Interview with the director of "Zombieland Saga: Revenge," Part 2! How was the moving live part created? Interview with Munehisa Sakai (Director) and Ai Kuroiwa (3DCG Director)

The anime "Zombieland Saga: Revenge" aired from April to June 2021 and attracted a great deal of attention. Director Munehisa Sakai himself storyboarded and directed the live scenes in this work, and 3DCG director Ai Kuroiwa played an active role in realizing the live scenes.

As readers of Akiba Research Institute know, the live scenes in the first season of "Zombieland Saga" attracted a lot of attention, but the live scenes in the second season, "Zombieland Saga: Revenge," were even better than the previous one in terms of camera work and the movements and expressions of the members, which made the hot drama more exciting each time. We asked two of the key persons to talk about how they created the live scenes with their best efforts.

⇒What did the staff choose to meet and exceed the expectations of the fans? Interview with Director Munehisa Sakai, Manabu Otsuka (MAPPA), and Nobuhiro Takenaka (Cygames)!

In the second season, our goal was to raise the average quality of the entire live part.

Q: For the 3DCG part of "Zombieland Saga: Revenge," did the director, Mr. Kuroiwa, look over it, and then director Sakai checked it?

Kuroiwa Yes, that's right.

Sakai But in the end, there must have been many parts where you said, "I'll do it! I'll do it!" in the end, didn't you?

Kuroiwa I did everything from cutting out the motion capture (abbreviated as "mo-cap") to delivering the film myself. I like this work so much that I couldn't help but want to do it.

Border But I had quite a lot of work to do, so I asked the production staff, "Mr. Kuroiwa, are you okay? (laughs).

I see. What changes have you seen in the 3DCG part of "Zombieland Saga Revenge" since the first season and what are your goals for the second season?

Kuroiwa: We continued to use the same body parts (face, hair, body, etc.) from the first season, so the challenge was how much we could improve the expression with those models. I myself participated in the second stage of the project after working on the first stage and the "Saga Incident" music video, so I had accumulated a lot of knowledge on how to adjust model quirks and looks.

So I shared this know-how with the team through specifications and other documents, and our goal was to raise the average value of the live part as a whole. If there were bad images, the immersive feeling would diminish, so we tried not to show any broken images.

Sakaide: I didn't have to say too much about the details, because the finished product was so wonderful. From the first phase, were there any points where you thought, "This isn't going to work"?

Kuroiwa: The part that we had been struggling with since the first season was how to make the singing part of the performance, so I was hoping we could pursue that in the second season.

Sakaisai So it was almost like you were hand-working the parts?

Kuroiwa Yes, that's right. The body was recorded by motion capture, but the faces were made from scratch, so the animators hand-touched everything from where to make the eyes pop to the timing of the lip-sync.

Q: Can all parts of the face be moved freely?

Kuroiwa: All parts can be moved. There are some restrictions, but this time we removed those restrictions (laughs). (Laughs.) That's why I can adjust them as much as I want.

Q: When you draw by hand, you can distort outlines and so on, but is it possible to express such deformation?

Kuroiwa: Yes, that is also possible. That is why I have been working on it as if I were making CG drawings, and I have been conscious of that since the first season.

Sakaida: You also do quite detailed movements. Like when you open your eyes for a moment and then squint.

Kuroiwa I love that kind of facial expression (laughs).

Saka I was actually going to talk about it at the meeting, but I thought I wouldn't be able to follow it all the way through. In fact, when I see various idols perform live, they are really good at making facial expressions, and they have a preparatory look a little before to make the most of those expressions. I thought it would be great if this kind of movement could be expressed in anime as well, and then they came along, so I was surprised (laughs). That made me feel that I could leave everything to them.

Kuroiwa: We had fun doing it (laughs).

Q: Did you also refer to videos of actual idols?

Kuroiwa: Yes. I used the live performances of Momoiro Clover Z and Halo Pro as a reference. Also, from the first season, the character designer, Ms. Kanzumi Fukagawa, told me that when she smiles, her face breaks up and her eyes narrow and her mouth opens wide. In the second season, I tried to express that well, so I moved all the parts larger.

In the second season, we included many scenes of stirring, which CG is not good at.

Q: First of all, the director gives you an order to create a scene like this, doesn't he?

Kuroiwa: Yes. From there, we had specific CG discussions about the scene, and we used a large space within the company for the discussions. Since many of us were working remotely due to the Corona Disaster, we thought about how we could share the director's image, so we installed an application called "Virtual Camera" and used a model to actually have the director move the camera.

Sakaide We used the entire room as a stadium field, placed a model of the PA booth in the middle, and shared how we would move the camera from there to the stage, actually showing it on the screen and explaining, "This is the kind of camera work we are thinking of doing.

Kuroiwa: I shared the images created by CG with the actual workers and told them, "This is the kind of camera work I want you to do.

Sakaide But it was also difficult to move the camera. I couldn't figure out the angle of view! (laughs). (laughs) However, we were able to discover camera work that we had not envisioned in the storyboards, so that was a good thing.

Q: By the way, when you were working on the CG, were you at the stage where you had already finished shooting the motion capture?

Kuroiwa: Yes, it was finished.

Q: Did you participate in the motion capture recording process?

Kuroiwa: Yes, I participated. Motion capture involves attaching markers (points that follow the model's movements during motion capture shooting) to the dancer's head and body, and then using cameras to track and digitize the dancer's movements. (The number of people recording at the same time was also a factor.

Also, in the second season, there was a scene in episode 4 where Tae Yamada plays the drums, so we had the drummer (Ms. Chisato Kawaguchi) actually play the drums for us. We also brought in the drums that she normally uses for the scene.

And this time, we happened to have an animator who has been playing drums since she was 9 years old, so we decided to leave this scene to her. Also, the person who created the drum setup (the mechanism to move the drums) was another person with drumming experience, so I think we were able to recreate movements that were closer to reality.

Sakaisai: How did you manage the gap between the actual drummer and Tae's reach?

Kuroiwa: The studio that produced the motion-capture video completed the gap and delivered it to us. After that, the animators make adjustments.

The drum set is modeled, a mechanism is created for the CG drums to move, motion capture is taken with the drummer's actual set up, the drum set is reassembled to match the drummer's set up, and the materials are then combined by the animators who have experience with drums. Then, the drummers would set up a drum set to match the drum set, and animators with drumming experience would combine these materials to create the animation. You have put a lot of work into one song, "Gekko Survive" (laughs).

Kuroiwa: The instrumental performance was a scene that did not exist in the first season. Also, going back to the change from the first season, before the second season started, the director told me that mobs would be very important in this season. I was careful to make the audience feel a sense of unity with the live performance.

Border I was working on the series knowing that there would be a stadium live performance in episode 12, so I also told Mr. Kuroiwa that I wanted to do this in the end and asked him to prepare mobs based on that, and if there were any problems, I asked him to squash them.

Kuroiwa: So episode 12 is the culmination of that. It is a good thing that we learned how to do it in the final episode and found it easier to do it this way.

Q: You mentioned earlier about advice on facial expressions, but what else did you and Mr. Fukagawa exchange?

Kuroiwa: In episode 7, a new character, YUZURIHA Maiyu-chan, was introduced, so I made a model and asked her to put facial (face and expression) marks with a pen, and then I made a collection of default expressions accordingly.

Also, in the live scene, we had her put bites on one or two cuts, and then we created the facial expressions in CG accordingly.


I have the impression that the expressions in the live scenes were always cute, but you were very particular about such details.

Kuroiwa: During the CG process, we decide which parts will be CG and which parts will be drawing.

Q: Even so, you were able to get very close to the camera, weren't you?

Kuroiwa: If it's just above the shoulders, it's fine! It's like that (laughs).

(Laughs) Sakai: We experienced the long shot of Sakura Minamoto in the 12th episode of the first season.

Q: I thought it was becoming almost impossible to distinguish between the two in terms of drawing.

Kuroiwa: The most important thing was to control facial expressions. In the case of mocap, the movements of the body are very detailed, so we had to make sure that the facial expressions were as fine as the body movements. I instructed them to include detailed changes in the face so that the facial expressions would not remain constant. So the amount of information on the face must be as detailed as that on the body.

Sakaide: The face is always a little bit out of place, isn't it?

Kuroiwa: Yes. When creating facial expressions, we go as far as we can by maximizing the use of the 3DCG model, but that is not enough, and it inevitably ends up looking like a doll.

There is the 3DCG work screen and the actual look from the camera (the final animation screen). On the work screen, the face looks messy, but when viewed through the camera, it becomes a cute look.

Sakaida: So you distort the model so much to get a look that is cute when viewed from the camera.

Q: So you make adjustments in anticipation of the final animated image?

Kuroiwa: 3D has a three-dimensional effect, but for example, in the case of Aori, the jaw can be very sharp. Therefore, adjustments are necessary in order to maintain the design of an animated picture. I was conscious of making sure that the design would look natural and familiar in an anime context.

Sakaida: Since the first season, one of the areas in which CG was not our forte was in the stirring scenes. That's why we didn't use them very often, but in the second season, we included a lot of aori from the storyboards (laughs). (Laughs.) I thought that if they said it was impossible, I could just make it into a drawing or think of another way to do it, but you made it all happen. ......

Kuroiwa: I like aori (laughs).

Sakai I think aori is soothing from a pictorial point of view, or it looks cool. When I watch idols' live performances and music videos, aori is cool.

Kuroiwa: That's right. It is also from the audience's point of view. I didn't want to create limitations in that area.

Q: From the live performance in the first episode, there are a lot of cuts of aori, aren't there? Did you create them in 3DCG while imagining how they would turn out?

KUROIWA: In my case, I first draw a picture as an atari. This is how I wanted it to look like. Then I create the cut while looking at it.

I create the work while keeping in mind that it should look like a hand-drawn picture.

Rather than making it right, I try to focus on impact, coolness, and cuteness.

Since we have already talked about the first episode, I would like to look back at them in order. In the first episode, "Revenge," the opening cut was cool, and it was an interesting scene with the audience fighting in the foreground and the live performance in the background. I heard that the director thought this was going to be a tough scene.

Sakai That's right. I had a CG model of the live concert hall itself, and then had the audience placed in the CG model and layout. From there, I asked them to come up with a guide for the drawing of the brawl scene, and I drew the pictures according to that guide. So the brawl scene was ultimately resolved by the drawing, but it was a cut involving CG.

Q: So it was a combination of CG and hand-drawn material?

Sakai: That's right. The stage was divided into CG and the audience in front of the stage was celluloid, so the CG was used to guide the camera work for the audience in the front of the stage.

Q: By the way, was the lighting on the stage processed by photography?

Sakaide: Yes, it was processed by photography, but how the light shines and where it shines is processed by CG.

Kuroiwa: Yes, we used CG to make the lights blink, etc., and then handed the material to the cinematographer.

Q: What about Junko's backlighting, for example?

Sakai: I had the CG do the shadows there. There are times when I ask the director to put on ordinary shadows and make them look backlit during shooting, but for this scene, I had the intention from the beginning to make the scene look like this with the light on my back, so I had the CG do everything from the shadows.

Q: So the director had that much of an image in his mind.

Sakai: Sometimes I have an image of the scene in mind when I'm making the storyboard, and other times I have an idea of what I want to do during the meeting, but generally I have an idea.

Q: There were a lot of camera cuts.

Sakaide: I put the camera in various places in my mind and made a storyboard while imagining how it would look like, but when I actually made the CG, it didn't turn out like this at all! But when we actually made the CG, it didn't turn out like this at all!

We could do camera work that could not be done in real life (laughs). (Laughs) We couldn't do shots from above in a small live music club, or fast camera movements, in real life.

Sakaida: Before making a storyboard, I was suggested, "Why don't you use a virtual camera to decide on the camera work before you start making it? I actually thought that was a good idea, but I wanted to include some false images. If it were true, it would look like this, but I wanted to make an image like ...... where everyone is in the picture with an impossible telephoto compressed perspective to give priority to coolness and mood (laughs). (Laughs.) But that would never come out if I were actually shooting with a camera, so I was really just daydreaming about it.

I had seen footage of actual idols, so I had a clear image of how this kind of image could be created, but I also wanted to include false images. I wanted to include false images as well, so I wanted to create a sense of depth through the connection between the two shots, even though the placement of the shots before and after were completely different.

Q: It looks like you used lenses in ways that would not be possible in real life.

Kuroiwa: I actually used different lenses for the characters and the background.

Q: So you were able to create live images that were possible only with animation.

Sakaide: I thought it would be interesting to make it that way. I wanted to emphasize impact, coolness, and cuteness rather than making it right.

Recommended Articles