For this first tutorial, let’s start with the AEC Routemaster, London’s legendary double-decker bus. Built from 1954 to 1968, it proved to be so resilient it remained in service up till 2005 on some lines. Thanks to this longevity, it established itself as one of London’s most popular symbols, alongside the black cabs and Big Ben.
Disclaimer: as with any tutorial, the methods presented here are purely subjective and are based on my personal taste rather than immutable rules. The process shown below is working well for me, but it’s certainly not the only way and I’m sure there are plenty of alternatives that are as efficient. Moreover, what’s shown here is pretty much a best-case scenario. More often than not, things are slightly messier and the steps tend to overlap.
From a technical point of view, I used GrafX2, but any paint program able to manipulate individual pixels will do the trick (to name a few: Photoshop, Promotion, Paint Shop Pro, GraphicsGale, GIMP).
1 – Reference gathering and initial sketch
Reference gathering is optional but rarely a waste of time. It will make sure that your end result is somehow believable even if it’s drawn in a cartoony style.
Drawing a sketch on paper isn’t always needed either, but I found it’s a good way to quickly try out several options. It also forces me to really understand the structure of the subject, which always proves helpful down the line, saving time and avoiding some backtracking.
2 – Cubism
I usually start blocking out the subject with simple shapes in basic colours, to get an idea of the overall balance, contrast and colour palette without getting lost in details.
At the same time I want to decide on the size of the picture, which is one of the most important decisions to take. Too small, and I will struggle fitting in the details I want. Too big, the workload will rocket and the end result won’t look like pixel art anymore. In the case of the bus, I chose the size using the wheels as starting point. I had a good idea of the size they can be (there’s only a limited amount of options for good looking circles with a radius of a few pixels) and the rest followed, even if other parameters had to be taken into account, such as the number of side windows (4 windows that are n-pixel wide + 1 pixel inbetween each + 2 pixels on each end = the length of the bus). To choose optimal dimensions it’s really important to think in terms of pixels at this stage.
3 – Outline
Obviously this step will only be needed for pictures that have an outline. When present, drawing the outline helps get a better idea of the level of details I can afford and also helps refine the silhouette, for instance the rounded parts of the roof at the front and the back. I’m not yet thinking about colours too much, but that’s coming next.
4 – First pass at colours
I’m now picking colours that are harmonious and that fit the mood, style and message I want the picture to convey. It’s never easy, especially as I also want to make sure the palette is well thought out: which colours are the most important and the most used (not necessarily the same ones), which ones can be shared between different elements, etc. Pixel art often means optimizing the colour usage with the goal of using as few colours as possible while keeping everything looking nice. Quite often the number of colours is a power of 2 (4, 8, 16 ou 32 colours for instance), mimicking the technical limitations of computers and consoles from yesteryear.
5 – Shading and details
I can now add shading that simulates the effect of the lightsource on the surfaces(originating from the top left here), as well as details. I make sure to include only details that can fit in the small resolution, and not cram as many as I can, otherwise the picture will end up looking noisy and unfocused. As Antoine de Saint Éxpury once wrote: « It seems that perfection is attained not when there is nothing more to add, but when there is nothing more to remove ». When creating small pictures it is very important to simplify the subject as much as possible to ensure it will remain perfectly readable despite its small size. Each detail must be justified and help to understand the overall image better. I also pay special attention to the colour contrast of small elements as they shouldn’t catch attention more than necessary.
6 – Antialiasing
I now add anti-aliasing to mitigate jaggies (stairstep-like lines) and soften shapes. The method is not overly complicated but requires a bit of practice and the ability to take a step back (sometimes litterally). It’s also a matter of finding the right balance between being too subtle in applying it (the end result will still look blocky) and being heavy-handed (the end result will be blurry and sloppy). Even more importantly, it really pays off to be consistent and apply the same rules everywhere. Beside that, it’s also the right time to finalize the colour of the outline and make the part exposed to the lightsource a bit lighter.
7 – Palette polishing
I’m nearly done. I tweak the palette by introducing a bit of randomness in the colour ramps – adding a bit of blue on one colour, then a bit of red on the next one, etc. This helps make the picture look more organic and less clinical. I also want the overall colour scheme to be warm, so I add a bit of red and green on the greys so they’re a bit brownish. Unless there’s a good reason for it, I always avoid basic colours (also known as “coder colours” !) and perfect greys which look too artificial. Finally, I adjust the picture’s contrast, brightness and saturation – even though the changes are not always visible to the naked eye 😉 This is also typically the only step for which I use Photoshop.
(via 16 Couleurs)