Interview with Satoshi Kon, Director of Perfect Blue

In Jojo’s Bizarre Adventures (1993), which I worked on previously, I had supervised the script, continuity and production, but sound, dubbing and the rest was the work of the director. I wanted to have a go at doing all of them and Perfect Blue presented the ideal opportunity.

I didn’t want to just read manga (comic books), but have a go at writing them myself. I was heavily influenced by Katsuhiro Otomo’s Domu (1980) and Akira (1982). I especially liked Domu, and I figured that if I were allowed to make just one movie from among all the manga I’d read, that would be it.

From the very beginning I enjoyed drawing pictures and my early scribbles eventually developed into manga. At length, one story I completed won the ‘Kodansha Manga Award’, and ever since then, I’ve been drawing manga not as a hobby, but as my livelihood.

Up until around the time I was in high school, I watched various Miyazaki anime and shows like Yamato (1974) and Gundam (1979). I bought loads of manga when I was younger, but after I entered university I guess I stuck almost exclusively to live action movies. I saw most of them on video, and made it my policy to look for scene setting, format and production as reference for drawing manga.

I can’t think of a particular film or a particular director who has really ‘influenced’ me, but I did gradually absorb the things I saw. For example, Akira Kurosawa’s format is solid and easily understood. However, just because I’ve seen it doesn’t mean I can imitate it, but it proved to be my best method of study for making movies. But it would be foolish to say just because of that I’d been ‘influenced’ by Kurosawa’ (laughs).

The most important influence on me at the time, I think, wasn’t a single film but the works of Terry Gilliam. Despite being fantasy, his depictions are quite bitter, his narration also throws ‘curve-balls’, and rather than covering every point in detail, he takes the staging off to a completely different point and plucks out a single, vivid theme. I especially like Time Bandits (1981), Brazil (1985) and The Adventures of Baron Munchausen (1989).

When I draw myself, I am quite naturally interested in whatever’s around me, so that there’s a feeling of starting from a realistic point of view, with which fantasy is then mixed, and finally finishing with pure fantasy.

What came to interest me enormously when drawing manga is of course the format, but also the layout panel by panel. You draw six or seven panels to a page, and when you open the page, there’s the question of which picture hits your eye, and when a double page is open in front of you, there’s the flow starting from ‘this’ picture and climaxing in the end with ‘that’ picture. It has to be ‘cut’, edited like a film.

When I was drawing manga, I put a lot of thought into guiding the reader’s eye by the positioning of the speech bubbles, should I do this if there’s a pause for breath, the character’s facing this way, and so on. For example, when depicting a movement such as falling, I would put the speech bubble above the character, and if the next dialogue box is placed diagonally to the bottom of that, the flow is quite natural. In my opinion, controlling such elements is the most interesting aspect of manga. Just please don’t ask if I’m any good at it! (laughs).

On Roujin Z (1991) I was responsible for set design. On anime set design especially, the most important thing is to make these places look lived-in, somehow real. Even vaguely placed background objects have to look as if they ‘have a past’, and came to be there through a process, not a man drawing them there. So, for example, if I’m depicting a dirty apartment, there’s someone who’s been living there for quite a while, the areas around where he stuck posters on the walls have to be slightly discoloured.

Well, a lot of animation is extremely ‘samey’. There’s only one tradition, one style of filmmaking, and one set of tools. In manga, we have many, many different genres, from children’s comics to material drawn exclusively for adults – there’s an infinite variety of subjects and genres. But it seemed to me that anime was almost nothing but science fiction robots and beautiful little girls, and it just gets boring after a while. Sometimes you hear creators complaining that they didn’t have the budget to do what they wanted, or that their best ideas were cancelled, but ultimately, I think that most of these people do what they do because that’s all they like. I’m not saying that’s a bad thing, but it is a little sad. That’s why I tried deliberately to handle Perfect Blue from a different point of view.

I wasn’t interested in putting a lot of effort into the stabbing or whatever. I was given the three keywords of ‘idol’, ‘horror’ and ‘fan’, and was completely free as a director as long as I stuck to those overarching themes.

First of all, I created images of the young woman protagonist who is trying to strike out on her own, puts her past behind her, and comes under attack. Then, when I was discussing things with the scriptwriter Sadayuki Murai, we decided to have a play-within-a-play – a construction sort of like a Chinese puzzle.

My first sense of unease came with the first plot suggestions, but we suggested loads of ideas to each other in order to press on for the time being. Because we had nearly all of the materials assembled, it was like we were editing something, not creating something anew. Because I’ve written a considerable number of scripts myself, I tried tying all of the necessary scenes together, and then looked it over. Whereupon, parts that had been written with great care and consideration became evident, and I discovered parts that needed to be dropped.

For example, Mima’s dramatic line, ‘Who are you?’ is something we repeat several times, to give a special meaning to an ordinary question. Every time we had a draft of the script, Sadayuki Murai and I would take perhaps five or six hours to discuss it. Whenever I understood what he was getting at, it would spin off into weird ideas of my own, so this turned it into a kind of negotiation. It was fun, and we ended up watching a lot of movies together, with movie scenes and shots becoming our mutual language.

The Chinese box construction of a play within a play was completed mainly at the scripting stage. The real changes in construction came at the storyboarding stage, because that allowed us to emphasise things that weren’t really brought out in the script. Things that can be done in writing are best completed in writing, but we ended up changing our portrayal of the villain at the storyboarding stage.

In the original script, Uchida (the fan) was a major character, but he hardly did anything, and we couldn’t expect the audience to get interested in a character who merely ‘acted’ suspicious. So, we actually changed him into a character that committed murders.

Our main problem there was that we had a lead character that didn’t do anything. She is always made to do something by somebody else. Normally there is a rush of activity around the actions of a powerful central character, but there is none of that here. From the point of view of the script, it is difficult to just have the events develop around her, and we agonised about that quite a lot. Therefore, we supplied a powerfully active character to revolve around Mima, and at one point deliberately constructed an action scene for Mima in the radio station.

Well, storyboarding is always fun, and when you’re revising and re-revising the images in each scene, sometimes it’s easy to get carried away. This project originally started out as a video, but ended up as a big-screen spectacular, so there were many cuts that we had to make. In the end, I think we had to throw out about 100 scenes.

The cuts were made because the cinema running time was limited, so nearly all the missing footage was just there to show the passage of time. We cut several scenes of escalating suspense that were originally there to emphasise Mima’s gradually growing fear. It’s very difficult to bring out the feeling of ‘gradually’ when you don’t have the time to do it, so that led us to become bolder and cut most scenes in which stuff crept up on you gradually.

We also used many jump cuts to link separate episodes and as an expression of mental confusion. We’d cut fast from one thing to another as if it were a fight scene, even if there wasn’t any action involved – it helped emphasise Mima’s sense of confusion, such as when we jump from her opening concert scene to the very different shots of her everyday life. It’s a scene that shows her with both her idol-mask and her normal-mask on.

It would be boring just to cut out little bits and piece them together, so I decided to create something along the lines of Slaughterhouse 5. I used imaging techniques to connect different scenes by piling action scenes or images one on top of the other. As the film progressed, especially in the later half, we decided to cut from scene to scene faster and faster.

With animation, there are many cases where the style is already fixed. For example, dream scenes have a pattern: when you get wavy lines on the screen, it means that you’re entering a dream sequence, or the scene switches to sepia tones, or cream flows onto the top of the coffee, creating a whirlpool, or there is a close-up of someone’s eyes.

But that kind of editing is totally boring, and I think that there are many more ways of introducing dreams and flashbacks. Even if the shot or the scene changes, they must be linked within the flow of the story and I thought that it would be interesting if the viewers did not immediately grasp they were watching a flashback or a dream. This kind of editing may be due to my theatrical influences. The vigorous scene changes of my generation’s theatrical groups are very speedy. Of course, in the theatre there is no ‘editing’, and if scene changes are performed one after the other in darkness, continuity can be disrupted. Clearly, modern theatre has put a lot of thought into expressing on stage what they had absorbed from the movies. I figured I’d try this in reverse, and see if it worked on film.

Viewers are too used to being treated kindly, so I’ve broken this pattern deliberately. For example, when the scene changes ordinarily, a long shot is used to let you know. Suppose there’s a cafe; we show the cafe, then we show the man and woman meeting there, then they start talking. I hate this sleepy kind of continuity, so there are almost no cuts to introduce the scene, as I fully intended to make a film that didn’t show you the scene changes.

Most of the film is set indoors. Mima’s room, Uchida the fan’s room, even the film studio. The other Mima also makes her first appearance from the PC in Mima’s room. To tell the truth, I thought long and hard about how the other Mima should best make her appearance, and it was Murai who came up with the idea of the Internet: the homepage Mima’s Room. In this way, we decided to make all of the rooms evoke the same thing. There are three ‘screens’ in the room where Mima lives: a TV, a PC and a tropical fish tank with the same 3-by-4 dimensions as a monitor. We also shot Mima’s room as if it was being viewed on a TV screen. This is because we wanted to give a diluted sense of reality, as if all of the events were taking place within a TV screen of some kind. Later, when the scriptwriter Shibutani comes into the parking lot, the entrance is framed as if it were a TV screen, with the same 3:4 dimensions.

There are many people who don’t know that the Japanese also have a culture in which B-grade idols are manipulated. Ours is a closed culture, and very small. It seemed slightly strange to me that foreigners could watch our films and find them interesting, while not understanding the background.

I firmly believe that Japanese animation culture is unique. If a robot appears and an explosive battle begins, it’s written off as ‘what else do you expect of anime?’ But if that were true, we wouldn’t have animated Perfect Blue, and it would have been made as a live-action film.

Just as we have an extremely broad manga culture, the same should be true of animation, or that’s what I think anyway. Moreover, although I didn’t aim for ‘realistic depictions’, everyone keeps telling me that it’s extremely realistic. Of course, I created the expressions from the fact that I wanted reality in the pictures for the movie’s sake, and I simply did what was necessary in order to tell the story.

I made this as animation because it’s my personal preference, and it was easier to film the virtual Mima with animation than with special effects and a real actress. Besides, if I had made a live-action movie, many parts would have appeared unnatural.

I’m determined to make more animations in the future. I want to eliminate my weaknesses and build on what I’ve learned from Perfect Blue, and supplement and augment it. In truth, so long as I can make it interesting, it doesn’t have to be animation; I don’t mind working in another medium. So long as I have interesting images within me, I’m sure that I can express them, no matter what the medium. However, my feelings for the moment are that most of my interest lies in animation.