“Of course, this was not a supernatural story,” said John. “The book described the wings as a type of cloth that looked and felt like cool fire. So a realistic but stylised look was decided on. We determined its colour and internal movement pretty readily - but how to give it a wing shape?” The size of the licks, flow, shape and direction were harder to settle on, requiring many versions to look effective at all angles, close up and at a distance, while flowing at speed behind the chariot. The apparent speed varied quite a bit depending on the camera angle, and affected continuity.

Katniss’ flaming dress used similar-looking fire. But these flames had to emerge from the small folds of the cloth as she spins around, amid sparks. The dress and its complex in-curves had to be completely tracked to determine precisely where the folds were located in the plate. Then they used this digital, tracked shape to plot just where and how the dress should emit flames in all spots in 3D space at any moment.

Adding Drama
Previs was extremely helpful as a starting point for such shots. However, while previs reveals the elements and basic angle for the live action, because so much of this environment was CG, the effects team often needed to ‘cheat’ the angle or expand on it to add drama, reveal the actual scope of the environment and tell the story of the sequence.

Similar techniques were also used in the Muttations chase, which takes place at night. The plates were shot from a cart with the camera attached, holding it about 2 metres from the ground, pointing down at where the Mutts would be running. This cart would move through the forest, capturing the background footage. However, using this camera angle would only have shown the creatures’ backs and tops of their heads. What needed to be at the centre of the frame and the audience’s attention were their awesome jaws and teeth to create one of the film’s most frightening sequences.

The team at RSP tracked the plate from the high angle, then re-created a digital environment from that track with the shapes of the main trees and ground plane, and projected the image from the camera angle back onto the geometry, while moving the camera down eye-to-eye with the Mutts. They could paint and add in more trees later on. The result would be the same basic shot as the previs, through the new camera angle. Completing this shot in particular involved close collaboration with the director.

Pack of Danger
The team built one lead Muttation, focusing on the size, shape and muscle structure. From this they produced the pack of six with colour and fur variations. Due to the story, the design process explored exactly how ‘human’ these creatures should be. In the book, they each represent one of the dead Tributes and have corresponding traits such as eye colour. Consequently, at one point during development they had much longer legs and an almost monkey-like design nearly able to run upright. But the book also describes them as wolf-like, prompting other versions closer to dogs.

Interestingly, in the Games Control Centre where every moment of the Games is watched and manipulated, we first see one of the Mutts in quite a different form, standing in a miniature wireframe version on the Controllers’ revolving maps of the Games terrain as they plot when and where to send the pack out. This form had its own challenges for the animators.

The team needed to convey an alert, alive, very dangerous creature, but in a wireframe model the size of a puppy. As they experimented with moves, they found that this small Mutt often looked ‘cute’, pawing the ground, and while its facial performance was right, it was difficult to read the wireframe features.

Out of Nowhere
Out in the forest, the sequence is very fast moving and frantically paced as the Mutts race through the dark trees to the Cornucopia structure to fight Cato, the strongest, most ruthless of the Tributes. For the sake of drama, the production felt that to stop at any point and verbally explain exactly who the Mutts are would break the momentum of the action. They chose less human looks, but wanted them to look capable of human thought.

Their faces especially were designed to resemble human faces in profile without the pronounced muzzle of a dog. The neck and shoulders were also jointed in a more human fashion. Boxers were a strong influence, instead of the typical Rottweiler attack-dog build. “We wanted them to be agile and light on their feet. VFX animation consultant Daniel Jeannette worked on the speed and motion of their performance as a principle way to express the danger and power they represent,” John said.

“He went through the sequence shot for shot deciding how to add conflict and build the threat. For example, when Katniss and Peeta scramble onto the Cornucopia with the dogs snapping up towards camera, that camera moves closer with each progressive shot. We had the lights pop on their teeth as they move into pools of light, as if out of nowhere.”  www.rsp.com.au

Words: Adriene Hurst
Images: Courtesy of Lionsgate Entertainment