The following prompt (in bold) was input into OpenAI’s GPT-3 Beta Playground using the Davinci model. The text in italics was the output for the prompt completed by GPT-3.
Instructions for a noise music performance for one person using contact microphones: 1. Find a location where you can make a lot of noise. 2. Make a lot of noise. 3. Make more noise. 4. Make more noise. 5. Make more noise. 6. Make more noise. 7. Make more noise.
Production date: February 14-20, 2021. Filmed: February 21, 2021. Performed by Chelsea Coon.
Initial Response / Development Notes:
A performance produced in my apartment under snap-lockdown restrictions in Melbourne, Australia. I chose to make my own decision for use of a similar, but distinctly different, Stadium “dynamic microphone.”
As I understood it, the GPT-3 instructions called for 6 phases of incrementally making more and more noise. I chose to develop a performance around the idea of 6 videos that would be stitched together. To make *more noise* over time, I decided that the volume of a chosen audio track would increase on the same computer desktop that I was recording the performance on. This sonically would layer in the video with the noises I made with my materials—significantly which were projectiling confetti and bursting balloons.
“contact” was filmed in my bed where I divided 12 party poppers and silver foil balloon letters into two separate piles. I oriented my body in the middle of these piles, with a microphone in front of my knees. I utilized “Arguing Robot Bass,” a Loop Packs sample from the application “Garage Band” as the performance track. The volume was incrementally increased across the 6 successive videos that comprise this work.
My actions consisted of repetitious cycles in which I broke the poppers to expel confetti streamers across the bed, then I held and ripped open the letter balloons with my teeth. At the end of each phase, in the microphone I mouthed the word “boom.” These actions were repeated in 6 phases over the duration of the work.