This research examines ASCII art as a functional background design system in narrative-driven 3D animation — not as aesthetic nostalgia, but as a visual compression language where each character carries ecological meaning and character density encodes spatial and narrative information ordinarily requiring photorealistic representation. The research materialised as Hollow, a 3D animated short in which semi-sentient mecha restore Earth's ecosystems after humanity's extinction, perceiving the world not as it is but as compressed symbolic data. A custom multi-layered Blender compositor pipeline processes depth, normal, material, and cryptomatte render passes through ASCII filters with an ecologically assigned symbol vocabulary. Mecha characters are rendered photorealistically while environments are encoded in ASCII — reflecting the machine's self-knowledge against its compressed perception of the surrounding world. Key contributions include a multi-layered Blender compositor pipeline for ASCII rendering, an ecological symbol vocabulary assigning narrative meaning to ASCII characters, and a theoretically grounded foreground/background compositional method for representing machine cognition. The work extends existing discourse on machine vision — referencing Paglen, Menkman, Ikeda, and McDonald — into the spatial, temporal, and narrative context of animated film. Keywords: ASCII art · machine vision · non-human perception · non-photorealistic rendering · practice-based research · 3D animation · Blender