What are the differences between the visual effects work that ILM is known for and making an animated movie?
GEOFF CAMPBELL: Well we thought the main difference with a hundred plus characters would be lighter meshes there was also a perception that animated characters should be lighter and cleaner because they're more stylized. It actually turned out they needed to be higher res than we originally planned.
Gore had us reference films like The Good, The Bad and The Ugly and The Man with No Name, and Crash brought up to ILM a wardrobe of old western-style clothing complete with sweat and salt stains under the arm pits and threadbare shirts and pants.
Although these characters were more stylized than the realism we had achieved on the Pirates of the Caribbean films, our approach to modeling was pretty similar. For example, we even shared between the characters a parts library of worn threads that we could apply to the edges of coats to break profile, along with ZBrush displacements and other tricks. The attention to photo-realism; hair, fur, feathers and bad skin was the bar, but it all had to work within the context of the characters.
Matching realism on effects films mostly comes down to precisely matching the reference of the actor both for modeling and facial work starting from scanned data of the actors. The characters in the town of Dirt were looser and a lot more fun because we were sculpting them to match Crash's artwork and not actors, with a few exceptions including Doc, played by Stephen Root. In his case, the artwork was designed with Stephen in mind, and Gore asked us to put a little more Root in the sculpt.
What are the differences between the visual effects work that ILM is known for and making an animated movie?
MICHAEL KOPERWAS: While we have lots of experience putting characters into real scenes, Rango represented a new frontier for us in terms of character and animation, and our facial pipeline had to fit. The 120-odd characters all needed some level of articulation and a whole lot of expression.
When we were initially planning the show, there were characters expected to be considered 'easy builds' because they only had one line, even though it happened to be full-screen. For something like that to be feasible in our budget, we had to adjust our facial animation system in a few ways. The goals of an animation system are to be easy to use, and easy to author for. FACS is a well-known facial system, but it's designed to help identify what muscles are used in a facial expression.
What we've done in our proprietary system, Fez, is flip the problem around and think about how someone intuitively thinks they should animate a face. Our modelers create animatable poses that move the face in ways an animator can intuitively interact with and that's standardized across characters. That allows us to transfer animation easily, but mostly makes it easier for animators to jump in and be familiar with the setup right away.
How were the characters in the movie inspired? How did they evolve from concept art to finished figure?
GEOFF CAMPBELL: Crash's artwork had a lot of humor in each drawing but also brought out a sense of desperation with each character in a town where the water had run dry. You could see in their eyes, the loss of their crops and livelihoods. You could virtually feel the dryness in their throats. The question would often come up as to what type of animal a particular character was but Crash would tell us not to worry about species but to stick with matching the 'character' of the artwork. That was an important part of keeping each unique and intriguing. You might not know or care what kind of lizard Bad Bill was but you'd remember the menacing look in his eye. In a way, we had to switch off the analytical side of our brains that comes with doing effects work and search out the qualities and character traits in the 2D artwork.
Crash and the ILM Art Department did a great job providing the DMS and look development artists with detailed references of everything from the wear and tear on the clothing to the types of pocket watches, ailing skin, fur, eye glasses, hair styles, guns of the old west etc. The research was incredible.
The general flow was to sculpt and texture a three-day ZBrush maquette for each character and after approval, start in on the asset. A first pass model was sent to rigging while the model asset continued to be refined, hair and fur splines added, scales added, UV's laid out etc. Once the finished model had gone to paint, the modeler would start on the facial library. We had a few weeks for each character model including wardrobe and two weeks on average for making our facial libraries.
Rango and Beans were special cases because they continued in a design phase throughout pre-production. Beans had an original design but Gore, Crash and Jim Byrkit (who was head of story) still didn't feel they had found their Beans. The design needed more exploration. I made a number of trips to LA to sit with them and do quick ZBrush sculpts based on their drawings. In Crash's original artwork she was pretty sullen and unattractive but we were searching for some of that Margarita Lozano sex appeal -- without going too pretty. She was already in shots when Gore had us add a gap in her teeth to offset her smile.
)
)
)
