SMFA at Tufts animators see their work spring to life with the help of high-performance computing
For Kayla Yewon Park, A24 (BFA), a 3D animator and recent graduate of the School of the Museum of Fine Arts at Tufts, waiting for a computer to catch up with her imagination used to be an exercise in patience.
“Rendering just a 10-second segment of a 10-minute animation used to take anywhere from 24 to 72 hours depending on the scene’s complexity,” she said.
It would either bog down her laptop so that processing anything else at the same time became a chore or dominate a desktop computer in one of the SMFA studios or computer lab.
All of that changed last semester when faculty and staff at SMFA teamed with Tufts Technology Services to launch the school’s first render farm that utilizes Tufts’ High Performance Computing Cluster. The cluster, housed at the Massachusetts Green High Performance Computing Center in Holyoke, Massachusetts, typically supports large-scale computing in the sciences and engineering; now it’s assisting the visual arts for the first time.
The SMFA render farm harnesses the collective power of multiple computers at once. Animations are a lot like the flip books many of us used to draw and then staple together as kids. Each frame in an animation is like a page of the flip book. Rendering or generating all those frames, particularly for 3D animations, on one computer takes a long time. A render farm speeds up that process.
The High Performance Computing Cluster typically supports large-scale computing in the sciences and engineering. Now it’s assisting the visual arts for the first time.
While it might take one personal laptop 100 hours to put together the animation, a group of 10 computers will take just 10 hours, because each will have been assigned a different batch of frames. Each delivers its frames back to the user, and they are automatically collated into order and ready to be viewed.
Ben Aron, who manages the SMFA studios supporting digital media, said that after students graduate, many go on to secure jobs in studios with commercial-grade render farms.
It was Professor of the Practice Kurt Ralske, the chair of media arts, who initially brought forward the rendering-time challenges that SMFA students faced and suggested the render farm as one solution. That’s when Aron and his colleagues got moving on bringing it to reality.
“Our studio spoke with SMFA’s dean, Scheri Fultineer, about preparing SMFA students for the workplace by being able to offer them experience working with Blender, a popular animation software used in professional studios,” Aron said. The open-source 3D creation suite is widely used by artists and commercial studios for tasks like rendering, simulation, modeling, animation, and editing. But Blender works best when backed by significant computing power. As a school that prioritizes interdisciplinary thinking and problem-solving, SMFA reached out to Tufts Technology Services to collaborate.
Park was one of the first students to pilot the new technology in the Advanced 3D Animation Class taught by Professor of the Practice Cristobal Cea last spring. There was definitely a learning curve. Still, even with a rocky first run, not only her time but the scope of work she could imagine for her final project, Dreamscape (2024), benefited significantly.
The piece was inspired by Park’s vivid dreams. “The film begins in a cloudy dream room, with the main character—a chair—guiding viewers through an abstract dreamland,” she explained. “The landscape constantly changes with time and camera movements.”
Using the render farm, Park leaned into Blender and After Effects software to add special effects that blurred the borders between surrealism and realism. As an emerging 3D artist, she hopes to use a render farm for future projects.
Meanwhile, Cea did a test run of the render farm with a group of Master of Fine Arts students in his Experimental 3D Animation course. They specifically used it for the technique of ray tracing, which simulates the lighting of a scene, predicting how it would travel as it bounces and scatters across virtual surfaces.
“To get good results with this method of rendering, you have to simulate hundreds of millions of virtual photons, a process that benefits from the technical infrastructure of Tufts’ High Performance Computing Cluster, which is equipped with powerful Nvidia hardware that is particularly good at highly parallel processes that require hundreds of millions of multiple simultaneous calculations,” Cea explained.
Cea sees this as just the beginning of an ongoing collaboration with Tufts Technology Services to give students and faculty the opportunity to try creative technologies that benefit from computing power, such as generative AI.
“This particular type of computational approach that is good at simulation is also particularly good at AI image generation,” he explained, adding that he can imagine combining Blender with AI tools “to make real-time AI animation, which could become a new form of theater.” Cea is already conversing with colleagues at Tufts’ Department of Theatre, Dance, and Performance Studies about the possibilities.
TTS will continue to work with the instructors of SMFA’s 3D and 4D courses, with a longer-term goal of eventually making the render farm available to the entire SMFA community.
