‘Demon Slayer: Infinity Castle’ Cinematographer Yûichi Terao On Developing The Visuals Of The Infinity Castle “To Withstand The Big Screen”

With how popular the Demon Slayer anime and the previous film Demon Slayer: Mugen Train have been, it’s no surprise that the Demon Slayer: Infinity Castle was met with just as enthusiastic a reception. As the cinematographer for both the series and the films, Yûichi Terao worked to blend 2D characters with 3D environments and action to create dynamic sequences the series is known for.

Taking place after the events of the fourth season, Demon Slayer: Infinity Castle is the first film in a trilogy that shows the Demon Slayer Corps’ final assault on the Infinity Castle, where Muzan Kibutsuji and the remaining upper rank demons reside. For Terao, the film presented a chance to build 3D spaces with unique physics and apply them to 2D characters.

DEADLINE: Compared to the previous film, Demon Slayer: Mugen Trainwere there any big differences in your job as cinematographer?

YÛICHI TERAO: One of the biggest differences is the background of the film. We had to make the entire Infinity Castle because there are so many scenes where everything but the characters were rendered entirely in 3D, which hasn’t happened before, and what we can do with the camera changes completely. As a stage, the Infinity Castle doesn’t really have an up or down and gravity can shift throughout the battle.

The added challenge to that is the characters in our film are all hand-drawn. If you have 3D backgrounds and 3D characters, all you have to do is place the camera and everything is going to work and look the way you want it to. But now we’re dealing with 3D backgrounds and environments, but 2D hand-drawn characters that are animated. So, we have to make sure the camera and the character’s animation match perfectly or they’ll slide around.

DEADLINE: Was it difficult to find the right camera movements within the Castle, since directions and gravity can feel meaningless?

will have it: Speaking towards my domain, the digital side, I would say in 3D, we would build out a scene and start placing the cameras in different angles and through trial and error, we would find the most intense dynamic way we can show this. Part of that is because when you place a camera, there is probably stuff happening outside the frame that can come into the frame, so it’s always important that we would place the camera but look at it more holistically to see how is this going to change and what’s going to come in and out.

DEADLINE: What would you say are the biggest differences in your job from working on the film compared to the series?

will have it: When working in 3D, the visualization process or the concept art portion really expands for the theatrical version, and that’s because in our pipeline oftentimes there were certain scenes where, until we finished the compositing and put all the pieces together, we wouldn’t know what the scene looks like. In the television version, that’s because until the very, very end, until I hear the music and all the other elements that are going to come together to create that drama, I can’t decide on how the visuals are going to come together. So, I’ll be making fine tuning and adjustments until the last second.

In the case of the theatrical version, however, because the scale of the project is much, much bigger, and we have a lot more animators working on this before we go into production, we have a lot more concept art. So, there’s a unified vision of what the final screen should look like as everyone goes to work. For example, throughout the film there’s a color script, which is kind of how the colors will shift throughout the movie and how that journey is going to carry the audience. While I’m concepting what the visuals are going to look like throughout the film, even in that process, I’ll have new ideas for how to light certain scenes. That’s a little different from the TV series version, and another difference is for the theatrical version, the Infinity Castle needs to be able to withstand the big screen. So, there’s a lot of detail throughout the screen, and there’s so many areas to look at that we didn’t want any area to feel like it hasn’t been fully developed.

DEADLINE: What were some of the more challenging scenes or sequences to get right?

will have it: I would say one of the challenging scenes is Azaka versus Tanjiro. In this battle sequence, one of our key animators would design action in a way where, in the span of one frame of animation, these characters would move or leap 50 to 100 meters. And in these action sequences where the character will leap a hundred meters at time, they’ll smash the environment or affect the surroundings, but the environments are rendered in 3D. So, we have to effectively use real-life physics to simulate some of the destruction. When you’re looking at this through the viewport, it looks like absolute chaos in our working environment, but when the audience sees it on the screen through the camera, it needs to make sense. So here we are taking this 2D action that defies all physics as we know it and we have to get all of our physics simulation to sort of match that. How the rubble and everything just shatters on screen needs to match what these characters are doing, and that was a lot of fun to design.