EverSong

EverSong is a sound installation concept where it is possible to hear the entire lifespan of a person in a song in realtime. It features a 5-year-old singing a lullaby, but by the end of the song they have gradually aged to 95. EverSong creates a universal experience that explores the fragility of life and its circularity, inviting people to celebrate the moment we are in right now. It questions our legacy as humans moving into an increasingly digital age.

I am collaborating with Mark Hanslip to realise this work. For the past six months we have been working with DDSP-SVC, a newly-developed process which builds on Google Magenta’s DDSP combining it with a diffusion-based vocoder.

We are currently recording 18 vocal datasets with people aged 5 to 95, and creating trained models my voice can be mapped onto in real-time; making it possible to hear myself in the past and future. For context, Holly Herndon’s Holly+ uses voice transfer. EverSong develops this concept, making it possible to hear a voice gradually age in real time. I am currently researching and composing the contemporary lullaby that forms EverSong. 

EverSong will initially be experienced in a tunnel or along a pathway, outdoors, installed sustainably in nature, with playback sensors and lights that follow people as they walk along its path. It will initially tour and end with a permanent home in an ancient site.