Welcome to Our Community

Wanting to join the rest of our members? Feel free to sign up today.

Sign Up

Interesting Beyond The Pixels : "making Of Lady Jecht" Or "how Not To Make Pixel Art"

Discussion in 'Site news & Updates' started by Zenaki, Aug 21, 2016.

↑ ↓
By Zenaki on Aug 21, 2016 at 11:23 PM
  1. Zenaki

    Zenaki Blackjack & Hookers
    Member Hall of Famer
    • 105/119

    Nov 1, 2015
    Likes Received:
    Trophy Points:
    CC Points:
    Introductory blabber
    HellooO everyone, Zenaki/Chris/some dude here. Welcome to Beyond the Pixels ! I'm here to walk you through the making of Lady Jecht, and tell you about my work process in general. So please have a seat, grab a snack, have a drink, it may take a bit.

    So first and foremost, there's a few things you should know beforehand. I'm an extremely messy guy and the way I work is most probably not the best of ways. And I rarely work with size or colour limitations, if ever, which allows me to get a lot of inspiration from more mainstream artists. My biggest influences are probably Stanley Lau (Artgerm) and Patrick Brown in that regard. And an unhealthy amount of Abysswolf pixel art wise (obviously).

    Okay, let's dive right into it !

    Making guidelines
    Just like when drawing, making guidelines is extremely important. Breaking down what you want into simple shapes helps you figure where things should go and to build the pose you want. Especially if you want something anatomically decent.
    When working with size restriction, I just do them in one go.

    But more often I just do the face/head first. The face is something that's pretty tricky, has a lot of details to put in basically 1/8th of the body. And it is the part that catches our intention first. And cramming those details into pixel art gets extra tricky for readability issues. That's why I pay more attention to it. Plus, who doesn't like pretty faces ?
    So I just put some lines into place till I feel the guidelines for the face are good enough.

    Then I do the rest of the anatomy. So basically the size of the sprites is dictated by the face I start with. Which is appalling in itself, but meh. That's why my works tend to be pretty big.

    With the anatomy done and the pose basically shaped, I draw the main details over it and turn it into a lineart.

    Shading tutorial - All about planes
    Now we're getting into the good stuff. Shading is what builds shapes, can make things pop out more contrasting with others, putting things on the foreground and others on the background. It also allows you to put extra details without having to actually draw them, like folds on cloth and muscle/bone definition.

    And this is where my process goes into absolute chaos. Most pixel artists have their palette readied up before tackling the shading, start out with the main colour, then work their way to the highlights and darker shades on the whole sprite. I just go in without a palette, try to come up with the colours on the go, change them when they don't feel right, add shades when I feel like I need more. And I usually render part by part because I'm impatient and seeing some part rendered encourages me to keep going. So making an explanatory step-by-step of this disorganized part would be absolute mayhem.

    But here are the main principles I keep in mind while shading.

    One plane = one shade
    As long as it stays in the environment, light travels in a linear trajectory. And if you go with far-away light sources, each individual light rays go the same direction. So every point of the same plane gets hit by light with the exact same orientation. Same orientation, same shade.

    Break down everything into planes
    Unfortunately, everything is not as simple as a cube. However, you can break down everything into basic planes. Remember low res early 3D models ? How can you forget those beautiful triangles ?


    But notice how on either Lara, the placement of shadows is basically the same, even if it looks blockier.

    Well do the same for every shape ! If it has about the same orientation to the light, then put them on the same plane and give them the same shade. And with the magic of pixels, you can smooth out the interface between two orientations, making it a curve, not blocky anymore. Et voilà !

    Of course, breaking things down into planes is just something to keep in mind, not something to actually do. Pixel art is already time-consuming enough as it is, dont have time to make low polygon playstation model each and every time.

    Edges are planes too !

    While putting two different shades side by side is enough to make an edge, you can make it more noticeable. Because, in the end, edges don't exist. No matter how sharp your edge is, if you zoom in enough, it is still a curve.

    And as said before, curves can be broken down into planes. And naturally, if the edge is facing towards the light source, there is bound to be a plane that is hit brighter than the rest. Especially brighter than the planes that form that edge. So from far away, you can use a brighter shade to make that edge pop out, making it look sharper.

    This is what I believe to be physically accurate. But the brain is dumb and easily fooled. So even if that edge isn't really facing the lightsource, making it brighter will simply make your brain go "Oh it's an edge, tis okay" even if that makes no sense.

    With this, you can make an edge even when you can barely see one of the planes, or even if you shade both planes the same shade (if they happen to be at perfectly opposite angles of the same lightsources, meaning they are about the same orientation, thus same shade).

    While not really necessary, this technique allows you to make an edge pop out, conveying the intended shape even more than just having two different shades collide.
    Examples of edges highlighted from Super Mario World, Mario & Luigi Superstar Saga and Metal Slug 5

    Reflected light is cool !

    White light is made of all the colours of the spectrum. When light hits an opaque surface, there's a part of the colour spectrum that is absorbed (converted into heat for example), and another that is reflected. The part that is reflected is what goes to your eye and gives it the colour you see. But that light doesn't just go to your eyes, it can also hit other stuff, which absorbs part of it and reflects the rest back to your eye again.
    That's why you may see brighter shades at the opposite side of the lightsource. Hell, sometimes it's just the background that just reflects that light. That's for the physics shenanigans.

    And I use reflected light a lot. Maybe too much even. But you know why it's awesome ?
    • It helps you define shapes even more. Particularly curved shapes, making it look even more 3D.
    • It helps you define how shiny a material is, depending on how bright the reflected light is.
    • You choose what you want to make stand out by making them shine in the darkness.
    And it does all that without needing extra shades ! So yeah, if I could marry reflected light, I would. Reflected light = best waifu.

    Examples of reflected light in Metal Slug 5 and Sonic Advance 3

    Get your priorities straight

    These are all just the principles I use to keep my shading kinda logical and accurate. However I tend to bend the rules when it can help me with my priority : readability while conveying shapes and details. That's why I don't always respect the lightsource, and my shading doesn't always make sense. That's why I overuse reflected light even when there's nothing that should reflect light that way. If I want something to pop out, make the pose more readable or just to make something look pretty, I'll use it.

    Hell, sometimes I bend the rules because it's more fun to shade something another way, or because it's easier.

    Shading progress

    So now that I explained the principles I keep in mind while shading, I can dump the various steps I kept from the shading process.

    Started out with the upper body

    Turns out I wasn't really happy with the face after all, so I tweaked it here and there. First was just shading over the previous lineart, then I changed the nose. The next steps involved tweaking the nose (top) and the cheek (bottom). Turns out I combined the two and ended up with the final face.

    Quick tip for making decent looking faces easily without having to make the sprite too big : make the bridge of the nose a straight line. Tilting it makes things everything more complicated and always ends up needing more space to look as good.

    And so I keep on rendering parts of the sprite until I'm done.

    #1 Zenaki, Aug 21, 2016
    Last edited: Aug 22, 2016
Tags: this article has not been tagged


Discussion in 'Site news & Updates' started by Zenaki, Aug 21, 2016.

Share This Page