Disney Wish: Disney's 100th Anniversary Film

TLDR: Character Modeling Supervisor on Disney's 100th anniversary feature, managing a team of seven through a fully remote production. Before production started, I was handed a multi-month manual cataloging task: tag and classify every character modeling asset from Encanto and its short. Over 2,500 assets, conservative estimate of eight weeks, and timing that would have eaten directly into the visual development window for Asha. Instead of grinding through it, I wrote a Python autotagging tool that classified assets directly from Maya scene contents, and the work closed in under a week. I used the freed time to develop a 3D interpretation of Asha, which earned the visdev assignment with Art Director Bill Schwab. The tool stayed as department infrastructure. The low-res cage plus creasing workflow I had introduced on Encanto found its permanent footing here and became standard practice in the character department.

DETAILS

• Studio: Walt Disney Animation Studios

• Project: Wish (Disney's 100th anniversary feature, released November 22, 2023)

• Role: Character Modeling Supervisor (shared credit with Alena Wooten-Loftis)

• Team: 7 character modelers, fully remote production

• Art Director: Bill Schwab

• Tools: Maya, ZBrush, Python

• Years: 2021-2022 (joined post-Encanto, transitioned to Meta approaching final sprint)

• Shipped: Python autotagging tool (2,500+ assets cataloged in a week), 3D visdev of Asha, low-res cage plus creasing workflow adopted department-wide

• Links: IMDb

THE CONTEXT

Wish was Walt Disney Animation Studios' feature for its 100th anniversary. A celebration film with the weight of the studio's entire history behind it.

I joined as Character Modeling Supervisor, managing a team of seven Disney character modelers, and before starting production, I had the huge privilege to collaborate again with Art Director Bill Schwab after our successful partnership for Encanto. The production ran fully remote, built from scratch after COVID. Unlike Encanto, where the character team had formed in-person and carried that connective tissue into production, Wish had to establish its beginnings from the best practices learned from Encanto in remote workflows, annotation tools, review rhythms, and creative alignment entirely over video and file sharing.

THE CHALLENGES

Right as I became supervisor and the visual development window opened, an urgent request landed: tag, classify, and catalog every character modeling asset from Encanto and its short (unannounced), down to each variant, accessory, outfit, and character prop. The studio's internal repository needed it, and the work had no clean path.

It was a manual task. Every model opened one at a time, inspected, labeled against the established taxonomy rules operating at the studio, filed, saved, and started over. Over 2,500 assets to tag. Conservative estimate: eight weeks, probably more, of tedious manual unexciting labor. No dedicated tech support and no pipeline investment were available to compress it, since the studio was running Strange World in full production at the time and every technical resource was committed to that show.

The timing was the real problem. Eight weeks of manual cataloging was eight weeks out of the visual development window for Asha, Wish's main character. Preproduction of visdev exploration doesn't pause, it can't, so if I wanted to get involved in early visual development, I needed to find a way to expedite that task somehow.

THE APPROACH

Building the tool

Instead of facing the task as torture, I framed it as a challenge. Instead of grinding through it, I focused on its challenging and problem-solving component. Challenge accepted! This was mid 2021, and vibe coding or any notion of ChatGPT or Claude didn't exist at all, so the fun was granted. I confronted the problem the old school way, rolled up my sleeves, sat down in front of VS Code (not even Copilot), and wrote a Python auto-tagging tool that classified assets by reading each Maya scene's asset contents directly in batch. Asset type, variant, accessories, garments, props: the scene structure already contained enough signal to classify automatically by geometry name, so that was a starting point, then a layer of programmatic interpretation on top, then funneling and adapting the data to the studio's established taxonomy, and all presented to the user, me in this case, in a UI that allowed me to make edits on the freshly auto-tagged assets in bulk, for final delivery to Dpix, the studio's content and asset library catalog.

Inside a week: 3 days for coding the tool, half a day for editing and supervising the tagging, and the work was done. All character assets from both productions were tagged and delivered. I felt very proud of the work done, the success of my self-established challenge, and my own execution time.

Additionally, the tool wasn't only a one-off script. I cleaned it up so other show supervisors could run it against their own libraries. What started as a personal bottleneck turned into department infrastructure that outlasted my time on the show.

Earning the visdev assignment

With cataloging collapsed from months to days, I used the reclaimed time to immediately jump into the anticipated collaboration with Bill Schwab on Asha. Bill's 2D designs were already in circulation, vibrant and unique character designs full of life and personality. Initial visdev concepts at early stages of the project are always a treasure. They carry the unique style signature of the concept artist. Their own voice, funneled toward the director's vision for the character. Additionally, Bill's visdev style is so fresh and dynamic, it's always a joy because it opens up space for collaboration and interpretation, rather than pure execution. This is the sweet spot for a 3D visdev modeler, because it allows the 3D artist their own interpretation, their own vision of what the Art Director started envisioning. This opens the door for a partnering collaboration. The 2D art inspires the sculptor, the sculptor's interpretation lets the Art Director push even more, or explore other options, or even challenge the Art Director, since lines can be expressed in volumes in many ways. Again, it was a nice collaboration between Bill and me. Sculpting, sharing the result with Bill, taking his suggestions, guidance, and draw-overs back into the sculpt, iterating until the silhouette settled and the character materialized.

Wish's visual target was a world inspired by watercolor illustration, deliberately flat and stylized, painterly, and graphic but without compromising the shapes too much. Ultimately, it was the movie celebrating the 100th anniversary, so the design language had to carry the studio's signature style.

Production

Team of seven fantastic Disney modelers, top notch crafters and fully remote. At that moment, my job was purely vision alignment: funneling what the directors and Bill needed emotionally from each character, and how that applied to shapes, forms, and silhouettes. Then translating that into direction clear enough that each artist could bring their own craft to the work. Remote made that harder sometimes, but at the same time allowed the limitations to flourish as a medium for an easy live drawover process, even during Zoom calls. A lot of small calibrations that would have happened at someone's desk in a studio hallway had to become explicit conversations over the shared screen.

During production time, I continued advocating for the low-res cage plus creasing workflow I'd pushed on Encanto. This method allowed the team and myself to focus more on the shape work, less concerned about topology cleanliness, since most of it would come almost for free due to the nature of its workflow, especially when combined with a solid universal base mesh and its very specific topological landmarks. On Wish it found its permanent footing, and the rest of the character department adopted it and benefited from its rewards during this production.

At some point approaching the last sprint of the project, I had the opportunity to join Meta, to face new challenges and step up, not only on project complexity in personally unknown territory and field, but also on my career as Avatar Art Lead. Due to this departure, I had the privilege to share my role and supervising credit with my talented friend, Alena Wooten-Loftis, who carried, guided, and took care of the team for the remaining work through to final.

Wish released November 22, 2023, as Disney's centennial celebration. One last bow for this incredible project, and for my time at my dream job at Disney, was the opportunity to celebrate its Wrap Party, not only with my friends and colleagues, but also with my older girl Julia, at the time 6 years old, born during the production of Ralph Breaks the Internet (Wreck-It Ralph 2), credited there as a production baby, and my +1 at the Wish Wrap Party. A forever Disney memory!