Released last fall on MTV, Linkin Park’s music video, “A Light That Never Comes”, loosely tells the story of Judy, a young woman attempting to navigate her way through a CG landscape in which each Linkin Park band member, as well as collaborator DJ Steve Aoki, perform in their own, separate districts. Linkin Park’s DJ Joe Hahn served as video director for the nearly four-minute video, which was created by Los Angeles-based Ghost Town Media.
The project is one of many that Hahn and Ghost Town’s creative director Brandon Parvini have worked on together since 2009 when Hahn saw the video that Ghost Town made for Kanye West’s “Welcome to Heartbreak.” Relying primarily on Maxon’s Cinema 4D and After Effects, the duo share a “self-taught approach” to VFX, Parvini says. “We like to find creative ways for getting things done without getting mired in a big VFX process with a team of 30 people. With Cinema 4D I was able to get a few guys in here who have really good taste and get them up and running really fast.”
Physical Meets Digital
Hahn and Parvini have long discussed how they might explore ways to have live images exist in digital space. For this video, they came up with an overall look that conveys the core concepts of a remix, Parvini says, in that “you’re taking something from somewhere else and repurposing it.”
Parvini scanned all of the band members in various positions and at different angles for use as raw 3D assets. In addition to the static 3D scans, Parvini and Ghost Town devised a system to capture sequential 3D objects, allowing the band to perform and have the performances played back in post—similar to how you would normally play back video in a non-linear editor.
Early conception consisted of finding images they responded to and weaving them together loosely to get an understanding of the look they wanted. Next, they created a map showing each separate district with Aoki’s district in the center. Ghost Town brought on director and 3D and VFX artist Noah Rappaport (http://vimeo.com/noahrappaport/videos) to lead the team in charge of planning and developing the city that serves as the main setting for the video.
Wrangling a Monster
During the two months they worked on the video, the biggest challenge Ghost Town Media faced was the sheer expansiveness of the piece, Parvini says. Because CG is ubiquitous these days, it’s easy to take for granted what it takes to create an entire digital world. But for a small production house, the reality is “you’re wrangling a monster,” he explains.
“We had to build the whole place brick by brick, and having to make everything, gather all of the assets and drop them in all scenes was really challenging. I think at one point we had about 70,000 buildings in our city setup, which is around the same as Manhattan.” (Watch the process video here.)
It helped that they were able to use Cinema 4D’s MoGraph to clone many of the buildings they needed.
Combining stock models and custom-built building structures, Ghost Town used cloned instances in Cinema 4D to allow for quick population of buildings across specific segments of the city. Using the instance cloner system after the buildings were populated allowed Ghost Town to quickly adjust positions and final angles.
Boosting Render Times
Rendering was another major hurdle to overcome. Ghost Town used Dell Precision Workstations for the heavy rendering of the project, but even with their systems completely built out to handle the renders, GI rending was simply too cumbersome and slow. After struggling with workflow for a few weeks, Parvini opted to upgrade from Cinema 4D R14 to R15, even though he knew it was ill-advised to make that kind of change in the middle of a project.
Fortunately, the upgrade paid off. “Once we got our hands on R15’s new light caching system we saw dramatic improvements to the timeframes and render times, so we could use Global Illumination to get the look we wanted in a quarter of the time,” he says.