Making of the Third Green-Screen Film ‘Rainbow Mist’: Updated July 26, 2021

Following Dark Demon Beast Wanigame-Eater and The Forest of Spirits and Kappa, the third installment in my green-screen film series is Rainbow Mist. This project is based on a flash fiction story generously provided by author Naomi Mishima, which I adapted into a screenplay for film.

Adapting the Original Flash Fiction

Naomi Mishima’s original story is a 2,000-character piece, blending the fantastical and the eerie. While preserving much of its core structure and development, I simplified certain aspects toward the climax to better suit the cinematic medium.

The first draft of the script led to the rough storyboards seen here. As usual in my process, these were revised after clarifying more challenging settings, and a finalized version of the storyboards was completed.

Creating Previsualization (Previs) Footage

Once storyboards are ready, my standard workflow involves scanning them to create previs footage. Previs acts as a prototype video to visualize the final product before shooting begins.

This method helps reduce inefficiencies like realizing post-shoot that a different shot composition or an omitted frame would have worked better. By previewing shots in motion with temp dialogue and backgrounds, these insights emerge early.

For dialogue, I use synthetic voices generated by Kantan! AI TALK3. It’s a handy tool that outputs human-like speech from text input, allowing for customizable speaker options. While the story features two male characters, I’ve temporarily assigned male and female voices to distinguish the roles in the previs.

Using Travel Footage as Backgrounds

A unique secondary theme of this project is repurposing travel footage as film backdrops. This approach could streamline future production by leveraging existing material.

For Rainbow Mist, I used footage captured at Lake Shinsō in Hadano City, Kanagawa Prefecture—a body of water formed during the Great Kanto Earthquake. Its lack of tributaries suggests it’s likely fed by groundwater. Locals enjoy fishing and light hiking around the lake, and its beech forest setting matches the story’s ambiance well.

The raw footage, initially intended as casual park documentation, was reviewed and matched to the storyboarded scenes. For non-POV sequences, static frames from the footage were extracted as background stills.

Evolving the Previs into the Final Product

As previs evolves in my “Masuda-Style Super-Previs Method,” its content is iteratively replaced with finalized elements. Characters filmed against a green screen are composited into previs backgrounds, creating a mix of “live-action backgrounds + puppets” and “storyboard placeholders.” Ultimately, this evolving previs morphs into the finished film.

Additional Practical Footage

• The van in the opening sequence was shot at a friend’s repair shop with their permission.
• For the moving car window shots, I mounted a camera with a wide-angle lens on my car’s passenger seat and recorded while driving nearby.

Wide-angle footage allows for cropping during editing without concern for resolution loss, as the blurred focus in background compositions masks any degradation.

Integrating Mammoths and Crocodiles

In the original story, a mammoth and a woolly rhinoceros emerge from ancient times. I’m repurposing a mammoth model in progress from another project, while the woolly rhinoceros has been replaced by a crocodile, using a miniature from a previous film. These models, green-screened, are being composited into previs scenes.

Moving Toward Completion

Once actors are cast, human performances will replace the previs puppet placeholders. This iterative process ensures a refined product while maximizing the efficiency of each production phase.