01/11/2022 20:00 – 23:00
Click here to access live radio player

In ‘None’ and ‘Latent Space Boradcasting’ Magnus became occupied with the vast worlds of sound in latent and liminal spaces that would normally be considered ‘silent’. Magnus Bugge recorded such sounds: room tones, barely audible field recordings, tape hiss, gear hum, and processed them in various ways to listen closer. He applied machine learning and trained a model on this desolate sound world. The sampleRNN model is ‘a novel model for unconditional audio generation based on generating one audio sample at a time’. One audio sample is basically the elementary particle in digital audio. It is a single point that is meaningless without the context of what becomes before and after, in itself it has no direction, and crucially, no sound. However, knowing the trajectory and velocity you may predict where it will travel based on an idea of what music is. In his case, the intelligence have never heard music before, only silence, and thus predicts based on this.

Magnus Bugge is a composer and artist working with sound art, electroacoustic music and related visual arts. Magnus works in an exploratory practice where works emerge from studies of concrete and abstract subjects, often inspired from science, technology and physics, in addition to purely auditory phenomena. With experience from stage art, concert, film, installation and fixed media releases, the conclusion is that media is not of importance, as ideas and works will develop in their own way according to their will, and the artist should follow. The tools, primarily synthesis, field recordings, programming, machine learning and tape music techniques are equally not important, but mere means for executing work, experiencing and observing. In addition to he’s solo practice, Magnus is part of the stage art collective Verdensteatret.