« VES 2006 Fact Sheet | Main | Pirates 2 all wrapped up? »

May 25, 2006

Melts in your Eyes!

June 2006 ANIMATION MAGAZINE www.animationmagazine.net

Tippett Studio creates warm and fuzzy chocolate characters for Milka commercials.
by Barbara Robertson

Why is chocolate so sweet and delicious? Because it’s made by cute marmots in the Austrian Alps. Ifyou didn’t know that, you haven’t been watching the Milka spots created by Tippett Studio for Frankfurt’s gilvy & Matherand its client, Kraft Foods. In a series of four 30-second spots directed by Frank Petzold, the crews at Tippett have mastered the art of chocolate delivery via marmot. Their marmot has transported the chocolate while riding on the back of a bear that slides down the mountain past a young couple on a hayride. He has flown in while hanging onto the feet of an eagle. He has hidden chocolate bunnies for Easter, and in a spot that will air at Christmas, placed chocolate Santas near a cabin door.

DOWNLOAD this Article - PDF

“It was a lot of fun working on these commercials because the characters were so happy—happy animals creating happy chocolate and finding great ways to deliver it,” says Will Groebe, lead animator. “But it’s very difficult, technical work. They’re always wearing a lot of stuff—hats and sacks of chocolate that they’re interacting with, and they’re interacting with each other. It wasn’t just ‘Bang out the animation.’ It was ‘Whoa! There’s a lot of stuff going on.’ Some shots have four characters. And, they’re furred.” Groebe helped design the animals, which needed to look real but also warm, fuzzy and cartoony, by painting on photographs in PhotoShop. Eyes grew bigger, the eagle’s serious expression changed into a smile, the bear’s teeth shrank and the marmot’s paws became hands.

Even so, the client continued to adjust the balance between realism and cartoon. Thus, knowing that the client might ask for shape changes even after the models were built, the team decided to add controls within the rigging that could scale facial features and body parts. An animator could grow limbs, broaden smiles, shrink ears and widen eyes interactively with the director, get approval on the look, and then give the other animators the new standard character. “By making it possible to do small adjustments in animation land, we didn’t have to send the characters all the way back to modeling and re-rigging,” says character setup Jeremie Talbot. “The rig allowed the animators to jiggle the model in ways they couldn’t with blend shapes.”

For the props and accessories, the modelers created separate pieces that easily attached to the characters because they incorporated the model parts. The marmot’s backpack, for example, which at tached to its spine and shoulders, had the marmot’s spine and shoulder modeled into it. The parts didn’t replace those already built into the character; the mirrored parts were constrained to the base model. “We had nodes in the backpack that attached to nodes in the character so the backpack would be attached directly,” says Talbot. “If the character moved his shoulder, the shoulder in the backpack would also move.” Similarly, when the backpack needed to move with the marmot’s hips, the riggers would attach it using the same technique. In addition, the animators could adjust the thickness and width of the backpack straps.

The backpack wasn’t the only plug and play prop; the crew also created flight helmet with floppy leather straps, a winter hat with earpieces, an investigator’s cap, a scarf for the bear and pouches filled with additional props: chocolate bars and bunnies. “Because we created all the props separately, the director could say that he wanted more chocolate in the backpack, which was often the case,” says Talbot, “and an animator could add as much as he wanted.” Animators could render the scenes without fur on the animals to check the animation and then hand the scenes to TDs (technical directors) who would add fur and lights. “Sometimes, once we’d see the fur, we’d see actions we didn’t see before,” Groebe says, “and we’d have to fix the animation.”

Tippett uses the proprietary fur tool called Furocious. In Maya, the TDs see guide hairs that they shape into a hairstyle. To control the characteristics of the fur—how scraggly it is, the color, how the colors change over the length of the fur and so forth—they used painted texture maps and parameters. As the fur tool runs, it applies these attributes to the thousands of hairs interpolated from the guide hairs that coat the character.

Lighting the fur was particularly critical and Tippett tested its global illumination (GI) toolset for the first time on these characters. On location, they took bracketed fisheye photographs of the environments and used them to build high definition range diffuse environment maps. “If a character is walking across a green meadow, we want the parts near the green to reflect the color in a broad, diffuse way,” says lead technical director Charles Rose. “Before, we did that using lights under the ground. We’d dial a color in and have it fall off as it got higher on a character. But those lights always come from specific positions and even though you may not be able to verbalize what looks wrong, you know it doesn’t look right.” With GI, the TDs apply diffuse light using data gathered by bouncing digital light rays into the entire environment. The lighters also used RenderMan’s deep shadows for the fur, which cause the light to fall off as it moves through the pelt. Calculating the specular light to high light the fur properly was a special challenge. Too much specular and the bear looked oily and the marmot looked sweaty. “If you want characters to look warm and fuzzy, you need the right specular,” says Rose. “It’s not hard to know where to place lights to get highlights on a hard shiny sur face, but because fur is just curves that simulate pieces of cylindrical hair, placing lights to get the specular right is not alwaysi ntuitive.” To test light placement, they would sometimes fly lights in a big circle around a character over a few hundred frames. Rather than simply using Maya’s dynamics on the guide hairs to make the fur look like it was blowing in the wind, the TDs baked out the dynamics pass and used it to generate animated texture maps that affected the entire body of hair. “It’s a slow and iterative process, but we got a better effect,” says Rose.

Animators did all the animation by hand except for such environmental effects as, in the first spot, particle-driven flower bits and pieces of grass tossed up as the bear slid downhill. “We didn’t have time to set up rigs for effects-driven animation,” Groebe says. On a larger show, we’d definitely want help, but because it’s such a fast turnaround, we were able to do some testing on our character rigs and the fur tool. There are a lot of really nice things that come up for short shows. We could do big, happy facial expressions and go overboard a bit on the animation, which we don’t always get to do. It was fun stuff.” Ssssssweeeet, indeed.

Barbara Robertson is an award-winning journalist who specializes in visual effects and computer animation. If you have any cool tips for her, you can e-mail her at brobertson@animationmagazine.net.


Posted by dschnee at May 25, 2006 11:13 AM