Performance capture for game development
Performance capture for game development is one of the most powerful tools at studios’ disposal. For years, Animatrik has been helping game designers around the world build realistic, believable motion right into their work quickly and without fuss. Which is crucial when, as every developer knows, you need to create hundreds, if not thousands, of unique assets for every game you make. From walking and jumping to fighting and falling off cliffs, performance capture has been the saving grace of animation pipelines the world over.
With games becoming more complex, leaning into sagas and stories that rival even Hollywood-level productions, we’re seeing the industry undergo a second coming. Even streaming services are finally making quality game adaptations. But it all starts with a coherent narrative replicated digitally and engaged with by players. And the key to all of that is believable, authentic movement. Let’s take a quick tour of how performance capture for game development has become a critical component of building out the latest hits.
Performance capture vs. motion capture
In the taxonomy of replicating bipedal movement, motion capture deals with the body — feet, legs, pelvis, torso, arms, hands, and neck. Everything that gives a person or creature locomotion. Performance capture simply adds detail to that foundation by recording the face and voice of an actor. So there is no versus, performance capture is just a leveled-up version of motion capture, which the master explains in this video from Wired.
Speed
Game developers have turned to performance capture in no small part due to its speed. What in years gone by may have taken weeks to animate, studios can now generate in a matter of hours by recording the movements of a real person. From simple walking to leaping, fighting, and falling, performance capture is a fast way to iterate through the motions required in your game. And it’s the iteration that’s key. Because data can be visualized live on set it’s possible for directors and producers to modify and build on movements for readability.
Types of movement
Character movement in games is driven in one of two ways: animation or script-based. Animation-driven actions are smoother, more fluid, but come with a reduced reaction time for the player. Script-based actions, by contrast, are millisecond fast, but because they’re procedural and generated on the fly they’re often more jerky (you’ll usually find them in first-person shooters where reaction time is make or break for a win).
Knowing which type of animation you’re using in-game is vital to a successful performance capture session. If you’re relying on script-based actions, you’ll need to do away with any flourishes or nuances from actors in favor of a stripped-back performance.
It all comes down to player control versus animation fluidity, and which is most important to the gameplay experience and how the player “feels” the action. In animation-driven environments, players are essentially choosing the next set animation for their character to perform, while script-based gives them near full, albeit less-smooth-looking, control.
Movement that reads
In gameplay, a character’s actions need to be readable from any angle on the x, y, and z axes. Performance capture gives us the ability to record unique movements for a character, watch them back, and go for a second take if, for example, the swing of a sword doesn’t come across well from behind. It’s this ability to iterate quickly and generate unparalleled amounts of clean data that makes motion and performance capture such powerful assets for developers.
And it gets all the more beneficial when you consider the sheer variety of motion that can be captured during a single shoot. Instead of repeating movement assets across characters, each figure in a game can have its own set of motions — resulting in near-real-life diversity of movement.
The face
Outside of bodily movement, it’s facial recording that transforms motion capture into performance capture. Here, reflective dots are applied to the face and cameras collect their movement information, similar to how optical motion capture works. Each of those dots is assigned to a muscle group, and VFX artists in post production use that information to drive the digital mask of the character. More recently, head-mounted cameras have been used to collect facial expressions from dots, and microphones can be added to record a vocal performance at the same time, which can be useful when matching lines to an animation, or even replicating a voiceover performance from another artist.
Rigging and retargeting
Successful performance capture depends a lot on the quality of your character rig. In game development, studios are often working with more low-fidelity models than we might see in films, but because a rig in its simplest form is a skeleton, these can usually be enhanced later if needed. However, the higher the quality of your rig the more believable your character will be.
What’s important is to match the actions of your performance capture session with what your rig is capable of. Humans are bipedal, so if you’re creating a four-legged beast or a thing with wings you’re going to have to retarget some of that motion data. The same goes for unusually large characters, where specialist real-life body rigs can come into play.
When considering performance capture for game development, knowing the rigging of a character ahead of time is crucial.
Workflow
Production schedules on games are tight. Often tighter than films, with entire character movement portfolios needing to be recorded sometimes in a single day. From repeatable movements used in gameplay, to story-driven sequences used in cutscenes there’s a vast array of requirements placed on a motion or performance capture shoot. Which means the quality and reliability of generated data can make or break a game.
It’s also what makes performance capture the ideal method of creating character animation. Game development requires hundreds — thousands — of movement variations to create a believable experience. By planning character choreography early in the production stage, it’s possible to come away with everything you need and sometimes even those luxury options that animators and designers live for.
The levels of realism that can be brought to a game through performance capture are truly unmatched. Traditional frame-by-frame animation can’t come close to replicating real human movement with all the nuances and subtleties that go unseen, but not unnoticed.
At Animatrik, we don’t just open the studio doors and expect you to walk in ready to go. We’ll work with you ahead of time to ensure that you get the data you need for your animation pipeline, building sets, sourcing props, and working with actors to help deliver the best possible results for game developers, studios, and players.