Stanford’s New Animation Simulator Is Focused On Finding Interesting Results
Simulators can jump-start animation work, but often return an overwhelming array of options for the animator to sort through. A new browser refines these options to a more manageable number.
See the video here.
By Andrew Myers
Computer animators who are tasked with bringing imaginary worlds and characters to life are aided by simulators that can model the many possible ways an object or fluid can move through physical space. Known as “solvers”, these simulators offer a significant head start on animation work. But there is a catch. As computers have gotten faster, these solvers often create too many options for the animator to efficiently sort through to find the correct one.
“A simulator can return thousands of options. It takes so long to sort through them that these useful solvers cannot be used to their full potential,” said Purvi Goel, a computer science doctoral student at Stanford, who, together with her mentor, Professor Doug James, has created a new approach . to refine the search and limit the results to the most promising options.
They call it “Unified Many-Worlds Browsing,” and they’ll be showcasing their approach at the upcoming SIGGRAPH 2022 conference in Vancouver, BC, the premier annual gathering for computer graphics and interactive techniques. The approach could both speed up the process of finding animators and, by limiting the options that simulators have to navigate, reduce computational time and expense. Eventually, the researchers hope, this will allow artists and scientists to interface with solvers with unparalleled ease and efficiency.
“We have this interesting conflict in the field that every year computers are getting faster, more parallel, and they can do more, but it’s humans who are maxed out,” said James, who is a professor of computer science.
Parameters and time
Solvers are controlled by input parameters – the physical dimensions of the object in the animation. In one example, the researchers used unified navigation across multiple worlds to animate an imaginary bowling alley. Settings include start speed and ball position.
A solver simply loops through the parameters incrementally, one at a time, and simulates the many potential outcomes over time. The key point is that there can be many input parameters, and the longer and more complex an animation is, the more parameters and computation time it takes to create simulations, and the more options are generated.
“If you’re an animator and you have an idea in mind, you have to adjust these settings manually and then sort through all the results one by one,” Goel explained.
Unified navigation across multiple worlds allows animators to create “queries” to refine options to facilitate the process of identifying very specific outcome options. Queries can, for example, limit results to those that most interest the facilitator – including those that are either “within” a desired range of possibilities, or, conversely, exclude those that are “not within”. ” this beach. Sample animations are available on the researchers’ website.
When asked to animate a stuffed armadillo falling down a spiral staircase, for example, the animator can create a query in the Unified Browser that says, “Show me only options where the armadillo falls up at the end and comes to rest on the lowest stairway. Tasked with animating Jell-O cubes leaping into a bowl, the animator could limit the options to just those where all the cubes stay in the bowl, or perhaps one, but only one, falls. A small blue ball thrown at an imaginary sandcastle might specify “give me samples where the front left turret is broken” or “top turret only”.
If the queries are well defined, unified navigation across multiple worlds could potentially reduce thousands of possible options to a handful that is attractive to the user, Goel says. Plus, the fact that this navigation framework can work with everything from fluids to smoke to frac solids – why it’s described as “unified” – sets it apart from previous options, which were limited in scope to a only phenomenon based on physics.
find the needle
James says unified navigation across multiple worlds can pay surprising creative dividends. The browser, he says, can increase the likelihood of finding that “needle in the haystack” option that the animator might never have found with the solver alone.
“We’ve found that it can help find unlikely outcomes that will surprise – or even delight – viewers with their novelty,” James said. “‘Oh, wow, the bullet went right between those two turrets and hit the other one perfectly.'”
The browser has clear implications in the fictional animation space, the researchers say, but could also be adopted to benefit engineers, physicists, biologists and other scientists who increasingly rely on simulations in their work.
“Unified navigation across multiple worlds certainly makes the animator’s job faster and easier,” Goel said, “but we think it also makes it more fun.”
James is also a member of Stanford Bio-X, the Wu Tsai Human Performance Alliance, and the Institute for Computational and Mathematical Engineering (ICME). This material is based on work supported by the Department of Energy’s National Nuclear Security Administration.