Pixel Art: Mayan Statue

I haven't played Child of Light, but I kinda want to. It has a surprisingly indie-esque look to it for a Ubisoft game, reminds me of LIMBO and Braid (among others) in its imagery.

So I did go do a bit of pixel art after mentioning it last time! And after seeing someone on Pixeljoint remaking an old piece of there's and jalonso mentioning how it's always a great exercise to do every couple years, I decided to do just that.

That is old! 8 years old, to be precise. One of the first pieces I felt particularly proud of. It was made as part of the Pixelation palette challenge. It's, uh, pretty bad in retrospect. No real depth, forms are all messed up, and the lightsource isn't exactly consistent. But very intensive.

This is when I remade it in 2009, three years later, which is still five years old. It is much better! These piece still stands up pretty decently. The top is all confusing-like, and the perspective on the tiles is ridiculous, and I was working with figuring out dithering at the time so it's pretty overdithered, but it looks appealing still.

Now it's the present. I've learned a lot! This time I didn't want to replicate it precisely like in the past, I just wanted to capture some essence. I did want to lock down the palette still though. I don't know how great of a decision that was, as it's one of the two major things holding back my piece. Let's just take a look at the final piece!

Among other things, I blew it up! Bigger size was one of the first things I wanted to do with the remake. Ultimately it leaves it feeling a bit empty. I might do a little more working on it to fill int he gaps or something. It's just a huge space, and I'm not used to working in such room. Plus it's been a while since I've pixelled independently. I'm proud of the statue, and the figure is fine (though he doesn't have hair pretty much so I didn't have to bother with character design).
Is it perfect? Hell no. It's wide and empty and there's large flat expanses. However, it was a very good experimentation for returning to some pixel art, and specifically for a key technique I've wanted to practice that I learned about during my time off, mostly during my pixel lesson-writing. That technique is pixel clusters!

This isn't a full lesson or anything, but I wanted to show a couple bits from the process to show how I was constructing stuff.

Most importantly, here's the palette set aside.

I have six colors to work with. Black (SIGH), dark purple, light purple, dark green, light green, and yellow. This is a sub-optimal palette to be forced into working with; solid black sucks, hence why I ignored it entirely in the 2009 rendition, there's not enough contrast between the purples, and the bright yellow is hard to drop in. On the other hand, the green's have great contrast, and the low saturation on the light purple lets it take on a role of buffer if I need it to. I really wish that black wasn't pure though.

This was step one! It, uh, was literally doodling over a grey canvas with red pencil, plotting out general areas. It's a mess. Don't worry about it.

And then I cemented it with color blocking. I eventually dropped the vines toward the end. I couldn't get the curves to look good and was tired of working. They may return to fill some of the empty space if I revise.

This is technically three separate steps compiled into one image - I composed the three colors separately. These formed the basic shapes that I built on to define the whole image. Purple came first, then yellow, then gray.

And here's the only mid-progress shot I had saved. I had it as a spare layer because I duplicated it before trying something - I think it was about the highlight ridge between the eyes, and it worked out.

Now let's take a look at my clusters! Specifically, the shaping of each color individually. This didn't comply with that challenge on Pixelation to try to only use clusters and have no lone pixels, but I kept it in mind that I wanted to avoid stray pixels. The main exception was when I wanted to anti-alias, but I also had a good deal of stray stuff that I used to add a little visual interest to the large expanses of grass and space, as well as to detail the little bits of grass and stars that would catch light.

Ignore the little tabs in the top right - those were from my palette and I forgot to remove them when I saved these. The main goal with my clustering was to be able to define shapes and regions with not only space, but with negative space - keeping both parts formed out of clean lines. With black, two of the best places are the torch inside the flame and the thin outlining around the top of the state and grassy ledge. Both clearly define what's happening, while leaving room for the other colors to augment, even though the outline is only the AA and not made of solid clustering. On the other hand, the figure gets cut up by removing the other colors - some gentle AA could help reclaim some of the edges. Also, this one makes it quite obvious how not-round the moon is.

This one's great the whole way around. Like everywhere. The curves on the statue aren't perfect, especially on the teeth, but are generally pretty clean!

And this one's just even better. I don't think I have anything wrong with this.

This is just both purples on the same image. It does a really good job of showing how the statue is defined by a clear shadow plane and the figure is constructed out of rim lighting.

Here's where things start to get messy. The upper region of the statue and the grass are fine, but the mouth area on the statue relied on green for anti-aliasing, which doesn't look great as clustering.

The light green starts to define the flame itself and the highlights on the statue. I like how the negative shapes from the grass can shape the figure and the statue's shadow side.

On dark so you can actually see it. The flame takes shape, the grass around the figure still implies its shape, and a lot of the front of the statue can be inferred purely from the highlights.

And here's the brights, all put together. All in all less clearly-composed than the shadows, but I still like it.

In general, I think I can be satisfied with my understanding of the positive and negative clusters and how to define forms with both of them instead of just the positive ones. Definitely might come back and do even more to the piece, but for now I'm happy with it as a warming-up exercise.

End Recording,
Ego.