eleven films and eight voices on how understanding is negotiated when experience cannot be directly shared
DOWNLOAD PORTFOLIO · PDFWe asked blind and low-vision friends to describe their worlds, then translated those descriptions into moving images. In the room the films are the size of a hand; here, they play at your size.
The captions are the translators' second readings — not quotations. Some films combine the descriptions of several storytellers. They record how we listened, not what was said.
Eight channels fill the dark from eight corners; they never pair with the screens. To see, you walk to an image; to listen, you stand where the channels balance.
Combined stereo reference, 11'09". In the room they play as eight independent sources. The only words spoken are the sentence itself — 世界是你的, the world is yours — nearly two hundred times, in different voices and accents, cut into eight channels: this is the complete transcript. The sound layer is complete in itself: for blind and low-vision visitors it is the entire work.
Field research (short interviews N=15 · bilingual micro-survey N=23, July 2026) drawn one person at a time, beside representative surveys and controlled experiments that find the same condition.
Field data is a small qualitative sample; the surveys and experiments are representative or controlled. Shown side by side, never interchanged. Denominators are per figure: FIG.01 draws on the first 20 survey responses (collection continues, N=23); FIG.02 visualises Keysar & Henly 2002; FIG.03 codes all fifteen interviews.
Interview consent and exhibition-playback consent are collected separately; storytellers keep veto power over the finished films. Blind collaborators are credited, compensated co-creators; braille assets pass three fabrication gates with blind proofreaders.
C1 — a six-ending taxonomy of how everyday understanding closes.
C2 — folk theory meets the literature: the honey gap as good-enough understanding.
C3 — “forget it”: the invisible endpoint where people stop asking to be understood.
We made the eleven films by listening to blind friends describe things, then trying to draw what we heard. The story wall lets anyone try the same thing from the other side. You finish a sentence from our survey. I read it. Then a program that has only seen the eleven films tries to draw it, and I pick one of its two attempts. It usually gets something wrong. We show it anyway, with a label saying a machine guessed this.






Five of these six came from answers on our quote wall; we used them to test the style before opening the wall. The sixth is the first thing a visitor sent in. Every picture is labelled as a machine's guess. The program was trained on the eleven films only, never on what visitors write. People agree separately to being shown and to being drawn; we don't ask for names; anyone can email to have theirs taken down; I read everything before it goes up.
A proposed staging at Warehouse XI: a dim buffer, a blackout star field of fifteen hand-height screens, a low-lit echo hall with a braille sentence wall and a card taken in the dark, a bright lounge where the card is finally read.





Renders are digital visualizations from the layout data — not photographs.
Online exhibition → world.yueyao.design · source: github.com/acrux-yueyao/wis
Full portfolio → download PDF · 2160×780 · six spreads.