BioWare’s Animation Conversation System
Author of this post: Beth A. Dillon | About Blog Authors »
I have to admit, what’s going on over at BioWare in terms of animation excites me the most right now, but that’s because I’m a game writer. With their tools for the conversation system, writers too can take part in the aesthetic elements of the game through camera angles and manual control of movements to create performances. Brad Kinely, Lead Animator on Mass Effect and Mass Effect 2, provided insights into BioWare’s process of development at the Vancouver International Game Summit this week.
Just to explain first, if you haven’t played any BioWare games (shame on you), the gameplay revolves a lot around branching dialogue and choices you as the player make during conversation. They’re improving the interactivity of their system with games like Mass Effect and Mass Effect 2 by merging live cinematics with conversation choices. So how do they mix it up enough to create the allusion to varied experiences for the player?
The branching tree system of dialogue is split up with a timeline, a procedural camera, audio, and the ability to add animation. The animation team started with a relatively small segment of actions (around 100) and resolved the issue of diversity by making an additive system. Basically, the system allows for combining actions of both the face and body. With all the various possibilities (for example, head nod and left arm movement), they have a rather robust set of options for the writers.
BioWare’s overarching system is made of two major pieces of software—RoboRad controls body movements, where as Face Effects strictly controls facial movements. The Face Effects system analyses a line of dialogue, looks at how long the line is, and goes through a preset of head gestures. In doing so, it ensures each time a line is processed, the head movement is unique. Characters are made to react procedurally to the system.
For the face, bone offsets are weight and treated in a certain way. With the body, small gestures are weight differently each time, which makes the content look broader in scope. Writers can go into the system themselves, select which emotions to use (given certain face and body weights and actions), and the system takes care of the rest. Once the base compilation is placed down, the writers can go in afterwards and add in specific gestures.
The team anticipated localization concerns by coming up with a process for dynamically replacing shapes of movements in the bone offsets. With this process, they were able to easily transition the content to different languages. Same went for interpreting meaningful gestures.
I got a sneak peak at a creature for the upcoming Mass Effect 2, and yeesh it’s nasty. With that one, they faced particular facial expression challenges, since well, he had no lips but instead a freaky piranha mouth. His behavior states are a little limited from the human spectrum—you know, he’s not exactly going to be “flirty.”
In every case for BioWare, as helpful as a procedural system is, the tools always allow for manual control that’s easy to apply.
