WHO’S OUTSIDE
WHO’S OUTSIDE is an experimental audiovisual NFT collection.
Each token contains a unique, semi-generative text (read by a Google AI voice), and is accompanied by an animated person on a custom background.
Images of the people and parts of the background were sourced via vintage Google Street View footage of downtown Los Angeles.
VISUAL COMPONENTS
Each animation was created from photos of people captured c. 2007-2011 in downtown LA via Google Street View. They were animated using an educational web application for children.
The backgrounds for each token were created by combining asymmetrical cutouts of outdoor scenes in LA (Google Street view c. 2007) and hyper-zoomed images of painted canvas (taken from Bradley Boboc’s oil paintings, which often incorporate images sourced from Street View).
The animation for each token was combined with a background using OpenCV in Python. The resulting video creates an immersive experience; part miniature film, part dynamic portrait.
TEXT
The text for each token was generated from handwritten fragments. These sentence components were combined into an aleatoric narrative via Python script. Each person’s narrative introduces you to their preferences and memories (“I am a connoisseur of Stephen A. Smith’s hairline and YMCA basketball games,” or “My earliest memories are of Patty Smith’s body odor and Superballs.”), and one of their favorite quotes (categories: excerpts of John Cage’s Silence, “Cageisms”; Rich Piana quotes, “Pianaisms”; “Bushisms”; Lil Wayne lyrics, “Weezyisms”; and fond family sayings, “Midwesternisms”).
Having provided you with their backgrounds, they leave you with their latest great realizations:
- “Giving my dog canine Klonopin wasn’t going to solve my feelings of existential dread”
- “Academics’ obsession with Eminem lyrics won’t solve the higher education bankruptcy crises”
- “Hell is Alanis Morisette being forced to relive Jagged Little Pill every day of her life”
- “I find the fleshy pink color of ham slightly disarming”
This departing wisdom, or the “hook” for each token is unique and composed by hand. The final texts were generated via custom Python script. An automated MadLib, if you will.
AUDIO
Each token’s text is read by an indeterminately selected Google Text-to-Speech voice AI. The sonified text provides contrast to its written connotations: The AI voice delivers cynical takes, such as “People who smoke Malboro lights don’t love themselves,” in a perky and upbeat manner, seemingly unaware of the meaning of the words it recites. At other times, it struggles with the pronunciation of avant-garde classical composers such as “Xenakis” or “Radigue,” while perfectly executing phrases such as “Sobe Lizard Lava” and “piss paintings.” The combination of the AI recitation with the human-composed text leans on the border between absurd and sincere, real and fabricated. Juxtaposed against the dreamlike setting of the animations, the text creates a surreal atmosphere in which the speaker jumps, boxes, and sprints.