DIRTGIRL’S WORLD

When they were ready to start producing ‘dirtgirlworld’ in 2008, series creators Cate McQuillen and Hewey Eustace were keen to design the main character Dirtgirl to be able to reach out and engage a young audience aged four to seven well enough to deliver the show’s environmental message. After seeing Hackett Films’ work on a music video for The Dissociatives using a composited, photomontage style mixing live action with animation, they thought the approach would be worth developing for ‘dirtgirlworld’.

Yanick Character Building
Production on the first series got underway at Hackett Films in May 2008. Not long before, the studio had also completed a pitch for a Chiquita Banana spot in the US. This pitch wasn’t successful, but the work they had done included extensive environmental designs, trees and plants laid out in 3D space in After Effects, which formed the starting point for dirtgirlworld’s environments.

Animation director Jean Camden described the production workflow and pipeline. “Dirtgirl and her companion Scrapboy are each composed of a live action body with live action eyes and mouth composited onto a large illustrated, animated head. Building the characters involves a live action shoot on green screen. Both Dirtgirl's and Scrapboy's bodies are shot separately from their heads. The head shoots have to be done one character at a time with the actor strapped into a chair to minimise movement. Then for the body shoot, also against green screen, Dirtgirl and Scrapboy both wear oversized boots and gloves to keep the characters' proportions looking cartoony, and balance with the large heads.”

The main reason for shooting the heads and bodies separately is the need to keep the eyes and mouth at a high enough resolution to enlarge and composite effectively into the heads. Therefore, the facial performances had to be shot as extreme close ups. Also, during the head shoots, the bodies had to be kept still and facing forward. “It was important to prevent the actors from moving and bobbing their heads around too much, and to avoid turning their heads too far around because of the very wide set of their eyes in the animated heads,” said Jean. “Their eyes simply slip off their faces if they exceed a three-quarter turn.”

Heads and Bodies
The actors rehearse and know the scenes well enough to give facial and body performances that align well. Also, as Jean directs the body shot, she has the head take, always done first, playing on set on a laptop. The actors can hear their own voices while performing, which helps with timing, and Jean can use the performance to prompt their actions. The body performance as well needs to remain facing forward as much as possible to blend correctly with the heads.

Once the performances are complete and the backgrounds are designed, compositing can begin. All compositing is performed in After Effects as well as tracking heads onto the bodies. “A principle concern has been the eyelines of the actors. Even when all four characters are to be in the one shot, each facial shoot still has to be done separately,” Jean explained.

“We had no reference points for the actors to use until we devised charts based on the storyboards to place in front of the actor. These show numbers representing the other characters in relative positions corresponding to where they will appear in the finished scene. As the cast memorise their lines, they also memorise which number they should look toward at certain moments in the script.”
Storyboarding is fairly extensive and the boards are followed closely. They used some animatics at the beginning but once they were familiar with the pacing and the time the actors need to deliver the dialogue, they weren’t as useful as boards. The crew shot the live action on Panasonic HPX3000 cameras with two AV Intra cards per shoot, which they would swap over when one became full, and give to the editors to download.

Photomontage
The effervescent jumble of 3D CG and 2D props, creatures, plants and environments for this series is no accident. The photomontage style had appealed to Cate and Hewey from the start. “We had a great deal of reference to begin with,” Jean said. “A lot of photos from Cate and Hewey's property helped us to grasp the overall feel, as well as all sorts of photographic vegetable and plant references. It was very important for the plants and veggies to be identifiable. Cate had also collected some fantastic references from treehouses and junkpiles to vegetable gardens, and kooky shaped beds made out of twisted branches.

“The majority of elements were created using Photoshop with a combination of photographic textures and vector illustration, always simplifying the form of the object. These 2D elements were then laid out in After Effects in 3D space, where most were assigned to always point at the camera, continuing to appear as though they were 3D and not turning to pieces of paper when the camera panned through a scene. Designing many elements with a cylindrical shape helped this illusion.
“Some props were built within After Effects’ 3D space, such as the chook wagon and Dirtgirl's house. At the storyboard stage we were able to identify whether a prop needed to be built in 3D in Maya, for example when Ken, our 3D character, needed to hold it or if it needed to spin around.”
The scenes are full of life and motion, largely due to the different z-depth each asset has in After Effects. Even a gentle camera drift can make a scene come to life. Jean wanted to give the environment an Australian look by maintaining a subdued colour palette. Once you enter the veggie patch it brightens a little, but overall they aimed to use desaturated colours and avoided rich blue tones.

Textures and Techniques
“The photomontage environment and look of the show, using a variety of textures alongside flat colour, also allowed us to build our creatures and characters in various ways,” said Jean. “Dirtgirl and Scrapboy have 3D CG heads modelled and animated in Maya. Grubby the grub has a completely 2D body made in After Effects with a live action face. Ken’s body is modelled, rigged and animated in Maya, and the chooks have bodies from Photoshop rigged and animated in After Effects. This blend of techniques also helps keep it a secret whether Dirtgirl's body is real or animated – which we like.”

Dirtgirl’s tractor was represented on set by a green prop. Any parts that she needed to handle such as the steering wheel and gear levers was built and painted, but everything else could be composited in as a 3D element. The same is true of all props that actors interact with. They had to be built, painted ready to remain in the shots, scaled up to match the characters’ oversized heads and hands. During storyboarding they tried to limit the need to hold objects. All such props needed to be photographed on green screen at different angles, to composite into neighbouring shots for continuity.

At the time they started shooting, a portion of the environment still had to be built to accommodate the new episodes. But by now, having completed the first series, they have gathered a large number of assets and references for Dirtgirl’s world. Also, keeping the same actors has been important because the characters' faces would change completely with different actors.

The animators are based at Spin FX in Canada with their own director, Hana Kukal, while Jean was working in Sydney with the live action team, carrying out the Photoshop work, rigging, building and setting up all characters and props ready to animate. They also laid out all the assets in After Effects, ready to go. Jean and Cate oversaw the animators’ work remotely by communicating with Hana via Cinesync software, in which all parties can watch the same QuickTime together and draw on top of it to share notes and ideas.
www.mememe.com.auwww.decode-ent.com
www.hackettfilms.com