Efficient animation production for TV shows, part 1

Guideline documents, pipeline diagrams and schedules

Guideline documents, pipeline diagrams and schedules

One of our previous animation projects demanded that the animators hit a quota of 40 seconds per week. In other words each Friday they were expected to produce 40 seconds of finished animation. Despite this being a fairly standard TV quota, it’s still a very tall order. The curious thing was that no other team was ruled by a quota. The modellers didn’t have to make ‘x’ number of meshes per week, and the producers didn’t have to write ‘y’ number of lists. (Or, actually, maybe we did, come to think of it).

Anyway, what must be implemented to prepare an animation team for a 40-second quota? Naturally, you need to have a ship-shape animation team, good leads and contingency in the form of floating resources or pickup animators. Coffee, headphones and peace-and-quiet are also prerequisites. Your animators need clear briefs and a ‘nearby’ Director to make decisions because there’s not enough time to get things wrong; to try things out and fail. Welcome to TV animation, people say.

The point is, though, that tinkering with the animation team will only get you so far. A demanding animation quota needs wider production changes. Luckily, I had great animation leads who understood this. If you want us to make 40-seconds per week, they said, then we need to be fed in a very specific way. So, collectively, Kevin McDade, Martyn Gutteridge, Navis Binu and I wrote a production document called “Efficient Animation Production”. Here’s a summary of some tips to hitting a high animation quota:

Script writers

Long and lingering establishing shots.
You can easily knock off a minute of animation time by simply have long establishing shots, slow pans, long cross-dissolves and any ‘filling’ other than character animation. Consider starting the dialog before we see the characters, say, while we’re slowly trucking into an establishing shot, for example.

Reduce the number of characters in a scene.
The number of characters in a scene/shot is the largest determining factor in fast animation production. They can take a long time to animate and scenes can become complicated for the animator. Logistically, too, they involve larger amounts of data and complicated downstream headaches for the rest of the production team. The files take longer to open, share, reference, render or archive. Where you can, cut the number of characters in a scene or shot.

Be clever with crowd scenes.
First rule about crowd scenes: avoid them if you can. Crowd scenes are particularly slow to do due to all the animation involved. The animator’s also constantly checking for intersections. They take longer to review and often contain errors when paused and reviewed as still images. Animators will waste a lot of time with crowd scenes.

Second rule: if you must do a crowd scene then cheat like mad. Use instances, crowd scripts, 3rd-party plugins, standalone software, talented TDs, compositing tricks, etc.

Exploit motion paths
If your show allows the characters to travel on motion paths, such as in a car, plane or via superpowers then use them. Characters chatting in a plane are fast and easy to animate.

Dialogue rather than action
It’s a useful rule of thumb to note that dialogue is much faster to animate than action. Great script writers balance the two.

Composition and storyboard

Embrace Virtual Production
The animation industry has recognised that merging storyboarding with layout is a time-saver. Virtual production - as it’s called - means you can see your 3D storyboard very early in the animation process. Look at any “making of” documentary from any recent animated feature. I’m sure you’ll see the animators act out scenes for the benefit of a mocap system and a real-time renderer. They’re capturing the Director’s instructions directly in 3D and in real-time. Virtual production can be a great tool for project efficiency and, crucially, it benefits hugely the creative story-telling process.

Locomotion
Making a character walk is time-consuming. So avoid ‘unnecessary’ locomotion such as walking characters in and out of shot. Frame a walking character from the waist up if possible, then the animator doesn’t have to animate the feet or worry about ground contacts.

Look at “Mickey Mouse Clubhouse”. Note how rarely characters walk into shot. Conversely, look how often characters are introduced by a camera move which reframes a scene. I’m willing to bet that this technique is all about reducing animation workload.

Reactions and groups
If one character is delivering dialogue then everyone else in shot has to react and act accordingly. If you can frame it so only the talker is in shot then it’s much quicker to animate.

Character design

Character rigs
Rigs should have a minimal number of controls. On one of our projects a character’s tail had about 10 controls on it. The animators had to spend ages posing this tail in every shot. With hindsight, this should, ideally, have be reduced to around 3 controls.

Two legs are better than four
Quadrupeds take longer to animate than bipeds. They’re also longer to model, texture and rig – but not by much.

Fingers
Less fingers = less animation time. Simple.

Secondary animation
Long hair, clothing, loose items and anything else needing secondary animation should be kept to a minimum.

Better yet: employ some dynamics to help with secondary animation. Can that character’s pony tail be simulated rather than animated? But, please, if you are going to put dynamics or simulation in a character rig make it bullet-proof. Set it up so under-pressure animators can use it easily, and someone else can debug it in your absence

Animation

Animation libraries
If you’re lucky enough to have a luxurious animation pre-production phase then set up pose libraries for your characters. We used a Maya system called “Studio Library”. These libraries can be accessed and used to shorten production times. As production rolls forward encourage your animators to save useful poses and clips to the library so that others can access them.

Reuse
I guess that under-pressure animators will already do this, but wherever you can, reuse animation from another episode. Professional pride may demand that you tweak it a wee bit but lets not be silly. There’s no need to make a fresh walk cycle for episode 17 - grab the cycle from episode 3.

Avoid realism and the expectation of realism
So long as your show’s world is set up to be stylised then you can get away with ‘cartoon physics’ such as pulling objects from behind the character’s back and cheating with gravity, scale or speed. A pursuit of realism is one way to bog down animation. Similarly, an exaggerated animation style will help with this.

Lip sync
Same goes for lip sync. Establish a fast and readable style with strong shapes rather than a pursuit of realism. On one project we used FaceFX inside the Unreal games engine. That was very interesting in that a novice animator lip-synched the entire 11-minute episode in 4 days, but its dislocation from the rest of the character’s animation was judged to be a mistake in hindsight.

Production dept.

The Hit-By-A-Bus test
Imagine your lead rigger gets hit by a bus. Can the rest of your technical team debug his rigs? What about if the Hit-By-A-Bus test was applied to your main modeller? Could others find her models, understand her naming convention, locate textures and carry on as if nothing had happened? The Hit-By-A-Bus test leads you to do many things on a production, all of them commendable: naming conventions, technical standards, show-n-tells, knowledge sharing, contingency planning, etc. Put these in place and you’re more likely to hit quota and deadlines.

Visibility
Visual projects - such as animating a cartoon show - benefit from artists sharing their work. Everything should be pinned on the wall or shared via ‘dailies’. It helps everyone keep abreast of what’s changing and how the show’s evolving. Discussions are prompted, secrets shared and tips swapped. Never underestimate how easy it is for an animator to become secluded in her own silo of information. Everyone benefits by seeing what’s going on in the other silos.

Producer ‘noise’
Some projects can suffer from a lot of Producer ‘noise’ which means that the animators are also expected to wrestle with change-lists, memos, notes and opinions. In the past we’ve lost time collating and reacting to notes. We’ve also lost time tracking down and discussing contradictions within the notes. An efficient and timely producer feedback mechanism saves a lot of time.

Leave the animators to it
Tempting as it is to scrutinise the minutiae, it slows everything down. Clear animation briefs, regular human-to-human contact from the Director and a limited but comprehensive review process is ideal.

Priority and severity
Fast animation demands that everything is reviewed in the context of importance. Consider a TV show intended for 6-year-olds. If an animation problem incurs the wrath of 6-year-olds AND a broadcaster’s lawyers then it’s high priority. If an animation sparks an intellectual debate between unaffected adults, or causes an argument about an intangible concept or constructed abstraction, then it’s low priority. Review your animation through the eyes of the audience. This doesn’t mean that 6-year-olds aren’t demanding nor that their scrutiny will be any less incisive, it’s just that it’ll focus your limited resources on fixing the most important bits of your show.