Recently, I had the opportunity to learn from the Master himself, but I’m not sure if the lesson learned is even something he did.
I’m currently working on a render inspired by Frank Frazetta’s painting for the cover of Karl Wagner’s novel, “Night Winds.” The painting features a naked girl hiding under a fallen tree from a nasty looking knight who hunts for her. I thought about switching it up, with a swamp instead of a mountain meadow, the girl on the left instead of the right, and having an ogre hunting for her instead of the knight. Like everything else I do, it grew, and grew, and grew. Until it was a full sized swamp with three girls. By then, however, the characters were so small as to be hidden by all that swamp.
It just looked wrong.
So, I went back to the original painting. That is when it hit me.
Frazetta and I approach composition from equally opposite directions. Or, it appears that way.
I approach 3D art like a model railroader. Those of you who have played with trains probably know what that means. For those of you who don’t, putting together a model railroad goes something like this: oh, this is a cool model stockyard and I like this Oldsmobile and that is an awesome tree and I have to have this set of crates and that pile of junk is great and look at that battleship and that building is fantastic and so is that building and that one and that Model T and that collection of figures and … uh, I should probably leave room to run a train through here, huh? When I design a 3D art scene, I pretty much do the same thing, assembling as many props as allowed by law into the most amazing set you’ve ever seen. Then, I bring in the characters and pose them in such a way as to still see the whole set.
The end result is the figures, the actual characters telling the story, are lost in the details. This is most apparent in my render for “Magical Yule.” I was so bent on that storefront and snowfall that it required 14 figures to fill the scene, and their stories are lost. Now, for “These Mean Streets,” it was deliberate. That scene was about capturing the look and feel of Daytona Beach circa 1974. The figures were details. Yet, in that deliberate composition, the figures become characters, all telling their stories at once, and it works.
Frazetta didn’t do that. Okay, so I really don’t know if he did or not. So far, I haven’t come across anything saying he didn’t, but it sure doesn’t appear that he did. In “Night Winds,” all we have are the girl, the knight, his horse, and the tree. Some vague mountains appear in the background. The rest of the set is there just to hold the figures together. It appears that he painted the figures first, large, up close, and in the viewer’s face. Then, he added the set. The end result is a sense of immediacy and intimacy. All of his paintings are like that.
I scrapped that beautiful swamp.
I brought in my four characters into the empty screen, positioned them, and posed them. Then, I moved the camera in close until they filled the screen. They take up about three fourths of the scene now. After that, I positioned the set around them, prop by prop. I positioned the props to the characters, instead of posing the characters to the finished set. It is a much more powerful scene now.
Here is the test of the swamp set, without the characters.
The girls will be on the left, and the ogre on the right. The space gives me plenty of room to pose all of them dramatically. The set looks packed, and it sorta is. I need that many props to make it look like a swamp. But the main reason it looks packed is all that swamp set I had before is now compressed into a small, intimate area, instead of being spread out. Consider it artistic critical mass. There isn’t much to the set beyond what you can see here. This type of set design was actually easier than my old grab it all and shove it in model railroading method. Everything is placed deliberately for maximum effect and interaction with the characters. Because the set is so small, it also saves resources on my aging computer.