Pixel Art - How to Build Objects - Part 2
Now that we’ve gotten into some basic shapes using pixels, I wanted to visit projection and show
With regards to pixel art, there are 4 basic projection methods that you can use to build your objects as listed below:
Flat-On, otherwise known as Orthographic Projection. Typically, it’s this type of projection that you gives that “cartoon” or “animation” aesthetic; because you’re only viewing a subject from a single angle, you can focus on more convincing and realistic proportions for your subjects without having to worry about depth or location in 3D space. In the early days of gaming, because of the limited technical capabilities of early hardware, this was the most popular and easy-to-understand form of projection. Examples include art in the sidescrolling genre of games, such as Mario, Kirby, Donkey Kong Megaman, Metroid, Castlevania, etc.
3/4 Top-Down. Incredibly useful for building out the length and width of subjects with while giving subjects the subtle illusion of height. Used a lot in 2D games sporting large maps, such as those in the action-adventure and RPG genres. Examples include the art in The Legend of Zelda, Final Fantasy, Dragon Quest, Frogger, Gauntlet, Metal Gear, etc.
Isometric Projection. This is where we begin to get into a sort of “pseudo-3D” projection, in that the subjects of this type of art are represented along all three axes. Because of the static nature of 2D cameras, the way it represents these axes is very precise and does not change. In gaming, examples of this artstyle include games such as Super Mario RPG, Kirby’s Dream Course, SimCity, Sonic 3D Blast, etc. This type of projection, because of its pseudo-3D nature, is very popular in 2D management and simulation games. *Note that in Pixel Art, Isometric projection is not TRUE isometric projection, which requires mathematically precise calculation. Isometric Pixel Art is typically more an aesthetically pleasing simulacrum of true isometric projection.*
Perspective Projection. What we would typically call “True 3D”. Incredibly hard to make convincing in low-definition pixel art, because of the lack of requisite amount of pixels.
Starting with the 8-bit/NES era, let’s take a look at how some of the objects rendered onscreen were constructed. Since we’ve done a brief look into the 1984 classic “Duck Hunt”, it seems to be a pretty good starting point.
Looking at Duck Hunt’s main background. Specifically, this tree:
How might I try to remake this tree? One technique I’ve made use of in the past: Blockout.
You can think of the blockout technique as drawing out a general shape for your subject in the very beginning, and slowly refining the shape to meet your needs. As an example, I’ve recreated a general blockout for the tree, with no leaves (so as not to confuse myself, we can add the general shape of the leaves later).
With the general shape of the tree down we can begin to refine the position of the branches as necessary, as well as proportions. Since this is just example, I’m not worried about making 1:1 reproduction of the Duck Hunt tree.
It’s beginning to take shape and become a lot more tree-like! The important thing about this technique is ensuring that the shape you’re making has an instantly-recognizable and distinct silhouette. I should be able to, at a glance, reconigize what is being shown on screen… or at least recognize the idea of what’s being presented on-screen. Or on paper, if working traditionally.
Some further refinement for the rough edges and some darker spots on the trunk to indicate bits of shadow:
The nice thing about doing this technique at lower resolutions is that you don’t need a lot of refinement to make your subject look good, just enough to get the point across. If you’re familiar with traditional art, this is very similar to the act of “thumbnailing”.
For the time being, I think we can leave the tree and begin to focus on blocking out the leaves on the tree.
And we’re pretty much done! Here it is without the grid:
Again, not perfect, but it gets the point across. And again, because of the low resolution, I could theoretically make many iterations of this tree and decide from there which design I would like best, similar to the thumbnail process.