The Battle of Dantooine
So where do you start when taking on the massively impractical task of trying to make an animated film by yourself? If you’re me, educated as an engineer and trained by lots of smart product managers over the years, you try and remove as many variables as possible and tackle the simplest possible version of your problem, while retaining the challenges that make your vision special. Unsurprisingly for me, this also meant Star Wars had to be involved.
For what I’ve been calling my “technical proof-of-concept”, I selected a 6 minute short that was part of the 2003 animated Clone Wars ‘micro-series’ made by the legendary Genndy Tartakovsky (Samurai Jack, Primal and other seminal classics). It’s all 2D, done in Tartakovsky’s hyper-stylized angular style and largely involves Mace Windu beating the crap out of a bunch of droids. No dialogue (so no lip-sync to worry about) and no hair (because Mace Windu — and simulating hair movement is hard). There are, however, flapping cloaks, and smoke, and shockwaves and a lot of character animation — but those are all things I want to learn how to do. The purpose of attempting to replicate this short in 3D is to develop a workable animation pipeline for my original short films.
A while ago (before I started doing this full-time) I tried to model Mace in 3D from scratch, and made some reasonable progress, but the prospect of modeling all the droids, Clone Troopers and vehicles from scratch was really daunting. The style I wanted was something like the Disney Infinity game, which famously united all their disparate properties with a stylized unified art style, and I spent some time searching for free models from the game I might use. I happened upon a pretty active game modding community that had figured out how to rip the models and textures from the game in order to modify them with alternate costumes etc — and they kindly gave me a script to do the same myself, and walked me through how to do it.
Suddenly I had access to beautiful models of every character (Star Wars, Marvel, Pixar, everything) in the game — not just the models, but their rigged skeletons and texture maps too. And since my project is totally non-commercial and entirely for education purposes (you hear that, Disney?), this shortcut seemed like fair game. The internet is a truly wondrous place sometimes.
It wasn’t all plain sailing: The Disney Infinity character models all have weird toy-like spinning arm joints, which in Mace’s case in particular mean his robe deforms in a very weird way. I spent a while trying to fix the model and re-rig the skeleton and quickly realized that the process of skin-weighting and rigging is something that takes years to get even moderately good at — so I began a search for task-based collaborators on Fiverr, Upwork and Behance. The funnel from prospects to qualified collaborators is sharp — from 10+ people that I spoke to, only one has delivered high quality work. This is an ongoing and complex part of the project. The Mace character is center-frame for most of the short and the model needs to be high quality and well-rigged to make the animation itself a reasonable process.
Also some models just aren’t in Disney Infinity. The short has a giant floating enemy vehicle (called a Seismic Tank) that creates massive shockwaves and generally causes havoc — but it’s an esoteric enough part of the Star Wars canon that it isn’t in the game. Fortunately the fandom has me covered here — there was plenty of reference material online and I was able to model and texture it from scratch.
So to sum up, I was able to remove a massive amount of work and variables for this technical POC. Script and storyboarding are taken care of by the original short — I can even use the original’s audio and not bother about that. The framing and shot composition is a lesson in animated cinematography for me. While I do need to do some work on all the models, a huge amount of the modeling and texturing is already done and ready to go.
Of course, that still leaves a truly impractical amount of work for me to do alone. I’m still working through look-dev (‘Look Development’ means figuring out the specialized aesthetic of your film) and learning lighting / materials at the same time. There are still a bunch of things I don’t know how to do, and even more that I haven’t even realized are going to be a problem. I’ve made simplistic efforts to summarize the open questions, but I know this is just scratching the surface.
And some of this is about to go out the window really, because I’m going to switch core animation engines from Unity to Unreal (for a number of reasons that I will detail in a separate diary entry). I’m not entirely sure how long this will set me back, but from the preliminary evaluations I’ve made, I think it’s the right call. Early experiments are promising.
That’s it for now. I hope this brain-dumpery is entertaining. I have found the scale and complexity of this work is hard to get across in conversations, and I’m starting to get used to people’s eyes glazing over the moment I say ‘skin-weighting’. Which is understandable, to be fair.
Follow me if you’d like to read more stuff like this. I’d love advice on pipeline and workflow, offers to collaborate or connect, and anything else. Sound off in the comments or shoot me a message.
