Pantheon: Reflecting on my role as Assistant Director and Visual Development Supervisor

I had the enormous good fortune to work on Pantheon for AMC via Titmouse. My role on the show was broad and I had a lot of latitude to explore. Everything from storyboards to design to post production. This is a download of my thoughts as a post mortem from my 2 years on this project.

Opportunity

Juno Lee approached me in November 2020 with the incredibly enticing offer of working on the first hard science fiction anime produced in the United States. My role was officially assistant director, with a directive to take ownership of anything that needed to be elevated visually, scientifically, cinematically, emotionally to rise to the standard of a hard science fiction. Eventually I would request a title change to “visual development supervisor” as this seemed to more accurately describe my contributions. Here was a chance to finally merge my deep love of hard science fiction, my esoteric obsessions with the sciences and engineering, and my die-hard cinephile beliefs.

Challenge

The biggest, most enduring challenge of the show was one of crew knowledge and interest. The crew was talented, bright, motivated… but came from the American animation industry. The most science fiction any of them had touched creatively was often none, and their native interests and passions were in other subject matter, mostly action, comedy and high fantasy, drawing heavily on video games and anime as their primary sources. I had to bridge the gap between their interests and what was a good faith attempt at depicting the aftermath of uploading human brains to the cloud.

So naturally, I created a document.

What Would Midori Asakusa do?

I knew that interest in anime was high among the crew, and I created a pdf for distribution that was a first attempt at communicating the mindset needed. As I was brought on after the storyboard process had already begun, I felt this needed to be communicated quickly and efficiently. The visual decisions being made in storyboards were already veering dangerously away from the spirit of the show and lacked much needed technical authenticity.

“Eizouken ni wa Te wo Dasu na!” - An anime about three girls building an anime club in a post climate catastrophe setting was making the rounds. I saw an opportunity.

1

2

3

4

5

6

7

8

EVANGELION LAUNCH SEQUENCE

EVANGELION REI PLUG EJECTION SEQUENCE

I realized soon after that while this document was motivating in some ways, there was another layer, an important thought process that informed everything I did on the show, that was not showing up in any of the visual and storytelling choices being made throughout the show. This was significant because, as a 44 minute show with 8 episodes in a season, I simply couldn’t be everywhere at once. I could hit the highlights, some of which I’ve shared below, but the writing created so many opportunities for pushing authenticity (or worse, shattering the illusion of authenticity), that I had to try and communicate these principles. Working on a TV production schedule can be very intense, with many moving targets.

The big problem was that American animation industry artists don’t have strong intuitions about what technology actually is, how it behaves, where it goes. This goes beyond needing merely sci-tech mimicry, and having a fundamental curiosity about the nature of technological progress as an expression of human ingenuity colliding with the constraints of reality. I was worried if this was something that could be taught. I explained it to my crew thusly:

First Principles

Depicting Deep Time

Technology is most often an expression of two driving forces - what a human wants to accomplish, and what nature permits. The biggest error in depicting the future, science and engineering is getting caught up in the superficial cultural symbolism of technology (holograms, mechs, space fighters, hexagons, etc) and confusing it with how technology often behaves.

Hyper advanced borderline magic technology is still based on fundamental basic physical principles and limitations.

These limitations of what is and isn’t possible in a universe are the constraints around which creativity explodes.

This is a crane:

So is this:

And so is this:

The basic idea of a crane has not changed in nearly 1500 years. Physical law constrains it, even if the above crane is computer controlled, with an air conditioned cabin and built using astounding leaps and bounds in metallurgical innovations.

This will constrain space elevators, space craft, robots, computation, even 100,000 years from now. A dyson sphere must still contend with tidal forces, gravity, orbital dynamics, and the limitations of atomic bonds in materials.

Those are the rules of existing.

A lot of subsequent energy was spent providing examples of artists that “get it”, creating targets that were visually strong and appealing while also technically accurate in unique and exciting ways. To what degree I succeeded in instilling these ideas into the crew, the reader can judge for themselves by watching the show (which I heartily recommend).

Blender

Throughout the production I had the incredible opportunity to work with my dear friend Chris Boylan, who streamlined and managed a pipeline for the creation of NPR (non-photorealistc) assets. 2D for 3D was a rarely used, expensive trick in the Titmouse toolkit, and we inverted that relationship. By the time post production was underway, Chris and I were daily demonstrating that Blender was a powerful force multiplier for inexpensively raising the production value of the show. Custom shaders were built to emulate the look of the art direction on landscapes, architecture, vehicles, even characters. Fixes and revisions that would have taken weeks to turn around were dealt with in days and often hours. It is my firm belief that CG, particularly the integration of Blender, should not be relegated to an on-the-fly decision for post, but ought to be integrated into the pre-production planning stage.

Most importantly, it let us pull some really crazy things off. The scale of shots and worlds ballooned effortlessly thanks to the procedural systems we built within blender to permit scaling and reusability.

Lessons

It was an incredible team effort, and we produced something truly unique, likely to be a cult hit. I had to choose my battles, I also discovered a very deep and personal satisfaction in scratching this kind of itch, that I only want more of. The experience confirmed a suspicion of mine that what you are working on actually matters, if you believe in a project and its goal, you will dig deeply to create. As a father of two small children, this was an incredibly challenging time, working remotely, managing a team, and taking care of my family. I was sustained by the mission. This will absolutely guide my career decisions going forward. I was fortunate to work with excellent people, who I would work with again in a heartbeat. But the pleasure and sense of purpose was multiplied by working with those excellent people towards a goal aligned with my artistic and storytelling values.

The danger here of course, is that this is a RARE sort of opportunity. I don’t know that I will be in a position to experience it again, and so I will cherish it and find myself pushing for it at every opportunity going forward. It’s worth it. The work was worth it.

SEASON 1

Animatic/Storyboard Selections

David’s Choice

This is the animatic I created for the first section of a fight in the 8th episode of Pantheon. My intention was to depict the unusual ways in which uploaded intelligences might bend space and time and indeed themselves in a conflict. This comes to a crescendo later.

David’s Sacrifice

Climactic moment in the season finale, finding the emotional balance point between a big fight and a girl losing her father was challenging and rewarding.

Chanda’s Espionage

The final result had to be compressed for time, but I had a lot of fun putting this together as a first pitch for this moment, including the temp audio selections.

Under the Hood

This was also shortened, but here is the animatic pitch for a depicting what happens “under the hood” with every weapon strike or attack between UIs. Quite a bit of research and interviewing software engineers went into this shot.

Design and Post Production

A significant part of my job after the animatic process ended was taking many of the insane shots and ideas I had developed, and “cashing the checks”. I often found myself implementing them from design to execution, which involved a close collaboration between myself and Chris Boylan, our CG supervisor, to build a unique pipeline for creating these 2D/3D moments.

“I already have.”

You can see my board above. I animated David and comped the shot.

David’s Anguish

Here I built the shot in multiple back and forth passes between TV Paint and Blender. I sculpted David’s face to get maximum precision for the subtle movements, and used it as a guide for drawing on top of. I brought the finished animation back into blender and projected the animation onto a curved surface, where I manipulated vertex weights and weld modifiers to produce the collapsing/pinching effect. This was brought back into TVpaint, and so it went until the shot was built up bit by bit.

Svalbard

This was one of the first fully CG shots I had to execute. The mountains and clouds are shaded procedurally, I built a custom shader node network to create the stylized effect matching the look of the show.

Becoming More

I designed this shot in boards, and then executed everything seen except for the BG painting at the start and the animation cels. The stretching of the room involved projecting the painting onto a 3D mesh. The design of the UI ‘skeletons’ involved a lot of thought and research, I wanted to create the impression of software objects in a debug, as though you are seeing the “rig” that controls the UI, including layers upon layers of GPT neural architecture storing their connectome and brain activity, arrayed behind them.

Brain Scan

I designed and art directed the robotic scanning apparatus, reboarded the sequence to show off its horrifying function. I did a full retake on the brain painting and comped everything together, including the laser ablation of the neocortex. This whole scene has many little touches of scientific authenticity that I researched and designed, including user interface details, anatomical and medical changes, and mechanism designs.