Rock in the Road
CGSociety :: Production Focus
6 November 2012, by Paul Hellard
In the fall of 2007 a small group of animation students at the Southern Adventist University in Collegedale, Tennessee worked to design characters and develop a story about a young boy struggling to move his cart around a roadblock. This was a simple story that grew larger than the tenure of most students working on it. After the narrative led to a 20 minute animatic, the edits brought it down to ten minutes. The following summer, more classmates joined the original students to finish developing the story, create assets for the production, and start animating the film. At the end of this last school year, they finished a five-year journey taken on as an animation program. The result is the twelve minute short animated film Rock in the Road.
The scope of the project grew as the team became interested in telling a longer form story. But as students who started it graduated, and other projects were begun, finishing the short became its own rock in the road. However, professors and students persevered, realizing the completion of the film would benefit everyone who was involved. A small, dedicated crew worked on it between other projects to finish the film at the end of the 2012 school year.
Emphasis is placed on developing talent at the School of Visual Art & Design, Southern Adventist University for the purpose of benefiting others through service. Character animation students are encouraged to tell narrative stories that reflect truths and concepts that focus on external means of change. Under the guidance of Disney veteran Hendel Butoy, instruction is focused on collaborative production where a studio-like environment is established during the third and fourth years of study.
The Struggle to Finish
Meanwhile, lessons learned from Rock in the Road were implemented into new films. Instead of cutting their losses, dumping the project and focusing on the new projects, a small team decided that finishing this film was important. “Finishing was challenging due to student turnover,” explains Zach Gray, one of the many students who took up the challenge. “There were so many great folks who worked on the project graduated and had to move on. And upcoming students developed their own projects and couldn’t commit time to the old one.”“We were staring at this huge mountain of work, but students Melissa Caldwell, Danny Cooper, and Yannick Amegan all kept pushing for another two years. Just like the kid in the story, they moved this giant obstacle by thinking of the benefit it would bring to others,” explains Zach Gray. “I’m personally so grateful to each of them for what they did.”
People
Melissa Caldwell is a graduating senior, focusing on simulation and rigging. She just completed a two-month internship at Disney Animation Studios in California. Melissa did almost all of the simulation, tech animation, caching, and scene assembly for the 120 shots in Rock in the Road. She comped the shots in NUKE and also oversaw the rigging for two other short films in production.
Daniel Cooper was a recent graduate of the program and is teaching the freshman animation courses this year. Daniel took over creative direction on Rock in the Road in 2010 and wrote and performed the score. He was instrumental in the animated short Danger Planet, and recently directed and completed animation on Gone Bananas.
Aaron Adams is graduate of the program turned instructor, then department coordinator, Aaron was instrumental in the growth and refinement of the animation program. Aaron has the artistic chops and the technical skill to back it up. After completing graduate work in animation at Savannah College of Art and Design, he coordinated production on Rock in the Road until he accepted a position at Disney Animation Studios in California.
As a student at Southern Adventist University, Zach Gray shared his passion for animation with administration who in turn supported the formation of the animation program. After helping start the program, he completed graduate work in animation at Savannah College of Art and Design. He directed a live action feature, and then took over coordinating production on Rock in the Road after Aaron Adams transitioned to Disney.
Hendel Butoy is a well-known animator, returning to the Adventist University in 2003, when the standards of quality in story and animation took an exponential leap. He studied animation at the California Institute of the Arts before accepting a position as an animator and director at the Walt Disney Company in 1979. He was employed at Disney for more than 20 years, working on such films as Fantasia 2000, Rescuers Down Under, and The Fox and the Hound.
Jesse Rademacher returned to teach rigging and production classes after working at ILM, Visual Concepts, and Raven Software. He’d previously graduated from Southern Adventist University then completed graduate work at Savannah College of Art and Design.
Story
The story is based on a fable the crew found called the Stone in the Road. While there are a lot of different versions, basically the king put a stone in the road. Sometimes it was a small stone and sometimes it was a big stone. There was either money underneath the stone and when someone finally moved the stone, the King said, “Hey, here’s your reward. Take this money!”, or the king was hiding in the bushes and when someone moved the rock, he sprung out “Surprise! You are the winner!” That was the original telling, and there was a greater value than money in the boy’s character. And what did the king really want? Did he want just one person or did he want to change the hearts of the kingdom? So the Aaron changed it in that way, as well as making the rock two stories tall.
“I didn’t want it to be a little idea,” noted Aaron. “We needed a big idea to get the creative engine going.”
The School of Visual Art and Design at Southern Adventist University is inspired by a studio model. For the first two years, its instruction, for the last two years it changes to production. “If we’re going to take that seriously, we’d better be doing something that is going to stretch us,” explains Zach. “What that meant at the time was doing something epic – really long, and much more complicated than what we had attempted in the past. It became epic, via the emotional spectacle that they presented rather than a grand scale. Without everyone living through that experience, it’s almost impossible to understand.”
Stylistic Influences
Aaron Adams had this India fetish, and wanted to set it in India. Since then, a lot of shorts have chosen India for their influence. They wanted to create a fictitious kingdom, more fantastical, drawing on an Indian style heritage; a foreign yet familiar place. “Even down to the design of the fruit, we wanted to make sure it wasn’t an existing fruit, but a mystery kingdom fruit,” Daniel Cooper pipes in. “We didn’t want to tie it to anything, but it was influenced by what you would find in India.”The class also wanted the simplicity and exaggeration of the characters in Iron Giant, a nod to Brad Bird’s 1999 animated feature, including a classic Disney slant with Jungle Book look. “Big ears, the big eyes, exaggerated features,” Daniel continues.
“One fun thing we did with the design of the main characters is ask what is your power animal,” explains Aaron. “The wise man is a monkey, the giant is the elephant, and the nobleman is a snake. We try to think about how their design and the animation would help complement that.”
Environments
When the crew looked at the position that Disney’s representatives suggested, there were a few ideas, but most of the work had gone into the character design and the story. Zach found that it was more about how to make the backgrounds live up to what had been established with the characters. “The environments got more complicated than any of us planned,” he says. “We did tests that took a more simple approach, but the characters didn’t really fit. We landed where we did when we applied the style guide for the characters to the environment.”“We thought we were taking a short cut by skimping on really detailed development artwork for the environment, but that ended up being the long way around,” he continues. “There was a production decision made early on to make this more producible by putting it in one set, staging it alongside a rock wall. How hard can that be? Well, it’s actually really difficult when you are trying to make a rock wall look interesting over the course of the short. It’s missing those big abstract shapes to lead the eye. In some way, it would have been easier if it had been in different sets and dressing to the shot instead of making one set hold up for that long. Ultimately, we decided it would be easier to make one set for everything instead of a making a high res set and a low res set. Thankfully V-Ray could handle anything we threw at it.”
It took months of after-hours work to figure out the look and get renders on the screen. But even with simple appeal and simple characters, production has a way of introducing complexity. There is so much shorthand is involved in making things appear simple.
Involvement as an Instructor
“Aside from Hendel Butoy, there weren’t industry veterans on the crew, so learning how to make this work was really a ‘grass-roots’ effort for the University team. They knew where they wanted to go. “We had CG generalists, and in story and animation, Hendel is a master,” said Zach. “We knew we wanted to make meaningful short films together, and follow more of a studio model than a classroom model, but it took some time to figure out what that looked like.”“We owe a lot of that to Jim Turner coming in and saying that this deserves to be bigger and better than what it is,” says Aaron. Jim Turner graduated in 2009, and is now working at Disney. “We did a story reboot. We had such talented group people coming up through the ranks; I knew that if we could give them something to sink their teeth into they could just go nuts. We had such great animators, modelers, and designers. All the talent kept compounding on itself as new students came. If they just got through this as if it were for a class, it wouldn’t be as rewarding an experience for them.”
Pipeline:
The most important thing Zach, Aaron and the crew learned from this was finding the emotional core of the story and developing character. The other thing that it allowed them to do was build out their pipeline and figure out how to make animated films. “As the scale of a project increases, 50% of the work is doing the art and the other 50% is in the art of production,” says Zach. “That is, building the machine that enables you to do the art.”Aaron’s initial goal was production organization. Getting departments settled, getting people functioning in terms of the work that needed to pass through their hands, getting the asset to the next department, and scheduling. “It started out as an Excel spreadsheet, and that was kind of working, but Shotgun was the real turning point for me,” says Aaron. “Getting the data where you could really visualize what needed to happen with it.”
“Without Shotgun, the handoff would have been really difficult,” agrees Zach. “Instead, someone could be up to speed in an afternoon and know what they needed to work on. This may seem like a small thing, but Shotgun gave us a picture of how far we had left to go. Without seeing those numbers slowly tick down and get closer to 100%, I don’t know if we would have had the energy to keep going.”
Melissa found herself staring at this huge mountain of 120 shots at one stage. “They all needed simulation of cloth and hair, animation fixes, and geometry caches. Some animation had been done with different rigs and there were one off simulation rigs for a few shots, not to mention all the junk the animators left in the scenes,” says Caldwell. “I had to get that all ready for lighting. My workflow was figuring out how to connect what the animator had done with what lighting needed. I hadn’t touched cloth or tech anim before I came on the project. It was a daunting task, but I just took it one at a time. If that meant working two weeks on a shot, so be it. I just tried to make it as pretty as I could.”
“Shot production in 3D, from my perspective is just a nightmare no matter what,” says Aaron. “You are working to get the data actually clean to pass on. Since folks rolled off the project, you couldn’t go have them make the adjustments. It’s like a safari, you have no idea what you might find when you open the next shot. Simulation and cleanup just takes time – a lot of time. Let alone the lighting and rendering.”
Rock in the Road was Daniel’s first project. Starting as a freshman, the last thing he did before graduating was the sound and music. “For me it was seeing the potential in what it could be if it got done. I got to work quite a bit with the Steve character. He’s the curly redhead carrying the pots. He’s the common man that the story needed. You had these larger than life characters in the old man, nobleman, and giant, but we needed someone like us in there. He’s just like the boy really. Steve represented what the boy could have become, but Steve also represents us, the audience.
“When you learn animation, you start with body mechanics, and there is plenty of that in the film,” explains Daniel. “But there are a couple of sequences that hinge on the inner life of the character and internal change. One of those moments is when the boy decides to help move the rock. We had a really gifted and experienced animator, Robin George work on that shot.” George graduated in 2008 and works at R&H now. That shot was about acting, not just the principles of movement in animation. “So we worked to cast the animators in the right way as they gained experience,” adds Dan.
Finishing
While the story fluctuated in size, the initial idea was three minutes, and then it became twenty minutes. Taking all that information and turning it into something more manageable was challenging. There came a point when animation had to start, and it went one sequence at a time. The major plot points were settled, yet some of the telling wasn’t ironed out until the very end. Animating took four years, finaling took two and a half years, and rendering took about a year.< br>Aaron did the research and development on the crowd system. The key to cracking it was to build 15 or 20 auto-rigged characters, build cycles, and randomize their shading. The only way to get that to render was with the V-Ray proxy system. V-Ray was a life saver for the project. We had enough complexity that we couldn’t introduce a pass system on top of the all that complexity.
Most technical aspect of the production
Aaron and Chris Wombold (graduated 2010, Nickelodeon) built a custom rigger and unified mesh crowd system. From that, they generated a series of animation rigs. They built cycles in their own scene files, exported those to V-Ray proxies. And created a crowd of 1,500 people in one shot, and they were able to render in a single pass using V-Ray proxies. “That led to a complication with the architecture of Maya,” explains Melissa. “We cached all the animated geometry out using Maya caches, and the system that we wrote split out a ton of files. We kept running into issues when we hit 256 caches, and there were many shots with way more than that, especially for the fruit.”Zach takes up the story. “We ran into a ceiling with too many open file handles per process between the caches and the mipmapped exr’s. The V-Ray developers helped us solve that by figuring out how to raise the limit in their plugin, and that carried over and solved the cache problem as well. Because of the turnover in artists and because we didn’t have an automated asset system at the time, we opted to render everything in a single pass. We took a hit on render time, but that meant all we needed to do for comping was put in a couple of backgrounds and apply the vector blur. Yannick Amegan (2011 graduate working at Magnetic Studios) lit the entire thing and Melissa finished the comping in Nuke.”
“I’d say that our tangent sculpt system for cleanup was pretty unique. It was originally built by Michael Hutchinson (2006 graduate working at Dreamworks). That tool allows us to put a corrective blend shape at the end of the deformer stack, but manipulate that target in world space. I can’t say enough good things about V-Ray either. I wrote a submission front end for the standalone version that interfaced with Qube! That let us just send a .vrscene to the farm. The memory management is great and the proxy system allows you to render massive scenes.