DETAILS


COLUMNS


CONTRIBUTIONS

a a a

Non-Photorealistic Rendering

Vol.32 No.1 February 1999
ACM SIGGRAPH

Manga and Non-Photorealistic Rendering



Sugano Yoshinori
Nippon Television Network Corporation

Translated by Hikasa Chiaki and Michael Arias


In focusing my efforts on non-photorealistic computer graphics (CG) imagery, I have been struggling to create pictures that are, in a sense, more realistic than those rendered using so-called photorealistic techniques. In the past, when directing television programs, I felt something distinctly lacking in CG visuals — CG tends to look monolithic and excessively clean-cut — I don’t deny that the perfectly integrated monsters and super-realistic special effects are quite captivating for viewers, but most of them are little more than just beautiful visuals on the two-dimensional screen or monitor. As technology improves and the visuals created by CG tend towards seamless reality, the more lifeless seem these pictures. The pervading opinion of late seems to be that CG should be used such that an audience can’t detect its presence — a mundane goal in my opinion.

When we produce a film based on a well-known story, inevitably some viewers will find the original more to their liking. Readers of the original text have the freedom to interpret characters’ movements and the author’s viewpoint themselves, an aspect not unique to, but inherent in, the process of reading. Thus, even the most complete visuals will fall short of providing an audience with this most satisfying dimension. Filmed imagery and hyper-realistic CG seldom allow the degree of interpretation a more participatory medium such as the printed word does to the viewer. This might explain why I feel so little for most of the CG imagery we see in popular media. At any rate, there can be little doubt that we collect information with our five senses and our brain interprets it. What is “real” for us is what is interpreted, not what is reflected on our retinae.

Perhaps then CG can facilitate this aspect of interpretation. Interpretation is how we arrange the world in order to make sense of it. That is, it is symbolization of the world performed by our brains. Much CG, though, aims at visual simulation exclusively and ignores the act of interpretation as an important component of our enjoyment of visual art. For example, when creating a realistic image of a glass through CG, a common step is to calculate how the light falling on the CG glass refracts and reflects. I think, however, that in trying to create images meaningful to our audiences, we might be better off starting with the question “how best to draw this glass so that it will capture the necessary aspects of a glass?”

It is with this in mind that I have turned my attention to non-photorealistic rendering in my work of late. A drawing, as we know, is essentially a symbolization of what we capture with our eyes. Therefore, if a computer could draw as we draw, perhaps the process of symbolization we perform might be elucidated. The reason why I decided to apply the “toon rendering” process to the Japanese manga movie, The Princess Mononoke (released in July 1997 in Japan), was that I wanted to see if it was possible to draw manga images with CG.

To explain why non-photorealistic rendering is relevant to creating manga via CG it helps to know a bit more about manga and its history. As you may know, Japan is the world’s greatest producer of comic books (manga). You can see Japanese manga anywhere: in Japan, you will often find office workers and students reading comic magazines while riding commuter trains. Japanese people like manga more than any culture in the world. This, I believe, is because manga is essentially a modern variant of a Japanese traditional drawing method.

Figure 1
Figure 1: Picture scroll.

Figure 2
Figure 2: Ukiyo-e woodblock print.

Japanese Traditional Drawing Methods

The most traditional drawing method we learn in studying Japanese art history is the picture scroll — Shingizan and The Tale of Genji are examples of works in this medium. The picture scroll is a form of drawing believed to have originated in the ninth century. In a typical picture scroll, text narrative is interspersed among illustrations and read from right to left. In short, drawings and Japanese text co-exist in a long, narrow manga. There are various common forms: one might illustrate a scene of the story, while another might explain a long scene by depicting several details together. In any case, the vivid line drawings of a scroll painting depict the shapes and faces of people at the time. By the twelfth century, scroll paintings began to feature stylized depictions of hell, demons and evil spirits, and in the thirteenth and fourteenth centuries historic battles, biographies of great priests and commemoration of temples and shrines became common fare.

By the seventeenth century, the scroll painting was transformed into the most famous Japanese art form: Ukiyo-e. Ukiyo-e are genre paintings that were most common during the Edo period (1600 to 1867), and were mainly popular in Edo (now renamed Tokyo). Around 1680 the word Ukiyo-e began to be used for paintings featuring prostitutes, actors and actresses — the inhabitants of the pleasure quarters and theater streets. The word ukiyo, “floating world,” if translated literally, implies vogue, pleasure-seeking, Epicurean ways, and was often attached to other words as a prefix at that time. Initially, as the origin of the word makes clear, the main subject of Ukiyo-e was people gathered around pleasure quarters. One popular genre, the Makurae, (“pillow picture”) depicts couples having sexual affairs. Ignoring physical principles and rules of perspective, couples are typically shown with absurdly pronounced sexual organs and exaggerated muscularity — the dynamic style of these pictures is the very essence of modern manga.

Ukiyo-e is also characterized by its method of manufacture, the wood-block print. At the time, the print was the only method of cost-effective mass production available. Ukiyo-e was perhaps viewed more like tabloid magazine or news photos rather than as fine art. Making a wood block print consists of three procedures: Drawing the original picture, engraving one or more registered blocks of wood and applying colors to paper by stamping with the wood blocks. The entire process is much like the one used to make today’s magazine; this is why Ukiyo-e died out so instantly when photography and photogravure became popular processes. Although many pieces of Ukiyo-e were seen in Europe and influenced modern art movements such as impressionism and art nouveau, Ukiyo-e never saw its restoration in Japan. Since then, however, the characteristic line drawing (of Ukiyo-e and the picture scroll) has come back to life again as manga in modern times. From the picture scroll to Tezuka Osamu’s manga and Miyazaki Hayao’s anime, Japanese people have continued to develop the artistic technique of representing space with line and solid color over more than a thousand years! You will notice when watching today’s Japanese TV programs and animated films, distance between two objects is illustrated not by gradation but by the contrasting colors and outlines.

Sugano Yoshinori is currently a Program Director at Nippon Television Network. Until last year, he was CG Director at Studio Ghibli, and responsible for integrating digital techniques into the cel-animated feature films of Hayao Miyazaki (which include Pompoko, Whisper of the Heart and The Princess Mononoke, soon to be released in the United States). He is currently producing a fully computer-animated television program for Nippon Television Network.

Sugano Yoshinori
Nippon Television Network Corporation
14 Niban-cho, Chiyoda-ku
Tokyo, 102-8004 Japan

Tel: +81-3-5275-4647
Fax: +81-3-5275-4520
Website


The copyright of articles and images printed remains with the author unless otherwise indicated.

Manga and Computers

For those of us in this manga-dominated country trying to create non-photorealistic imagery, rather than pictures such as those a camera might capture, manga are an obvious source of inspiration. Lately, it has become something of a fad to use computer graphics to create for manga in Japan. As the production of these images becomes more streamlined and straightforward, our techniques will become more popular in the Japanese manga and animation industries. However, it is important to remember something in our use of non-photorealistic rendering techniques. A computer-simulated visual is still a pattern of dots or lines or colors on the screen or on paper which represent a particular shape, and it is still impossible for a computer to imitate the symbolizing process that is essential for us to “draw” a picture, and then for others to interpret its meaning. Here I offer a cliched comparison between a CG image and an “attractive” person. The person may be attractive at first glance, but a partner will start to consider not only their looks but also their character and their mutual chemistry as the relationship progresses. The same thing will certainly happen with computer graphics. Clever graphics attract people at first, but at some point they fail to stimulate viewers. The meaning of drawing manga with computers lies not in simulating manga but in capturing the symbolic essence of manga drawings themselves.

I said in the above that non-photorealistic rendering is the abstracting of concrete visuals by computer, and the final impact of the process depends on how others interpret the lines and colors we use to represent the subjects of our work. How we collect information from the world, how we pick up its characteristics and how we emphasize them will determine the impression left by the imagery we create. This is what artists of picture scrolls, Ukiyo-e and today’s manga have always known. With the benefit of CG technology, what we must do is not to imitate using the computer, but rather to use the computer to express characteristics of a subject. Thus, we may someday be inventing images which are neither photorealistic nor manga-realistic, but CG-realistic. Perhaps we will find methods of abstraction, representation and symbolization which are only possible using the computer. CG artists may influence the course of other media, as we are now influenced by the world of manga. Perhaps this will only happen after CG has accumulated a thousand years of tradition — something exciting to imagine! The main reason I am so attracted to non-photorealistic rendering is that it provides opportunities and possibilities to create new styles of visual image.

Manga as a Medium

Lastly, there is another point useful in understanding traditional Japanese visual arts such as picture scrolls and wood-block prints; scrolls and prints were not fine art, as such, but popular art used as a means of communication. What is most important in these media is not the picture or its characteristic style but the subject depicted therein. This has not changed today with manga. Manga is a medium, not art. What is symbolized by a line drawing is meaning, not a picture. Some genre of Western painting have discarded meaning as an essential component and made pure graphic style itself into content. Manga artists, however, have drawn visual images in order to convey meaning. Manga doesn’t require the strict representational accuracy and one could say that in manga, anything goes; the conventions that exist for drawing manga are only those required to convey meaning to readers. Therefore, a paradox of manga is that drawing manga has nothing to do with skill at drawing pictures. What is important is that what you are representing with your manga can be understood.

The same thing should be true of CG. CG need not follow established conventions of representation. I think that computer graphics and non-photorealistic technologies offer us chances to create images which are actually more real to viewers than those created using photorealistic methods.