The History of Mushi Pro – 1.5 – Atom through its storyboards

In the previous article of this series, I stated that Tetsuwan Atom’s production was “centered on one document, the storyboard”. Although the production pipeline of anime has changed a lot with time, the storyboard’s central place has remained constant. It is, alongside the layout, the lifeline followed by most of the staff, the central document which has to be both adapted and interpreted. It is therefore very important to understand the history of how anime storyboards, or ekonte, appeared, evolved and were used.

I initially planned to include the contents of this article in the previous piece on Atom, but as it became increasingly long, I decided to make it a separate piece. The writing will be less formal, I’ll cite my sources less systematically, and the goal will be different: not to narrate certain events, but to analyze a certain object, the storyboard. This is experimental for me, as I’ve wanted to analyze storyboards for some time but never found the good occasion or way of going around it; hopefully this piece works out and proves methodologically interesting.

Its very existence is, naturally, made possible by the access I had to actual storyboards. My main case study, the storyboard for episode 90 of Atom, is available in its entirety (except for 2 pages which have been lost) in the first volume of the Osamu Tezuka Storyboards Collection. Other storyboards could be found online, in the Atom DVD box booklets, and the catalog of the exhibition The World of Yoshiyuki Tomino, containing 10 pages from episode 96. 

The advantage that these have is that, perhaps except for the ones included in the Tezuka Storyboards Collection, these are all originals. Indeed, when storyboards are published, they often go through a process of editing – for instance, the boards published by Ghibli aren’t perfect reproductions of the ones actually used on-site but clean copies. This does not seem to be the case for those used here, which should all have been taken directly from Mushi Pro’s archives.

An introduction to the ekonte

While I don’t know when exactly ekonte first appeared in Japan, we have material proof that they were used from the earliest days in Tôei Dôga, as pages from Hakujaden’s boards remain to this day (at the time of publication, some are being shown at the Tôei Animation Museum, although I haven’t had the opportunity to go see them yet). In his autobiographical novel, Yamamoto also mentions that he did ekonte in Otogi Pro. Tôei’s storyboards were big, unpractical, crowded objects on B4 paper with around 10 panels per page. They remained so until at least the late 1960s. Mushi’s boards were also on B4 paper, but included less panels, first 6, then 5, which remains the current standard in the industry. The storyboard paper seems to have generally been yellow, with copies being on white paper.

Modern storyboards usually contain 5 columns: from left to right, cut (shot) number, pictures, written description of each cut, lines, and number of seconds & frames. Mushi’s boards were almost identical, except for the last column, broken down in 7 sub-categories: background, music, special effects, photography, genga, douga. These are quite clearly indications for each section, and this space must have been used to indicate the length of each new element as well as bank cels, backgrounds, or sound effects.

Storyboards were essential documents, but they didn’t circulate easily at first, because of the unavailability of copy machines. It seems that, in Mushi, they were made a little at a time and that the finished parts were immediately sent to animators. As the studio grew, it acquired copy machines, and full storyboards could more simply be circulated among the staff. Naturally, the early copy issue meant that Mushi’s animators rarely had a grasp of the full episode, especially if they worked without a script either. This issue would have been solved in the early stages of the production by the fact that storyboard artists and animators were the same people. 

Naturally, each storyboarder used the storyboard differently, with some having very detailed drawings and few written remarks, and some others being the opposite. Moreover, as I have shown in the previous article, the exact function of storyboards on Atom evolved alongside the work of those who wrote them, the episode directors. However, the impression that I get from all the Atom boards I could see is that drawings are generally clean and detailed – which makes sense, since all storyboarders except for Tomino were animators. As we will see, some shots seem like they were directly copied from the storyboard to the key animation; still, the amount of detail of each panel of the storyboard does not necessarily determine how detailed and directive the storyboard is as a whole. It is essentially a case-by-case basis, which is why I’ll immediately turn to my case study.

Osamu Tezuka, Atom #90

To provide a concrete example of how ekonte worked on Atom and how we can use them, I will turn to episode 90, as its storyboard is available almost completely in the Osamu Tezuka Storyboard Collection. Sadly, the full credits for this episode have been lost, and we don’t know who the animators were; but the script and direction were by Tezuka, and the assistant director was Yoshiyuki Tomino. Given all I have said about Tezuka’s working method, this storyboard may not be very representative of how other directors went about storyboarding; but I hope it is still better than nothing.

Let us start with a brief synopsis of the episode. A group of robots is accused of having abducted a human child, leading humans to besiege the robots, hidden inside a dam. Various fights occur, with Atom on the side of the robots. Then, Atom discovers that the lost child has disguised himself as a robot but, when he tries to bring him back to his parents, the child falls into the pipes of the dam. With the collaboration of other robots, Atom saves him, thus showing the robots’ goodwill and resolving the conflict.

A quick first look at the storyboard for this episode reveals a few things. First, the written indications are close to nothing: most of the text is the dialogue and nothing more. While this episode is already well into the “third phase” of the production, it’s possible that, as was his earlier habit, Tezuka wrote no script and the storyboard represented his initial ideas. Second, if we look at the few indications that there are besides the dialogue, we notice that the technical vocabulary of animation storyboarding is already well in place: words such as “PAN” or “T.U.” (Track Up, when the “camera” approaches an object) and “T.B.” (Track Back, when the “camera” moves away from an object) regularly appear. On the other hand, throughout the entire storyboard, there isn’t a single indication for the music or sound effects. Such isn’t the case in the other storyboards I could access, illustrating a clear visual-centrism on Tezuka’s part.

Similarly, the indications for backgrounds, just next to those for sound, are very vague and generic. They are, most of the time, one-word descriptions such as “sky”. The drawings don’t provide much more precisions, which would have made layouts – or “background blueprints” – even more necessary. Each new background is numbered (backgrounds are collected in something called the “book”) and, whenever a background is reused, it is written in the storyboard: “repeat x”. 

If we get in more detail, there is one interesting element. One of the most arresting stylistic aspects of Atom is what seems to be a repeated use of depth-of-field, especially in indoor scenes: characters are in focus but backgrounds are out of focus and blurred. However, a close reading of the storyboard reveals that this effect wasn’t achieved by just using the camera focus, but also with the background art: see, for example, scene 3, cut 35 (pictured below) in which this effect is apparent. In the storyboard, the background indication is “n°46, brush background”. In other words, a camera effect was imitated through the use of what I presume would be airbrushes, though I don’t know what the actual process was.

Staying on the written indications, it is notable that bank cels are not explicitly mentioned in the storyboard. For example, at scene 2, cuts 39-40, Atom’s flight is bank animation, but the only things written in the storyboard are “sky” (?) for cut 39 and “multi [multiplane] on the clouds” for cut 40 – in other words, indications for the background and photography teams (note how the speed or number of frames at which the clouds are supposed to move is not indicated). It’s only when cuts from the same episode are repeated (in this case, the entirety of scene 7) that “repeat” indications are given. In other cases, it would likely have been the role of the assistant director or production team to realize that bank cuts would work at a certain place and insert them.

Now, let’s move on to the relationship between storyboard and animation proper. The storyboard drawings are rather detailed, and some close-ups appear directly copied from it. In this way, it is very easy to conceive how settei would have been drafted from the storyboards rather than separately from them. On the other hand, the animators took some liberties with crowd characters: such is the case, for example, at scene 4, cut 25 or some of the “ghosts” at scene 16, cut 43.

In fact, this latter scene seems to be the one where the animator took the most liberties. Just one cut prior, one of the humans is mistaken for a ghost and violently hit by one of his comrades with a hammer. The storyboard only indicates the hit and then the victim on the ground, whereas the animator added a brief “puffing” motion of the head, which grows into a huge bump under the hit. In other instances, the animators would add smears where they’re not indicated or add small movements.

On the other hand, Tezuka would sometimes provide detailed instructions for how the movement should go. Such is the case in scene 6, cuts 3-7, in which the machines used by the humans to attack the robots parade on. At cuts 5, 6 and 7, Tezuka provided rough drawings to indicate how the group of soldiers should come out of the truck – but these indications weren’t respected, and an additional profile shot was added between cuts 7A and 7B, probably based on one of Tezuka’s roughs.

Yoshiyuki Tomino, Atom #96

This second case will be less detailed, as I sadly do not have the full storyboard for Tomino’s debut episode, #96. I will therefore particularly focus on the opening scene. However, what few pages I have should be enough to show a few things: how Tomino’s approach differed from Tezuka’s, and how he exemplified what I’ll call “technical” storyboarding, perhaps encouraged by his background as a production assistant.

First, some general remarks. Despite Tomino’s own comments about his supposed lack of drawing ability, the drawings in this board do not compare unfavorably to Tezuka’s. As I mentioned in the previous piece, Tomino’s clear drawing skills, at least for drawing storyboards, probably played a role in his promotion to director. Since this is his debut episode, it’s not like he had received any formal training in the studio either – he just had spent all his youth drawing, and then learnt about storyboarding techniques during his time as a production assistant.

Another thing that jumps to the eye, especially when compared to Tezuka’s board, is the abundance of written indications. This is especially the case in the second page (cuts 4-8), where descriptions of the contents of each shot are inserted, alongside additional drawings such as one specifying how the dial lights should look. While I’m not sure I can decipher everything, the descriptions for cuts 7 and 8 seem to read like this:

“The (??) move towards the robot’s body. Every time, Faye’s (?) body lights up / Overlap 2” / Leave the robot’s body on screen / Transparent disks enter the robot’s body 5 by 5”

As the term “overlap” indicates, there is technical terminology everywhere. For a good example of how Tomino pays attention to small details that Tezuka completely ignored, we can turn to the first shots, two pans over the same background but in different directions. The instructions are extremely clear: “Fade In 2” + 12 frames / Overlap 4” / The pan for cut 2 is the same as the one for cut 1”. These are extremely precious, because, alongside the length of each shot (indicated on the far right of the storyboard), they make everything clear  for the background, photography and editing team. While I can’t read anything besides the “quietly” between parentheses, indications for sound effects are there as well.

Things aren’t always as detailed: indications for the backgrounds remain rather vague, and I don’t know what the letters “C.B.B” stand for in the following cuts. Moving on to scene 2, portraying a running bullet train, it’s only written “background pan” without any details on the speed of the movement. It’s only when a cel element enters the frame – Atom in cut 4 – that indications suddenly become more intricate: “In 10 frames / 1”+12 frames”.

I have no particular conclusion to offer to this piece, because it was largely experimental and mostly descriptive. This may not be true in all cases, but in this one, I don’t think studying the storyboards made me approach the “creative intent” of the directors who wrote them. It has rather made clear how essentially technical a storyboarder’s work is. Whatever they communicate, storyboards are, first and foremost, instruments of communication. What I hope this short study has shown, then, is the various modes through which such communication can happen, according to various levels of technicity. I also hope such a close study of storyboards can establish one thing, which is that, in terms of basic storyboarding techniques, there has been no major evolution since Atom – the technical terminology, different approaches to write (or draw) a storyboard and the various back-and-forths between storyboard and final episodes were already in place.