick a side to give prominence to, and allow the other to be less prominent.

-Shadows rarely bisect form. One cartoonist who understands this is Mike Mignola, creator of Hellboy. Look on the graphic above how there's more black in his subject than white. He's given it over to darkness. On a more micro level, he's also given the bat over to darkness, with only a couple triangles of light poking through. The skull has been given over to light. See how the delineation doesn't occur at the halfway point, but somewhere to the other side of it. A pleasing ratio indeed!

-local colors // In the alien graphic below, we see what local color breakups look like when they're divided in half (1 & 4) vs. when they aren't (2, 3, 5, 6). Can you see the difference?



-cool to warm color // You want to have a mixture of both, but no absence of one or the other, and certainly not a 50/50 split.

-straights to curves // An image that's all curvy is rubbery and weak, and requires some angular straights to give it integrity. Conversely, an image that's all straight is coarse and rigid, and needs curves to give it flexibility. All in all, though, you want an image to be MORE curvy than straight, or MORE straight than curvy. Include enough of the opposite to give it variety, but stop there! The direction you choose will depend largely on your subject matter, and your opinion of it.

-In most pleasing compositions, it's typical for the subject to be slightly (if not drastically) offset, so as not to divide the canvas into two equal parts. Balanced compositions are ones with pleasing ratios.



-A quick word on tangents : Tangents are where two objects seem to just barely touch. What is this, if not something that's halfway to having depth? In example A (above), we can't quite tell what the deal is with Tangent Man. Is that rock about to hit him? What's going on with his neck and his nose and his ear? They all seem to be encased in that egg-shaped head of his. In examples B and C these issues are dealt with via the power of overlap. By putting things in front or in back of each other, "picking a side" as it were, we create a much greater sense of depth. Sometimes, however, it may be best to "clear" the tangent completely by spacing objects apart from each other, as in example 3.

Rendering



I don't want us to think of rendering as "beautification" or "finish" but rather the simple showing of material or form. In the above graphic, Sean "Cheeks" Galloway used 3 values and 6 colors to show what his version of Panthro is composed of. For him, that's all it took.

Cast shadows are employed to show form and create contrast in an image, but their overall pattern should not be random, but rather a planned "design within a design." (Arnold Tsang taught me that these shadow volumes should not cut the canvas/subject into equal dark and light halves.) You can see how, in the posterized version of Frankenstein, most of the image is dark. Graphically, the light pattern is strong, and shows the planes of his face clearly.

Occlusion [Apologies to the froggy artist. I couldn't find your name to credit you.] was described to me as the places on/in an object where no light can reach, even under very bright conditions. These shadows serve to separate objects from each other. The slit of Frankenstein's mouth is an occlusion shadow, as well as the creased area in and around his nostrils. Think of them as the deeper nooks and crannies where dirt would collect. Unless in close proximity to another object, the outer edges of things do not have occlusion.



The quickest way to show the volume of something is by giving it a strong silhouette. You can think of it as a kind of "pre-render" during which you pose the question, "If this was as far as I was going to take this, would it still communicate my idea clearly and stylishly?"

Artist Nick Carver routinely examines his Maya meshes with a flat, black shader, so he can check his silhouettes for maximum read and appeal.

In example A (above) we have a symmetrical jack-o-lantern where the features are fairly aligned with each other, the "gutter" around them fairly uniform. Watch what happens in B, when we begin to break symmetry, twisting the stem, moving the eyes and eyeballs inward and expanding the mouth. Now look at C, where the features are offset from their "halfway" points. The silhouette is still flat, but the design appears more dimensional and interesting.

---

In this next part of the talk, we're going to look at how we can use different rendering methods to "find the design." Here we have three identical silhouettes...



Let's see what slapping a couple different local colors on them does. In the graphic below, we can now see a division between different areas -- skin, nylon fabric, and, I dunno, leather. It looks like it could be some kind of alien spy. We are now one step closer to finding our design, and we haven't drawn a single line.



...But say we did want to start with lines. Below are three more contiguous treatments, where the general silhouette was adhered to, but broken in a few places. In example A, we've reversed the direction of the head and come up with a completely different design, subtracting from the silhouette where needed. In B and C we've added to it. Such massaging of the silhouette keeps us flexible. We are, after all, finding the design!

Note : Lines that seem to repeat shapes are like echoes -- a melody. To add harmony, occasionally oppose them with marks that go in the opposite direction.



Here's a graphic that demonstrates the power of rim light. Typically rim light isn't applied until later, but if it helps you find the design, use it early! Also note how rim light has a gradual intensity falloff; it's not the same value throughout.



Here's a demonstration of how different rim light and occlusion shadows can alter the same design in drastic ways. Play around with concave, flat and convex shapes. Find a solution that pleases you!



It's the same with cast shadows. The light pattern can drastically change how the interior planes are seen. Note the difference in edge treatments, and how each has a particular role in describing form. (Note, too, that the silhouettes have been altered in the finding of these designs. Nothing is set in stone!)



Highlights, which are also typically reserved for the finishing touches, can be used to find design. The greater the intensity of the highlight, the more reflective it is. Notice how, even without any other details, we can see a suggestion of the interior planes. From here, filling in the gaps is easy.



Parting note :
When I moved into my first house, I didn't have a clue what to do with it. I'd only lived in apartments before, and wasn't used to a space that required as much furnishing or upkeep. My mother gave me some great advice : "Just live in it for a while. The house will tell you what it needs." So it is with designing. Not all of our ideas come fully furnished, but if we live in them for a while -- if we're open to finding rather than chasing -- they will tell us what they need.

Now go forth and find!