LOREM

The book Jerusalem reads, “Around them, from horizon to horizon, several different eras were all happening at once. Transparent trees and buildings overlapped in a delirious rush of images that changed and grew and bled into each other, see-through structures crumbling away and vanishing only to reappear and run through their accelerated lives over again, a boiling blur of black and white as if a mad projectionist were running many different loops of old film through his whirring, flickering contraption at the same time, at the wrong speed.”

This belting pace of the visuals keeps up with the cranking drums and contrasts with the drawled vocals. There is something very off-kilter in the mixed images of eyes, deserts and cities. LOREM’s own image research comes together with machine learning that facilitates the melting between them. Francesco D’Abbraccio, who steers the hybrid individual and collective project, explains that he has been employing machine learning techniques since 2016 to produce MIDI, audio, visuals, and screenplays. This predates the boom of text-to-image pretrained models, focusing on the dataset provided to the computer, combining them, and training models to generate new content. LOREM admits, “I am hesitant to work with newer tools like Midjourney or Stable Diffusion, as they are significantly less connected to my own data.” Sticking to his own research the musician has full control over the final result.

That dataset, and therefore aesthetics, is heavily inspired by the book we read a passage from above. Along with some other authors from the weird literature and graphic novel world including Mervyn Peake, Kafka, Dino Buzzati, China Miéville and Charles Burns. On Jerusalem LOREM lets us in on the story and his interpretation, “Characters and events weave through time, intersecting and overlapping across various ages. The novel's structure is shaped by this eternalist notion, where every moment in history is viewed as an individual, eternal node, endlessly repeating. The human perception of sequential time is merely one of the many possible arrangements of these moments.”

It’s mind-blowing stuff, investigating our experience of time. Perhaps a tree sees each moment racing by as fast as this video. Perhaps a butterfly has a slower pace of experience, due to its lifetime. Mixing references of old and new, overlaying and jumping between time and space LOREM creates a visually astounding experience that mines our ideas of linearity as well as pace of time. This is achieved by employing hypercollage. This concept was born through sampling. That’s of records, pasted fragments of photographs or writing a Dada poem. LOREM illuminates the idea, “You take things, chop them up, and glue them together with other fragments of things. A piece of the quote unquote original object—if that means anything—is always present in the new one. Now, hypercollages apply statistical tools (and Machine Learning) to traditional collages.” The video and music for At The Wrong Speed are an exploration. The music is derived from diverse sources like Italian prog rock, film sound clips, and iPhone recordings these “seeds” as LOREM puts it, have grown into a strong oak of a single. Now, we enjoy.