Includes high-quality download in MP3, FLAC and more. Paying supporters also get unlimited streaming via the free Bandcamp app.
Purchasable with gift card
name your price
Cassette + Digital Album
Pro-dubbed yellow tape in transparent case with folded J-Card - three versions. Edition of 50
Includes unlimited streaming of The Well-Tempered Diffusion Model
via the free Bandcamp app, plus high-quality download in MP3, FLAC and more.
ships out within 7 days
edition of 50
€7EURor more
Cassette + Digital Album
Pick up three tapes from wabi-sabi's catalog (unfortunately some of them are sold-out)
Use the message box during the order to specify which tapes you want.
3x download codes will be send as well.
Includes unlimited streaming of The Well-Tempered Diffusion Model
via the free Bandcamp app, plus high-quality download in MP3, FLAC and more.
"Pretty much all of my musical output is an attempt to answer a series of questions of the form 'what would it sound like if...?' Not the most profound motivation, but the point is that it's driven primarily by curiosity rather than, say, ego or avarice. I've been using text-to-image software for a while, where a trained computer model creates novel images based on descriptive text prompts, so the specific question that inspired this work was 'what would it sound like if I got AI to reproduce not what Bach's Well-Tempered Clavier sounds like, but what it looks like?' " (more infos about Cahn Ingold Prelog's process below)
--
In recent years, much has been written about artificial intelligence and how it can be used. In the art world, these apprehensions have quickly turned into workers' struggles which, beyond criticisms of AI as a tool, have brought to the forefront the flawed relations of production between art commissioners, work instruments and artists. This syndrome of the gargantuan appetite of the capitalist hydra within the cultural and creative industries has, however, left in its blind spot the sensitive experiences of a joint adventure with AI.
Cahn Ingold Prelog's work in The Well-Tempered Diffusion Model can be likened to that of an interpreter, an attentive translator forging a poetic dialogue with the machine. The titles of the Bach pieces used as prompts reveal a more abstract nature than the practical exercises for young musicians devised by the German composer. Here, rhythm and harmony follow tortuous patterns, gravitating around the rule rather than applying it to the letter.
The palette of virtual instruments chosen by Cahn Ingold Prelog offers the listener multiple, ambiguous universes. In N°11, a half-gamelan, half-tubular bell creature wanders over the spectrum of a string orchestra lost in the middle of some kind of Tristan Murail's process. The vocal ensemble of N°9 seems in constant contortion to accompany the singularity of the percussion. Percussion forms a capricious thread running through the entire album. They appear impromptu, disassociated and elusive, finally freed from the yoke of metre.
Cahn Ingold Prelog is an artist based in Wales. Under various aliases, he invites himself onto the parallel timelines of a fragmented musical history. Metamorphosing constraints into creative setbacks, he cultivates a spontaneous musical language that questions the place of narrative in the act of creation. The Well-Tempered Diffusion Model is just one page in a dense collection of work the artist has begun with AI, which has led him to recreate 34 of the UK's best-selling singles or reinterpret releases from the Waxing Crescent Records label.
--
"Text-to-image diffusion models are trained to understand key visual concepts with billions of tagged images. It's a similar (but much more rapid) process to how we learn visual concepts. We know what cats are, because we've seen lots of different cats and understand their common features. And if we want to draw a cat, we use our generic knowledge to create a recognisable picture of a cat, but one that is (very likely to be) slightly different from any other cat picture that we've ever seen.
If the diffusion model is trained specifically on spectrograms, however, rather than pictures of cats, fruit, people etc, then something interesting happens. Spectrograms are visual representations of sound: plots of frequency vs time, so the machine comes to learn that certain keywords have certain spectrographic similarities. It knows what generic techno or punk 'look like' in the same way that it knows what a generic dog looks like. You can then prompt the machine to produce new spectrograms in any style you can think of. In this case, the model was prompted with the titles of the first 13 parts of The Well-Tempered Clavier.
You convert these generated spectrograms into little chunks of sound, and you end up with some new music, produced by a model that has absolutely no concept of sound, only of images. The thrill of seeing the spectrograms emerge as they're rendered is second only to the thrill of hearing the corresponding audio for the first time.
The raw output is generally surprisingly listenable, given how it's arrived at, if a little alien in its gently unconventional use of rhythm and melody, but in the case of these pieces, where everything is solo piano, it gets a bit grating after a while. So I thought, why stop at one level of artifice? Why not see what else can be done with this material?
The audio was therefore converted to MIDI info and wired up to various virtual instruments. All of the original MIDI info is untouched, unedited and as-is - the only creative decisions I've needed to make are which instruments and instrument combinations to use. My goal was to employ the vocabulary of conventional electronic music, but with the strange new AI-created grammar, as if composed by some extraterrestrial lifeform that doesn't fully understand earth music."
Cahn Ingold Prelog
credits
released February 9, 2024
composed, mixed and mastered by Cahn Ingold Prelog
artwork by Mona Chancogne
printed at toner toner, screen printing workshop at Grrrnd Zero, Vaulx-en-Velin
