Ubisoft Nano – Squeezing Big Games into Small Spaces
Ubisoft Da Nang officially opened its doors around one year ago, and since then has been steadily building its team and producing the Ubisoft Nano collection of Instant Party games such as Rabbids Wild Race, TrackMania Blitz, and Hungry Shark Arena. These are social multiplayer experiences based on big Ubisoft brands, and feature simple controls, colorful graphics, and strong replayability, all stuffed into a small amount of memory and playable on mobile devices, tablets, and desktop computers.
What is your background, and what is your current role?
Maybe seven years ago, I started to learn 3D modelling by myself, and then about three years ago I took a degree in programming. I wanted to know how to dig deeper into videogames and how they work, and use my background in 3D. I started looking for some relevant roles, and I found a position here in Vietnam for a 3D artist at Gameloft. I sent an application, but they saw I had a degree in programming and asked if I would like to join as a technical artist.
Can you describe what a technical artist does?
JD: Technical artists can cover a range of specializations, like animations, visual effects, or tools, for example. Mainly it involves acting as a bridge between art and technology, and as technical artists we need to keep both sides in mind – how to create assets with a consideration towards how to implement them, and finding optimized solutions for adding them to a project. Much of what I do is creating tools and scripts that help our artists implement their assets into our game engine.
What is the usual process to create 2D animations?
JD: We have three kinds of 2D animations that we work with here. The next is using a timeline editor inside the game engine to create animations by moving objects around the screen or scaling and rotating them. Finally, we also use software called Spine, which is a bones-based animation tool. It’s similar to how a lot of 3D animation tools work, but in this case it works in 2D. You create a kind of skeleton from a series of bones, which we call a “rig,” and attach the object you want to animate to it.
How does the tool you have developed simplify the process? How does it work?
JD: When we make an animation in Spine from scratch, the animators just need to think in two dimensions – the x- and y-axes – whereas in 3D animation you have a third z-axis to think about. This puts a lot of pressure on the game teams to find efficient solutions on areas that require a lot of manual work, like animation and 2D drawing. For our first season of games (2020), we were creating assets, rigging them, and manually creating animations by ourselves.
We were given access to 3D assets with models and animations that were already created, but as we work in 2D, we thought we couldn’t use the content. Then we realized making a tool that could create 2D animations from 3D assets could be really useful, and that’s how my tool Max2Spine was born. It’s developed in the Python programming language, and what it does is look through a virtual camera in a 3D space within 3DSMAX, the 3D modelling and animation software some of our teams use.
What are the challenges of “flattening” a 3D animation into a 2D one?
Spine and 3DSMAX are not related in any way, so we have to export something readable for the 2D software. Basically, this is done using math; using the perspective of the 3D camera, I can project the position of each bone in the rig onto a 2D plane, and then convert those points into a format that is readable by the 2D software.
I can then recreate the 3D bones in 2D using the saved points and positions from each frame, recreating the rotation and translation of those bones during the animation. We use a lot of trigonometry to do all this, and it’s quite funny, because when I was in school, I didn’t like math and I didn’t understand trigonometry all that well; I thought it was pointless.
Are tools like this a replacement for traditional animators and animating techniques?
It does save a tremendous amount of time, though, and the more animations you have in 3D that you want to convert to 2D, the more time you save. Talking about just one animation with the rigging technique, we can now do in about one hour what could take an artist one or two days. Then you multiply this out to, say, 50 animations you want to convert, and the time savings become even greater.
But we still need animators to clean things up and polish, because we lose the z-axis and the depth in this process, and we still need people to go in and make tweaks or changes to give it the final quality we want. We always need animators, but they spend less time working on each animation.
Apr 13, 2021 at 07:48