Digital Oshii: from Patlabor 2 to The Sky Crawlers (2)

Mamoru Oshii
Born in Tokyo on August 8, 1951, Mamoru Oshii is one of the most remarkable personalities in modern Japanese filmmaking. He introduced introspective philosophical speculation into the world of animation, influencing at the same time movie creators all around the globe with his visionary style. Oshii joined the animation industry in 1977. His main works are Urusei Yatsura 2: Beautiful Dreamer (1984) the epoch-making Ghost in the Shell (1995) and Innocence (2004, nominated for the Palme d'Or at the Festival de Cannes). Oshii has also directed a number of live-action features, including Avalon (2001). The Sky Crawlers, was nominated for the Golden Lion at the 65th Venice Film Festival, selected in the 33rd Toronto International Film Festival and was greeted with three awards at Sitges 2008. In 2010 he directed Je t'aime, a short film in collaboration with the Japanese super rock group GLAY, that had its world premiere at Annecy 2010.

Part 2

What is your personal opinion about the usage of CG, aesthetically, philosophically, and in terms of practical considerations?
Aesthetically speaking, CG obviously expanded the expressive potentials in movies. Philosophically speaking, it did not solve the same old problem, because any technique is subjected to what you want to create. I could exploit CG potentials in full for Innocence because I had a clear vision of what I wanted to realize with that movie. The method and the theme of the movie were in harmony with each other. In a movie centered upon a dramatic story, or family entertainment, 3D could turn out as an obstacle rather than a help. This is why for The Sky Crawlers, I used a completely different approach. In The Sky Crawlers, all available 3D techniques money can buy are thoroughly applied, so we reached a fairly high level in terms of motion and texturing, but while for Innocence the technique was somehow the subject of the movie, for The Sky Crawlers I was focusing on the recreation of an emotional drama, although I indulged myself with some air battles among reciprocating engine planes.

Practically speaking, I could probably say that nothing actually changed: you are surely able to incorporate a huge amount of information into your pictures, but eventually the individual sense of the staff that is actually doing the job remains the primal issue. I believe that producing animation making a good use of 3DCGI is still a difficult task.

Hand-drawn Japanese animation, especially during the 90s, seemed to aim for a very photorealistic effect. In this respect, why not completely use CG?
As far as I'm concerned, I started focusing on photorealism back in 1989 with Patlabor: The Movie. For that movie, we used several optical tricks, such as blurring pictures or modifying exposure rate, in order to obtain a more realistic feel. This approach underwent a natural evolution with Patlabor 2: The Movie, where we achieved a high degree of photorealism. But movies like that are a minority, though perhaps a noticeable one. The mass of Japanese animation is still relying on codified symbols and flat images that are a direct inheritance from the ukiyo-e figurative style.

However, if you want to pursue realism with 3D, you'll soon bump against a very tricky hurdle. 3D may be suitable for puppet-like characters like in Shrek, but if you try to render realistic characters with 3D, you inevitably step into a minefield, because the more the characters become realistic, the more they end up looking like moving corpses, no matter how many details you add.
Nobody so far succeeded in creating realistic 3D characters getting rid of that peculiar, creepy feeling. It's not a technical problem, it's merely psychological. Could be a matter of personal aesthetics or cultural training, and I'm not rejecting the idea of making a full CGI movie a priori, but I don't fancy the idea of converting characters into 3D, as long as 2D animation is clearly more effective on expressive level.

How do you position the experiment made with Tachigui: The Amazing Lives of the Fast Food Grifters in this context?
I simply chose the inverse method: I took realistic characters and I reduced them to 2 dimensional objects. So I had pictures with the same quantity of information as a live-action movie, but their motion had the same quantity of information as in animation. I was eager to explore what kind of representation could be possible by combining these two ideas.

As someone who has pioneered the use of CG in Japan in both live action and traditional animation, do you think there are times when the use of 3D CG is particularly appropriate/desirable, and times in which it is inappropriate? For example, backgrounds, characters, special effects, and emulating camera effects.
It all depends on the project. Arguing that animation is supposed to do only this and live-action must be like that is simply pointless. Everything depends on the director's vision in using those techniques. Only with a clear vision you can break the barriers of live-action and animation using CG.

How do you feel about having both Ghost in the Shell movies placed so highly in the 3D Magazine poll, and therefore voted the most influential of all time by over 5,000 3D professionals?
Although I may sound immodest, I am not completely surprised with this outcome. Application of 3DCG in Hollywood movies is, to some extent, rather primitive. There are potential applications that are not explored at all. I'm confident that the impact that Innocence brought to the concept of 3D applied to movies has yet to be surpassed. And it's not a matter of budget or software, but it's due to a human factor. And since it's virtually impossible to gather the staff that made Innocence once again, that movie will remain one of a kind.

(2 - end)

Interview by Jasper Sharp. Partially published in 3D Magazine, January 2009.