Click on the image above to see a bigger/higher resolution
I watched Fullmetal Alchemist: Brotherhood from April 5th, 2009 to July 4th, 2010. It was a weekly ritual: every Sunday a new episode would be out, and since I’d read the original story beforehand I would practice my Japanese listening skills by watching episodes without subtitles the first time around (afterwards, I’d wait for an official subtitle version by Funimation to see what I’d missed). If there were a few things I looked for in each episode, they were directing, pacing, music selection, and animation quality. And if there was one animator that stood out the most, it had to be Yoshimichi Kameda.
Kameda’s style is instantly distinct from the rest of the production’s animators for a few reasons:
• Multiple angles/perspectives
• Generous use of close ups
• Frequently dutched/diagonal horizons
• Pencil-like line art in the finished animation
• Quick cuts that created a lot of dynamic within the scene
• Follow/tracking shots of movement (often paired with close ups)
Unsurprisingly, Kameda’s most noticeable work involves action sequences. In a video compilation, you can see why his style can be simultaneously enthralling and dizzying. Here are some screenshots of his work as seen in the linked video:
Greed/Ling vs Wrath/Bradley
Alex Louis Armstrong
Original Greed
Envy disguised as Maes Hughes
Roy Mustang
Lust
Roy Mustang
Riza Hawkeye
Hobo man Scar
Jerso (sp?)
Envy vs Ed
Ling vs Envy
Ed, Envy, and Ling (left to right)
Roy Mustang blowing up Envy
Roy Mustang still blowing up Envy, with Riza Hawkeye in the foreground (right side of the screenshot)
The above screenshots are excellent examples of Kameda’s strength and weaknesses in animation, especially with regards to conveying action. I’ve talked before about framing a fight/action scene and how you could storyboard movement to create differing effects. With animation, you can push boundaries even more since your characters aren’t bogged down by the physics that govern live action actors. Because of this, animators like Kameda can experiment even more with close ups, tracking shots, dutch angles, skewed expressions, shaky cam, artistic renditions of the environment - anything you can come up with.
However, there are moments when auteurs of animation may go too far with these kinds of experiments. Technical skill is one thing, but making sure what’s happening on screen is cohesive is another trick: if no one knows what’s going on for long periods of time, an animator’s attempt to be original may fall short of anything artistically effective.
With Kameda, his greatest asset and problem is the use of tracking close ups (at multiple angles) but forgetting (or choosing not to) establish the horizon, respectively. This style creates a lot of visual movement, and is best when there’s at least some sort of establishing shot – even for a second – to put everything into context. One of Kameda’s best examples is the scene where Envy reveals his true form to Ed and Ling, as seen here (Kameda’s work ends at 1:30, and afterwards a different animator/director takes over):
I like this particular sequence of Kameda’s because while there are a lot of close ups and movement happening on screen, there’s always something establishing the characters into context and location. In this case, Envy’s massive size puts Ed and Ling into context by default: when the camera tilts up from our heroes to Envy stretching himself after the transformation, we instantly get a feeling for how small both Ed and Ling are as compared to the massive homunculus. Also, in almost every scene there is a horizon established, if even for a second before the camera cuts away to different angle and occurring action. The horizon is defined by the sea of blood that all three characters stand upon, and since there are no walls or major objects in the background (and the “sky” is pitch black) Kameda can get away with multiple tracking/close ups because the viewer will see the horizon (sea) of blood somehow (the infinite sea of blood (“ground”) is like a physicist’s wet dream of infinite planes and infinite dimensions - an animator can pull almost anything and it’s unlikely the viewer will get disoriented during the entire action sequence). Kameda and the director do a nice of job of creating lots of camera movement with close ups, tracking shots and cutting while maintaining a sense of cohesion of the entire fight sequence.
Conversely, Kameda’s excess use of close ups and tracking shots loses its edge when he forgets (or chooses not to) establish a distinct horizon within each cut, as seen here when Greed/Ling fights Wrath/Bradley in the Fuhrer’s office (the fight begins around 17 seconds, where I assume Kameda takes over):
The main problem in this fight scene is that Kameda doesn’t establish the horizon very well in most of the cuts. The horizon is defined by the floor that Greed and Wrath walk/run upon, not by the walls or ceiling. When we don’t see the floor between multiple cuts, the grounded physicality of the two sparring characters seems less apparent (they could be floating on air and we wouldn’t even notice). This becomes problematic over longer periods of time (“time” as in seconds and number of cuts) because the viewer starts losing a sense of relativity with regards to the ground. Compared to the Envy vs. Ed and Ling scene, where the environment was immense and primarily defined by the sea of blood that all three characters stood upon, in this scene with Greed and Wrath the environment is much more detailed and constrained comparatively (the Fuhrer’s office is a defined room and space), so it’s even more important to establish a sense of horizon more frequently so the viewer doesn’t become disoriented.
The few times that we actually do see the floor help reestablish what’s going on, but the part where it matters the most - when Wrath pins Greed on the floor - is the most lackluster of the entire sequence: there’s no unique angle, no dutch, no skewing – just a plain old linear horizon and framing that places both characters dead center – the hell hole of the Thirds Rule.
If there’s one thing you should avoid almost entirely, it’s placing the characters dead center screen at a un-angled horizon during any climax of an action sequence, and especially in animation. With live action, you could possibly get away with this since the actors are always breathing and twitching and their clothes are always ruffling, regardless if they stand dead still; with animation (especially on lower/tight budgets) the artists can’t spend precious extra minutes drawing these characteristics, so the characters on screen will look especially flat and boring if you don’t frame or angle them in an interesting way.
As a last reiteration, it is always the ground that defines the horizon – not the sky, not the ceiling, not the walls, not even a cat. To maintain any sense of cohesiveness during an action scene, establishing shots are a absolute must (without them, kiss your chances of cohesiveness good-bye) and are the most effective when you incorporate the horizon upon which the characters stand. Cuts, multiple angles and perspectives, tracking shots, and dutch angles are friendly tools you can use, but overusing them without establishing a horizon is cohesiveness suicide.
Kameda does good work throughout the series, but he’s not without his faults and demonstrates some prime examples of the strengths and weaknesses when animating action and characters. His work is a great reference for any aspiring animator and for anyone looking for different ways to frame action sequences. To see the episodes of Fullmetal Alchemist Brotherhood, please visit the official Funimation website (or search Hulu or YouTube for full episodes as well).
…
Additional Links/Readings
Yoshimichi Kameda profile on Anime News Network – shows a list of credits that Kameda has worked on.
Talented Up-and-Coming Animators: Yoshimichi Kameda – a more comprehensive overview of Kameda’s work and the notice he’s received in his career so far.