Tiger & Bunny: Criminally Bad CG or Heroic Effort?

Tiger and Bunny is generating a lot of buzz this season for its unique Japanese spin on the superhero genre that we’ve grown to take for granted in the west, as well as for the schizophrenic condemnation and approval of consumer culture. But there’s something else more garish than the Soft Bank logo splayed across the eponymous protagonist’s chest. Amidst the heated conversation, we’re ignoring the cell-shaded elephant in the living room. Because let’s face it—for everything it does right, Tiger and Bunny’s CG parts stand out like, well, a guy in an a cheap CG mecha suit.

Not everyone can pull this look off, especially not Wild Tiger.

But I’m not here to knock on an otherwise entertaining show for aesthetic compromises made to compensate for shrinking budgets and talent pools. Rather, we'll use Tiger and Bunny as a case study to explore exactly why it is that CG in 2D anime produces such a knee-jerk reaction, even when done well.

To cut to the chase, it’s a problem of integration. For reasons I'll expand on below, 2D and 3D elements have a hell of a time doing the dance without stepping on each others feet. When handled correctly, the results are a masterful mix indiscernible by the untrained eye. Yet slip up for even a single frame, and your brain snaps back to reality, awoken from the spellbinding suspension of disbelief the artist slaved to create.

Problem 1: CG Animation is Too Smooth

From the opening sequence of the first episode, I was already scratching my head trying to figure out why CG characters such as Rock Bison and Fire Emblem didn’t sit right.

Fire Emblem makes me feel funny on so many levels.

Something about their movement was off in comparison to the rest of the cast—was it too smooth?

 My gut reaction was that these heroes were being animated on 1’s while the 2D parts were animated on 2’s. Clicking through the action frame-by-frame, I found out I was wrong—characters are also done on 2’s (for the most part).



Woah, back up for a second—1’s and 2’s?

Anime is created at 24 frames per second. This means that for every second of footage, there are 24 frames, or animation cells. The question of smoothness is then answered by how many unique drawings are contained within this second, with more drawings creating the illusion of lifelike, flowing movement.

Example of a walk on 1's (From The Animator's Survival Kit)

If each drawing within a second of footage is unique, that’s animating on 1’s—24 drawings per second (also known as full frame). The next step down would be animating on 2’s, for 1 unique drawing every 2 frames, or 12 drawings per second (also known as half frame). You can further subdivide from 24, leading to animating on 3’s for 8 drawings per second, and so on.

The same walk on 2's (From The Animator's Survival Kit.)

Televised anime is done on 2’s at best, or still frames at worst. Your brain fills in the gaps and never misses those other 12 frames, making action on the 2’s believable, though technically imperfect. More importantly for studios, being able to halve the amount of drawings without noticeably affecting quality allows productions to stay solvent.

This is a holdover from the limited animation workflow of animating on 3’s introduced by Tezuka back in the 60’s, a herky-jerky ghetto that divided the industry between artists like Gundam’s Tomino who just wanted to tell a good story regardless of the quality, and realism junkies like Miyazaki who pushed for greater immersion though lifelike (and expensive) animation.

But that’s a story for another day.

What does it matter to CG animation if its 2D counterpart is going number 1 or number 2? Like I alluded to earlier, the two are trying to dance, and need to keep on the same beat.

By default, CG animation is all on 1s. Once you set your key frames and breakdowns, the computer extrapolates the in-betweens. This is an overly simplified explanation of the process, (and I don't mean to suggest that the computer does the animator's job for them), but my point is full frame animation commands no additional cost in CG. In fact, while 2D anime cheats costs by using long (non-animated) holds on characters, dropping frames like this in full CG is artistic suicide, as the character reverts to a lifeless puppet the instant they stop moving.

So, in theory, we’d end up with the 3D parts on silky smooth 1s, while the 2D is done on comparatively choppy 2s, 3s, or lower. The result: A jarring mismatch of objects moving at different speeds, like walking through a stationary train as the tram across from you begins moving in the opposite direction. Your brain might not miss those extra frames, but it sure notices when certain objects on screen have them while others don’t.

Of course, this is all Animation 101. CG elements in 2D shows are generally done on 2’s in an attempt to synch up to the paper drawings. But knowing is half the battle. It took the industry years of trial and error to make the leap from pencil, and from paper to stylus and tablet. For a CG specialist, going back over the divide to purposefully create “jerky” animation requires similar mental gymnastics.

Meaning: Even with 2D and 3D dancing to the same beat, they still feel off. Why is that?

Problem 2: Data Overload

It comes down to an information gap and the issue of clashing animation styles. The former plays into the latter, so let’s start there.

In terms of the amount of information contained within a single image, CG lords over 2D, no contest. But don’t discount 2D too quickly. It’s information is hidden beneath the surface, implied between the lines. This visual shorthand acts as a sort of secret aesthetic message that our brains decode. With 2D, less can be more.

The Temple of Seven Golden Camels explains better than I can how simple and complex elements either complement or hinder one another.

This lack of raw materials means that a 2D artist needs to stretch each resource to its limits. In other words, no wasted poses, no superfluous movements. Color treatments should be spartan with simple two-tone shadows.

And on the other side of the wasteland, languishing high up on a throne of abundance, CG blows the Horn of Plenty without a care in the world. Therein lies the problem. Remember that we’re trying to unify two competing visual elements. CG needs to cast its worldly opulence behind in order to move down below the information poverty line and square off with 2D. The reverse, frankly, would be impossible.

Easier said than done.

The lack of a proper outline from the image-flattening Toon Shader is another dead giveaway, but too technical to go into here.

As you can see in the image above, Rock Bison sticks out like sore thumb against the other simplified characters drawn in visual shorthand. He's got too much detail, too many lines, too complicated of shadows.This problem is persistent on all the CG characters.The gradients on Fire Emblem’s jock have the same effect as David Bowie’s monstrous cod piece in Labyrinth. You can’t stop staring! Too much information.On the other hand, Lunatic is nearly seamless. There are a number of scenes where it’s tricky to pin him down as being either 2D or CG, and that’s amazing! The audience shouldn't be in a position to ask this question in the first place. They’re supposed to shut up and enjoy the fireworks, and Lunatic’s design allows us to do exactly that. His suit is simple enough to work in either style. His face mask means no awkward “cell in a suit” hybrid animation like with Wild Tiger and Barnaby. Conceptually, he transcends hero and villain. In execution, he transcends 2D and CG.

Of course, this applies to still frames. What happens when the characters start to move?

Before we answer that, let’s think about the process behind drawing the cells themselves.

Anime techniques have their roots in western animation. Hanna Barbera in particular developed a cost-saving cheat that forever changed the face of production. Their trick was a type of limited animation where they separated a character into a number of layers, each animated separately. So if you have a character speaking and gesticulating, an animator only needs to draw the moving arm, mouth, and chin, while the rest of his body remains static, thus recycling the cell.

First we discovered that we can cut drawings in half by animating on the 2’s. Then we figured out how to optimize the drawings on a cell, dropping costs, realism, and information even lower.

Easy with the finger curls there, buddy.

TV anime doesn’t have the budget to always redraw the entire character for each movement, so they’ve adopted this cheat with great aplomb. And while this works great in 2D, it clashes with the CG. Flame Emblem, Rock Bison, and the power suits have a tendency to overact, either through too many poses, vaudeville hand gestures, or moving body parts that would otherwise be static on a hand-drawn image. This makes them appear much too busy, and when placed alongside their taciturn 2D counterparts, like a spastic child that can’t sit still.

Problem 3: Perfect Scaling

As we can see, 3D does certain things too well for its own good. Full frame animation and image information are fairly obvious problems, though there is also a more subliminal element working to pull us out of the action.

CG allows for perfectly calculated perspective. This comes off feeling cold and robotic next to a show’s inherently wonky hand-drawn elements. It’s somehow fitting that CG elements tend to be equally lifeless mecha and vehicles. Keeping these rigid objects in proper perspective as they drive away towards the vanishing point or zoom around in dizzying space battles is difficult enough to make a traditional animator want to snap their pencil in half.

CG solves the problem by providing models that, by default, create perfect perspective regardless of the movement or camera angle. However, this perfection makes the image appear sterile in relation to the the slightly janky drawings around them.

A smart animator will work around this limitation by purposefully distorting every few frames to give the impression of hand-drawn wonkyness that keeps the image alive. To bring up Lunatic again, this close-up shot from the end of Episode 8 illustrates how effective “messing up” a drawing can be.

There's devilry at work here!

Pay attention to how his eyes distort as he pulls his head back. They wobble, and the screen right iris even disappears for a frame, giving him Macross googly eyes. Based on previous shots, I want to believe that this is CG, but the actual image has all the tells of a traditional cell. Perhaps they switched over to 2D for certain shots? In any case, nice trickery Sunrise. I take my hat off to you.

To recap, integrating CG with 2D raises the following problems:

1) CG animation needs to drop its frame rate to match 2D.
2) CG designs need to be simplified to match codified 2D drawings.
3) CG needs to purposely distort models to counteract otherwise “too perfect” perspective and scaling.

That’s one hell of a series of flaming hoops to jump through. Why even bother?

The answer would be an entire post in itself, so I’ll just sum up the situation as I understand it.

1) Lack of skilled animators domestically.To rephrase Mutant Frog’s translated article on how the industry is hollowing itself out, by outsourcing in-between drawings to other countries, Japan has lost a training ground for new talent that would eventually graduate to become key animators. A true ouroboros. CG animation has since stepped in to fill the void.

2) Designs are complicated. (Related to Problem 1.)

Assuming there is still a small pool of talented artists in Japan, their simply aren’t enough to go around. Once again, CG is there to bring complicated mecha and monsters to life so the remaining animators can focus on other elements, such as characters and acting.

To give a recent example, Gundam 00 breaks out the big guns with 2D drawings for the hero's mobile suits, then falls back on disposable CG for background mecha and spaceships.

3) You have to stay on the cutting edge.

CG may never replace the raw emotion and exaggerated poses that are 2D’s bread and butter, but it can supplement it in other areas. It allows directors and artists to explore new means of expression and rewrite the rules of the medium. So long as there are new toys, people will want to play with them.

Whew, we've barely scratched the surface and already this post is too long for its own good. This covers what I've noticed to be the most jarring elements of CG in 2D anime. There's plenty more to discuss, such as backgrounds, camera movement, and the pros and cons of a Toon Shader, but I wanted to shoot close to the hip and keep to more the fundamentals of animation rather than dig through backend processes.

Despite the flaws and odd design decisions, Tiger & Bunny is relatively well done, which makes it all that more fun to pick apart. There's no denying that persistent ear flick you feel at times, but at least now you have a better sense of what causes it. And with that understanding and enough time, you may even learn to ignore it.