Tuesday 21 November
is a soundscape continuously generated by one deep neural network, trained on 16 hours of soundscapes collected online from 28 locations around the world in late April 2022. The stereophonic soundscapes computed by the deep neural network compress and normalize the biophonies, geophonies and anthropophonies of the planet. They echo the terraforming of AI, whose computational infrastructures intensively exploit socio-material resources over the globe, and whose media flows generated throughout the Internet divert our attention from the planet’s ecosystems.
For ISSE 2023, I propose a static version of Deepscape: Transversal, as a 15-minute field recording led inside the deep neural network’s latent space. More information can be found in my full paper Deeply Listening Through/Out the Deepscape, presented at the International Symposium on Electronic Art (ISEA 2023).
Hugo Scurto is an artist, designer and researcher, born and based in Marseille. They completed a PhD at IRCAM entitled “Designing With Machine Learning for Interactive Music dispositifs” (2016-2019), were post-doctorate at EnsadLab (2020-2022), and co-founding member of w.lfg.ng, a post-AI music collective (2018-). Their practice consists in creating, listening and performing with learning machines, in cooperation with a plurality of people, to reveal and reshape our musical entanglements with our environments. Hugo currently pursues a research-creation residency at Locus Sonus Vitae / ESAAIX, where they are working on Deepscape: Longitudinal, a transdisciplinary practice based on deep listening and deep learning of sounds and images collected during a year in the calanques of Marseille.