Animating by numbers: workflow issues in Shane Acker’s 9

I read an article by Matthew Teevan of Starz Animation. He gives a short overview of Starz Animation’s production process of Shaker’s Animation feature, 9. It looks at the challenges this production had to make, with a small budget and limited time frame. Even so the animation kept a high level of invention and creativity to address the issues without compromising the films quality. In this article it address some ways in which they approached problems and how they developed solutions. The article mainly looks at character design, art direction and workflow in the production pipeline. For a smaller studio, it offers insight on seeking a Hollywood style animated feature that can stand up against major studios such as Pixar and Dream works.

“Starz (now known as Arc Productions)  is based in Toronto, Canada, and employs around 275 people. We are now making our sixth feature film. What differentiates Starz from major studios like Pixar or Dreamworks is that we function as a ‘studio-for-hire’, so if a movie requires some computer-generated characters, for example, we might do them for a whole film, but we can essentially handle a film from development through to post-production. We operate in downtown Toronto amidst a lot of other animation studios. In this article, I want to talk about the making of an independent film and the workflow process that informs it.”


Starz Animation Studio – now typical in being configured as computer work stations, and populated by staff working on specific projects, and having a dedicated aspect of production to carry out in relation to an overall workflow pipeline. Screen Grab from Article

The animated short 9 created by Shane Acker was noticed by film director Jimmy Lemly. He is an American film and television producer based in Paris best known for his work on the hit action-thriller film Wanted. Lemly recognise the potential of 9 for a feature development. He turned to Director Tim Burton and Timur Bekmambetov to help raise funds for the production.


jim Lemley.jpg

Jim Lemley

 “They in turn took it to Focus Features, who are essentially an art-house live-action studio, but who seemed suitable in recognizing the art-house/commercial crossover of 9.They liked the quirky approach and styling of the film, and worked on it for a year and a half before bringing in Starz as a service house to work on some scenes already in production in Europe, which were falling a little behind schedule. We ended up taking on more work and ended up doing the whole film – some 1550 shots – even reworking some scenes that had already been completed, so that the aesthetic of the film remained consistent.” MATTHEW TEEVAN Starz Animation

Matthew Teevan talks about how they wanted to keep the design true to the original short film created by Shane Acker.

“Much of the initial design stage was based on the original visual concepts in the short film. We needed to visualize what this world was going to look like. These weird little puppet people were living in a post-apocalyptic environment, and though it was being done in CGI, it was very important to Acker that the whole piece was characterized by a very tactile feel, that it had a sense of ‘old school’ stop-motion, and that the movement in the environment responded to an authentically material world. Some of the technical aspects that had been acceptable in the short needed changing for the feature, so a lot of work had to be done to make the puppets functional, yet dress them in the same way as Acker required – the seemingly hand-made, amateur-looking ‘dressing’ had to hide a more sophisticated and robust ‘skeleton’. This was hard work as the characters are essentially a net, with a limited pattern, and this posed problems with moving them, so we needed an extensive pose panel or pose chart – initially of about 15–20 poses identified at the beginning of the project, but added to across the production – to see where the material would stretch, how we would achieve balance, etc. The pose charts also helped different animators achieve consistent movement or emotional gestures across scenes – if we achieved a great scene with ‘5’ looking sad, we would direct another animator to look at the pose chart on which it was based to imitate that emotional performance.”

As the characters are very similar they where able to use the on rig for half of the main characters. For they bigger characters they had to customise the rig to adapt, this also went for the smaller characters. This helped alot in keeping the budget as low as possible,  aswell as time limitations. They really wanted to get that ‘Focus’s usual independent art -house feel’ as it was not a high concept movie.

Screen Grab in Article 


Screen grab from article

As this was not a children’s story, but very dark and violent, it allowed them to be very creative, but still kept to the original story of Shane Arkers, wanting to get as much of his content on screen as possible with high quality.


A problem they faced was the characters design for animation. In the original short there was very little movement in the characters for them to animate. They had to put alot of attention into the mouth of the characters, aswell as the eyes. The irises would dilate and the mouth contracted and expanded, but there was little else.

“This was problematic because there were a lot of sophisticated emotional moments and a lot of backstory to convey. It was quite hard for the animators, therefore, to find the emotional performances in the characters, and this was crucial as the backstory has a lot about the characters caring for each other. They had to achieve a lot through body language and key poses, and a great deal through what could be suggested through the eyes and mouth. Once we got a turnaround, surfaced model to a certain level, for example, we would try and place it in a key environment, so we could judge how it could be lit, and how particular kinds of specularity in the ‘eye-cans’ might communicate something. Using a rough set-up was extremely useful in seeing how colour and lighting might communicate emotion. Also, this was important in helping to differentiate the characters, 6 and 8 particularly, who posed particular problems because they were essentially white. For example, 8 looked bright blue in some environments so we had to adjust him almost on a scene-by-scene basis.”

Another challenge they faced was the ensemble nature of the characters. Since there are essentially 9 main characters in this movie which where all on screen for alot of the time, they first thought they had to create all the characters before any work could be started.

“Usually, you might have one or two central characters that can be plotted through scenes, and a workflow allocated accordingly, but here we had a number of the characters on-screen most of the time, although, quite helpfully, we saw in the storyboard that as one gets discovered another one often gets killed !! So, we developed a workflow looking at when the fewest or the same characters were involved. We built 9 because he was in the first sequence and the only one in it, then we added 2 as he emerged, and so on, and this helped us create sequences. Because 5 looked a lot like 2, we could build him really quickly, and work on a range of sequences with those characters. Throughout, ‘animation’ was really chasing ‘rigging’ and it was a very intensive process. The beast characters were entirely different, and highly complicated, as well as time consuming. The ‘Seamstress’ and the ‘World-Building Machine’ were highly complex designs and rigs. They were the last characters completed. Basically, the workflow was created on the basis of how we could match technical requirements – effectively the number of built characters – to when they emerged in the story, saving as much time and money as we could. Though the film was, therefore, made out of sequence, it did not matter, as we had a nearly fully developed storyboard.”

Screen grab from article

One of the more complex characters in design and animation, the Seamstress, who captures characters in the sewn mesh of her own body.

For the character modellers, they recieved  orthographic views and technical guide information. This gave them another challenge as they had to think about how they where going to shoot the materiel:

“This threw up the issue of ‘scale’ as we had to put together elements that seemed to have different scale references. We had to decide how to actually shoot the material. We had to decide if we were shooting 10-inch tall characters, and photographing them with a heavy depth of field, which can be hard to watch. We decided upon a creative conceit in which we would pretend that there was a miniature cameraman, who was also 10 inches tall, and had a 10-inch tall camera, and the laws of physics, which would be applied to the photography, would be scaled down accordingly. So, essentially you become part of the scale of their world, and everything is in proportion. This becomes clear when you see some of the dead people lying around or compare the characters to other objects and bigger environments.”

Screen Grab from article

“In relation to a lot of design and layout much of this work is done in 2D – it is drawn, Photoshopped, painted – and we build as many of the assets in advance before we get into shot and sequence production. We decided to geometry-optimize as many of the assets as we could, which means that we could adjust the shape and view of something to make it look different, so that within any one shot we could save on rendering. There was so much rubble, for example, we merely rendered a few pieces then optimized placement and shooting issues. This meant going to layout as early as possible and using as much pre-visualization as possible to facilitate trajectories and rendered elements. If the camera was placed in a certain position we could judge very early what was going to be seen and what would be less seen, and plan what was required accordingly. Thankfully there is always something being knocked down or rebuilt in Toronto, so we could always go and photograph a lot of reference material. We had to do a burned-out car, too, so we repurposed some footage from a movie called Everyone’s Hero (Colin Brady, USA/Canada, 2006) which featured lots of 1930s cars. We retextured it, of course, and did a lot of visual distressing, and that served our purpose well, and it serves as a good example of using extant or previous acquisitions and assets. Reusing materials by turning it around and shooting from a different angle was also another strategy. The art department goes through every shot and creates a colour ‘thumbnail’ too, in order to establish colour tone, hue and saturation. All of this is considered in what we call ‘The Blue Book’ reading, which considers every shot from a technical point of view in relation to the assets, assembly and animation. We then allocate tasks, and define specific work, because in the time frame we had we did not want to create anything that was not actually, ultimately going to be on-screen”

Three Main Challenges

“There were three essential challenges. First, the character animation, especially in regard to the design and the physical storytelling. Second, the construction of this world, which though it had a homogeneity as a post-apocalyptic environment, there were lots of diverse places they had to go to – everything basically blew up, was burned down or destroyed by machines. Much of that was about looking at the frame, and deciding how it might look most effective, so we used a lot of matte painting. The characters remained in 3D, the foreground environment remained in 3D, but thereafter, it was often rough geometry, and matte painting on a variety of planes. There was the benefit of immediacy in this; for the most part we had a clear sense of what things were looking like. The third concern was the dark palette of the film, and whether this would work not only in the cinema, but when it translated to television. At one level we exploited the colour script to both hide some things and suggest others, and the darkness was ‘forgiving’ in some scenes with the large machines. A lot of work was done in post-production, particularly in colour correction, to anticipate what its readability would be in all formats, and to make sure it could be viewed easily. For a low-budget crossover movie, we believe we met the challenges, and created an unusual and appealing film.”


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s