What do you think of the hardest anime ever made?
You must be wondering, what do you mean by 'hardest'?
Is it technical, or is it artistic?
How about both?
I have watched a lot of anime from the 80s until 2023, and I stopped watching it after 2023, so basically, all anime that was released after 2023, except for the movies.
A movie shows the technical capability of production and shows how a director sees the world in art.
As I watched a lot of anime, the only title that I considered the most groundbreaking and the hardest anime in history is Memories (1995).
Memories is an anthology anime with three episodes; each episode was directed by a different director, but the story was based on a short manga by Katsuhiro Otomo.
The anime was produced by two studios: Studio 4°C and Madhouse. The production team consisted of legendary members; I was talking about Satoshi Kon and Kouji Morimoto as the script writer. Masao Maruyama, Hiroaki Inoue, Shigeru Watanabe, and Yoshiaki Kawajiri as Producers. Kouji Morimoto, Katsuhiro Otomo, and Tensai Okamura as Directors. Katsuhiro Otomo and Satoshi Kon as Art Director. And other legends in different sections. Basically, this anime was like a brainstorm of breakthroughs, the edge of animation at that time.
The episodes consist of different approaches by each director to tell the story.
You must be wondering, what do you mean by 'hardest'?
Is it technical, or is it artistic?
How about both?
I have watched a lot of anime from the 80s until 2023, and I stopped watching it after 2023, so basically, all anime that was released after 2023, except for the movies.
A movie shows the technical capability of production and shows how a director sees the world in art.
As I watched a lot of anime, the only title that I considered the most groundbreaking and the hardest anime in history is Memories (1995).
Memories is an anthology anime with three episodes; each episode was directed by a different director, but the story was based on a short manga by Katsuhiro Otomo.
The anime was produced by two studios: Studio 4°C and Madhouse. The production team consisted of legendary members; I was talking about Satoshi Kon and Kouji Morimoto as the script writer. Masao Maruyama, Hiroaki Inoue, Shigeru Watanabe, and Yoshiaki Kawajiri as Producers. Kouji Morimoto, Katsuhiro Otomo, and Tensai Okamura as Directors. Katsuhiro Otomo and Satoshi Kon as Art Director. And other legends in different sections. Basically, this anime was like a brainstorm of breakthroughs, the edge of animation at that time.
The episodes consist of different approaches by each director to tell the story.
- Magnetic Rose
- Stinky Bomb
- Cannon Fodder
I will not talk for each episode because, again, every episode has a different view; for example, Magnetic Rose consists of Cell Animation and 3D CGI, yet Cannon Fodder consists of Cell Animation, and the director Katsuhiro Otomo wanted the episode to use a traditional technique, no CGI.
So I will talk about Cannon Fodder.
Cannon Fodder is probably the anime that is hard to replicate because it's a ONE-SHOT anime.
What is One-Shot?
One-shot filmography features movies filmed in a single, continuous take or designed to appear that way, offering intense, real-time immersion by eliminating traditional editing cuts. Basically, in the episode, there aren't any cuts.
Before we go to the One-Shot, I want to talk about the cut transition. A cut is the most fundamental film editing transition, instantly switching from one shot to the next to control pacing and narrative.
In 2D animation, cut has a big con: it's background art. Basically, for each anime cut, the background art must be different, yet some productions cheat it with the exact position of the camera.
If there are two people in dialogue, they just use 2-3 shots of background, and after that, it just keeps going back and forth with small movement and lip movement of the character, so the burden is on the animator, not the background artist.
However, there is a studio that is so eccentric about the cut and camera position. It's SHAFT.
Well, the above isn't from Shaft, but it's from a former staff member of SHAFT; it's Tomoyuki Itamura. When I watched Yofukashi no Uta for the first time, I thought it was from Shaft because of the camera placement and the eccentric shots, and it wasn't, yet it was from Tomoyuki Itamura, a former staff member/director at Shaft.
So I break down the one scene as a scene breakdown; you can see the left side, which is the breakdown of the scene on the right.
The yellow is the camera currently used. The red line is the 180-degree rule line; the pink is Nazuna, and the other is Ko.
However, do you see how many cameras they used for one scene?
They used more than 20 shots or 20 cameras in 1-3 minutes, so basically a 2-4 second cut. Why does Itamura use this?
Basically, dialogue is the most boring part and the easiest part; you can't prevent what Itamura did if you want to show a dialogue scene, you can just use 2-6 camera angles, not 20 shots. However, the problem is how to make the audience keep following the dialogue.
Shaft has a way to do that, or I can say had a way, as now Shaft isn't Shaft anymore, but you can still see former staff or director Shaft still using the style.
Animation isn't like a film or live-action movie, where camera placement needs to be adjusted for realism. So, they put the camera in the most eccentric place or way, and they use transitions in the same style too.
For example, Monogatari and Arakawa Under the Bridge. Usually, they start with a wide shot or establishing shot where the environment is so big that it makes the characters small as an illusion. In theory, if you want to move the camera to the dialogue or the character, you usually move it slowly, basically 2-3 shots from the environment into the character. However, Shaft doesn't do that. From a wide shot, they can immediately jump into a close-up after that wide shot again. This makes a jittering transition that catches the audience's attention.
So they used a lot of cameras to catch this jittering transition and illusion of space in the environment. It feels like you are being dragged into the conversation.
It's not just the camera; they used the best color theory.
However, as you can see in the breakdown, there are 20+ cameras, which means 20+ background arts. This is the con of too many cameras. For an animator, it's the cheap way or the easiest way; basically, you can just animate small movements and mouth movements of a character. Yet for a background artist, this is a little bit challenging because for just one scene's dialogue, they need to draw each shot's background art, and each shot of this style uses eccentric angles.
What happens if there is no cut? It means it's just one piece of background art, so it's easy, then?
No, it's much worse in Memories (1995), the Cannon Fodder episode.
Katsuhiro Otomo has the most eccentric way to see how he wants to direct an animation. Basically, when he drew the first scene/shot of Cannon Fodder, he had a thought about what would happen if I keep the camera? Because in his previous film, Akira (1988), he felt he had directed it with too many cuts.
So he drew a storyboard and kept the camera.
This is known as "The Impossible Shot."
Now, in this era of technology, After Effects can recreate that shot, but in 1995, that shot needed to be studied first, and it took 2 directors, Katsuhiro Otomo and Sunao Katabuchi (Assistant Director of Kiki’s Delivery Service).
Katabuchi one day just visited the studio and saw the storyboard in 1993. Katabuchi had an experience with special effects from Little Nemo Pilot: Optical Compositing. With Katabuchi helping the project, now they saw the storyboard from Otomo not as a scroll that the camera kept moving through but as a block; there were 30-40 blocks.
Basically, One Shot can be exactly not one shot, but they can use others to show the transition; it will create a one-shot illusion. For example, a live-action film, 1917, used a hidden transition to create a one-shot.
It's like a jigsaw puzzle or domino puzzle; basically, each block is shot separately, and after that, they compile them into one.
So Cannon Fodder used the exact technique; even with this technique, the background art is still long.
Background art in 2d animation will follow the camera movement if the storyboard shows the camera movement is panning, so the background must also pan. If the storyboard shows the camera is tilting, then the background must be at a tilting angle. In 2d animation, the cut used to not make the background art so long, like a scroll.
In a modern way, you can make a long background without any problem; the problem will be the size of the file. But in an analog way, this is a challenge.
The Director of Photography, Hiroaki Edamitsu, stated that when the material (background) arrived, he lined up the BG on the floor to see how it would be connected, but it didn't fit in the room.
So he needed to find a way to make this background scroll fit the photograph; he tore down/cut the background into blocks.
The worst way: he needed 3 days, 15 hours a day, with 90 minutes to record to take a single second of animation, and the episode was 20 minutes long. The entire project is all or nothing, because you can't retake a single shot in this way. If you watched the episode, you would notice there are several shots that had different brightness; it's because of the room light.
Back to "The Impossible Shot"
That shot was used in 3d assets, but not in a digital way.
To create the impossible shot, Katabuchi and the CG artist Hiroaki Ando scanned the cells and background into their computer; after that, they combined them with 3d assets. However, the results were usually printed back onto physical film and spliced with the traditionally shot sections. Sometimes, Katabuchi printed the paper, cut it out with scissors, and pasted it onto the cells by hand; after that, they were photographed.
Katabuchi stated, it's not like something computed and generated them; we decided them by hand, based on what we saw with our own eyes and our own color scene.
And they achieved it, the only One-Shot Anime in big production, not one-shot manga, but ONE-SHOT.
The workflow of this project, Cannon Fodder, was being kept, learned, and expanded by Katsuhiro Otomo Team for the next project, Steamboy (2004)
Every field has the same challenge to achieve One-Shot animation.
3D, the challenge is the environment and heavy computers; they need fast computers to render them. Imagine you have an entire city's assets, and each animator needs to have the same fast computer to animate.
The problem of 3d animation and games is 3d animation isn't created for optimization, but created for looks.
If a 3D director wants a continuous shot sweeping from a character's bedroom, out the window, and across a sprawling futuristic city, the computer often has to hold that entire, high-fidelity world in its active memory. The burden is on the animator; if the scene file contains an entire unoptimized city block of millions of polygons just to accommodate the camera's unbroken path, the software will crawl to a halt, making the actual art of animating agonizingly slow. In 3D production, the character and the environment are in different file path, after that the layout artist/animator use link to combine it, basically, if the file master is changed, the environment or the character file will be changed too, but the animation is still the same. It can be used as a tool.
If the shot starts with a bedroom, the modeler needs to model the bedroom first in the master file; after that, the layout artist or animator uses that bedroom model to animate with the character. As the animator animates, the modeler can expand it.
The animator can start blocking out the character's movement in the bedroom while the environment modeler is still building the futuristic city outside the window. As long as the file paths remain unbroken, the scene updates dynamically.
Stop Motion, the challenge is the environment; they need a big space. To build an entire episode or movie, they must have a big space to build a diorama or miniature. Secondly, the scale must be the same; if the character is a 20-30 cm puppet, then the diorama must have the same scale, so the house must have a 40-60 cm height or higher based on the scale 1:3 or 1:5. Studios like Laika literally have to rent out massive warehouses just to physically house the dioramas for their sweeping shots.
And the animator's burden is so high, basically, the animator needs to animate the puppet in the set; if they bump the diorama or the miniature, it will disaster.
I had a stop motion project, camera and diorama are the big challenges, for example, if I was in the process of animating, and one of the guys bumps the camera, then that shot was useless. I needed to animate it again. The illusion cut or transition in stop motion will hard to find than 3d or 2d.
One-Shot or continous takes are hard to achieve, this is why Memories (1995) the Cannon Fodder episode, I see that episode as the hardest anime in history, because the episode can't be replicated, it needed to be studied by two directors, and it used computer, where computer was still slow in that time, and blended it with traditional cells animation by printing the assets.
The 80s and 90s are the golden era of anime, which shows how small studios or directors see art or animation in a different way.
This is why TV animation and OVA animation at that time had different quality; TV animation at that time is cheap and bad quality than the OVA animation.
OVA animation at that time was like a playground of studio, to try and achieve the technique. However, now it's different, with the rise of production companies, the studio has a low take on creativity in it. Some director prefers to direct short animation rather than TV animation, because of the lack of creativity in it. Even the A-1 studio lost 1 million dollars.
The loss is attributed to high production costs for top-tier animation and a production committee model where the studio receives a fixed fee rather than a large share of the direct, massive streaming profits, which instead go to companies like Aniplex. So now, the production company has big shares, and the studio who creating it has no shares.
I can talk a lot about how the 80s and 90s anime as golden era of anime technical way and the difference of nowaday, but for now, I think this thread will be over right here.
So I will talk about Cannon Fodder.
Cannon Fodder is probably the anime that is hard to replicate because it's a ONE-SHOT anime.
What is One-Shot?
One-shot filmography features movies filmed in a single, continuous take or designed to appear that way, offering intense, real-time immersion by eliminating traditional editing cuts. Basically, in the episode, there aren't any cuts.
Before we go to the One-Shot, I want to talk about the cut transition. A cut is the most fundamental film editing transition, instantly switching from one shot to the next to control pacing and narrative.
In 2D animation, cut has a big con: it's background art. Basically, for each anime cut, the background art must be different, yet some productions cheat it with the exact position of the camera.
If there are two people in dialogue, they just use 2-3 shots of background, and after that, it just keeps going back and forth with small movement and lip movement of the character, so the burden is on the animator, not the background artist.
However, there is a studio that is so eccentric about the cut and camera position. It's SHAFT.
Well, the above isn't from Shaft, but it's from a former staff member of SHAFT; it's Tomoyuki Itamura. When I watched Yofukashi no Uta for the first time, I thought it was from Shaft because of the camera placement and the eccentric shots, and it wasn't, yet it was from Tomoyuki Itamura, a former staff member/director at Shaft.
So I break down the one scene as a scene breakdown; you can see the left side, which is the breakdown of the scene on the right.
The yellow is the camera currently used. The red line is the 180-degree rule line; the pink is Nazuna, and the other is Ko.
However, do you see how many cameras they used for one scene?
They used more than 20 shots or 20 cameras in 1-3 minutes, so basically a 2-4 second cut. Why does Itamura use this?
Basically, dialogue is the most boring part and the easiest part; you can't prevent what Itamura did if you want to show a dialogue scene, you can just use 2-6 camera angles, not 20 shots. However, the problem is how to make the audience keep following the dialogue.
Shaft has a way to do that, or I can say had a way, as now Shaft isn't Shaft anymore, but you can still see former staff or director Shaft still using the style.
Animation isn't like a film or live-action movie, where camera placement needs to be adjusted for realism. So, they put the camera in the most eccentric place or way, and they use transitions in the same style too.
For example, Monogatari and Arakawa Under the Bridge. Usually, they start with a wide shot or establishing shot where the environment is so big that it makes the characters small as an illusion. In theory, if you want to move the camera to the dialogue or the character, you usually move it slowly, basically 2-3 shots from the environment into the character. However, Shaft doesn't do that. From a wide shot, they can immediately jump into a close-up after that wide shot again. This makes a jittering transition that catches the audience's attention.
So they used a lot of cameras to catch this jittering transition and illusion of space in the environment. It feels like you are being dragged into the conversation.
It's not just the camera; they used the best color theory.
However, as you can see in the breakdown, there are 20+ cameras, which means 20+ background arts. This is the con of too many cameras. For an animator, it's the cheap way or the easiest way; basically, you can just animate small movements and mouth movements of a character. Yet for a background artist, this is a little bit challenging because for just one scene's dialogue, they need to draw each shot's background art, and each shot of this style uses eccentric angles.
What happens if there is no cut? It means it's just one piece of background art, so it's easy, then?
No, it's much worse in Memories (1995), the Cannon Fodder episode.
Katsuhiro Otomo has the most eccentric way to see how he wants to direct an animation. Basically, when he drew the first scene/shot of Cannon Fodder, he had a thought about what would happen if I keep the camera? Because in his previous film, Akira (1988), he felt he had directed it with too many cuts.
So he drew a storyboard and kept the camera.
This is known as "The Impossible Shot."
Now, in this era of technology, After Effects can recreate that shot, but in 1995, that shot needed to be studied first, and it took 2 directors, Katsuhiro Otomo and Sunao Katabuchi (Assistant Director of Kiki’s Delivery Service).
Katabuchi one day just visited the studio and saw the storyboard in 1993. Katabuchi had an experience with special effects from Little Nemo Pilot: Optical Compositing. With Katabuchi helping the project, now they saw the storyboard from Otomo not as a scroll that the camera kept moving through but as a block; there were 30-40 blocks.
Basically, One Shot can be exactly not one shot, but they can use others to show the transition; it will create a one-shot illusion. For example, a live-action film, 1917, used a hidden transition to create a one-shot.
It's like a jigsaw puzzle or domino puzzle; basically, each block is shot separately, and after that, they compile them into one.
So Cannon Fodder used the exact technique; even with this technique, the background art is still long.
Background art in 2d animation will follow the camera movement if the storyboard shows the camera movement is panning, so the background must also pan. If the storyboard shows the camera is tilting, then the background must be at a tilting angle. In 2d animation, the cut used to not make the background art so long, like a scroll.
In a modern way, you can make a long background without any problem; the problem will be the size of the file. But in an analog way, this is a challenge.
The Director of Photography, Hiroaki Edamitsu, stated that when the material (background) arrived, he lined up the BG on the floor to see how it would be connected, but it didn't fit in the room.
So he needed to find a way to make this background scroll fit the photograph; he tore down/cut the background into blocks.
The worst way: he needed 3 days, 15 hours a day, with 90 minutes to record to take a single second of animation, and the episode was 20 minutes long. The entire project is all or nothing, because you can't retake a single shot in this way. If you watched the episode, you would notice there are several shots that had different brightness; it's because of the room light.
Back to "The Impossible Shot"
That shot was used in 3d assets, but not in a digital way.
To create the impossible shot, Katabuchi and the CG artist Hiroaki Ando scanned the cells and background into their computer; after that, they combined them with 3d assets. However, the results were usually printed back onto physical film and spliced with the traditionally shot sections. Sometimes, Katabuchi printed the paper, cut it out with scissors, and pasted it onto the cells by hand; after that, they were photographed.
Katabuchi stated, it's not like something computed and generated them; we decided them by hand, based on what we saw with our own eyes and our own color scene.
And they achieved it, the only One-Shot Anime in big production, not one-shot manga, but ONE-SHOT.
The workflow of this project, Cannon Fodder, was being kept, learned, and expanded by Katsuhiro Otomo Team for the next project, Steamboy (2004)
Every field has the same challenge to achieve One-Shot animation.
3D, the challenge is the environment and heavy computers; they need fast computers to render them. Imagine you have an entire city's assets, and each animator needs to have the same fast computer to animate.
The problem of 3d animation and games is 3d animation isn't created for optimization, but created for looks.
If a 3D director wants a continuous shot sweeping from a character's bedroom, out the window, and across a sprawling futuristic city, the computer often has to hold that entire, high-fidelity world in its active memory. The burden is on the animator; if the scene file contains an entire unoptimized city block of millions of polygons just to accommodate the camera's unbroken path, the software will crawl to a halt, making the actual art of animating agonizingly slow. In 3D production, the character and the environment are in different file path, after that the layout artist/animator use link to combine it, basically, if the file master is changed, the environment or the character file will be changed too, but the animation is still the same. It can be used as a tool.
If the shot starts with a bedroom, the modeler needs to model the bedroom first in the master file; after that, the layout artist or animator uses that bedroom model to animate with the character. As the animator animates, the modeler can expand it.
The animator can start blocking out the character's movement in the bedroom while the environment modeler is still building the futuristic city outside the window. As long as the file paths remain unbroken, the scene updates dynamically.
Stop Motion, the challenge is the environment; they need a big space. To build an entire episode or movie, they must have a big space to build a diorama or miniature. Secondly, the scale must be the same; if the character is a 20-30 cm puppet, then the diorama must have the same scale, so the house must have a 40-60 cm height or higher based on the scale 1:3 or 1:5. Studios like Laika literally have to rent out massive warehouses just to physically house the dioramas for their sweeping shots.
And the animator's burden is so high, basically, the animator needs to animate the puppet in the set; if they bump the diorama or the miniature, it will disaster.
I had a stop motion project, camera and diorama are the big challenges, for example, if I was in the process of animating, and one of the guys bumps the camera, then that shot was useless. I needed to animate it again. The illusion cut or transition in stop motion will hard to find than 3d or 2d.
One-Shot or continous takes are hard to achieve, this is why Memories (1995) the Cannon Fodder episode, I see that episode as the hardest anime in history, because the episode can't be replicated, it needed to be studied by two directors, and it used computer, where computer was still slow in that time, and blended it with traditional cells animation by printing the assets.
The 80s and 90s are the golden era of anime, which shows how small studios or directors see art or animation in a different way.
This is why TV animation and OVA animation at that time had different quality; TV animation at that time is cheap and bad quality than the OVA animation.
OVA animation at that time was like a playground of studio, to try and achieve the technique. However, now it's different, with the rise of production companies, the studio has a low take on creativity in it. Some director prefers to direct short animation rather than TV animation, because of the lack of creativity in it. Even the A-1 studio lost 1 million dollars.
The loss is attributed to high production costs for top-tier animation and a production committee model where the studio receives a fixed fee rather than a large share of the direct, massive streaming profits, which instead go to companies like Aniplex. So now, the production company has big shares, and the studio who creating it has no shares.
I can talk a lot about how the 80s and 90s anime as golden era of anime technical way and the difference of nowaday, but for now, I think this thread will be over right here.