Spectres is the debut Supersigil album: 13 experimental tracks produced over the last seven years—the 28th release in my discography but my first ever completely solo full-length LP. While working on this record, I delved deep into Jungian psychology, concepts from which became central to the project. All 13 tracks are named after archetypes—innate, universal forces that shape our experience of the world.
Spectres is also a visual art project. For each of the 13 tracks, I trained a bespoke AI GAN on consensually crowd-sourced images submitted in response to the corresponding archetype. By limiting the training time of these custom image-to-image models, I found I could achieve results that mirrored the formation of archetypal symbols as I imagine them deep in our collective unconscious: elusive yet powerful forms emerging from abstract swirls of colour. I think of this process as filtering the archetypal forces through contributors' personal psyches before channelling them back into the realm of the universal. For the album cover I trained another GAN on outputs from all the previous GANs combined, resynthesising the individual archetypal energies to forge an integrated whole.
Spectres is more than a collection of WAVs and JPGs; it is a dialogue between the personal and the universal, a multidisciplinary exploration of the forces that connect us all. The album has taken me on a transformative seven-year journey through sound, imagery and psyche. In the resulting work, individual and collective experience harmonise, revealing the profound depths of human interconnectedness.
Before its official release on 11/23, I'm rolling Spectres out across the web3 ecosystem, with drops on Sound, Zora, Lens and other emerging platforms. In a novel release structure, collectors of individual tracks gain instant access to stream the full album months before its official release. The goal is to push the limits of onchain interoperability to explore what a web3-native album release can look like in 2024.
0xa79a
Spectres is also a visual art project. For each of the 13 tracks, I trained a bespoke AI GAN on consensually crowd-sourced images submitted in response to the corresponding archetype. By limiting the training time of these custom image-to-image models, I found I could achieve results that mirrored the formation of archetypal symbols as I imagine them deep in our collective unconscious: elusive yet powerful forms emerging from abstract swirls of colour. I think of this process as filtering the archetypal forces through contributors' personal psyches before channelling them back into the realm of the universal. For the album cover I trained another GAN on outputs from all the previous GANs combined, resynthesising the individual archetypal energies to forge an integrated whole.
Spectres is more than a collection of WAVs and JPGs; it is a dialogue between the personal and the universal, a multidisciplinary exploration of the forces that connect us all. The album has taken me on a transformative seven-year journey through sound, imagery and psyche. In the resulting work, individual and collective experience harmonise, revealing the profound depths of human interconnectedness.
Before its official release on 11/23, I'm rolling Spectres out across the web3 ecosystem, with drops on Sound, Zora, Lens and other emerging platforms. In a novel release structure, collectors of individual tracks gain instant access to stream the full album months before its official release. The goal is to push the limits of onchain interoperability to explore what a web3-native album release can look like in 2024.