Found Sounds #8

A few months back, I was given a mixtape—an actual, physical cassette recorded spontaneously and in real-time by someone who wanted to share some recent favourites with me. While I was listening to it, I looked into the artists that I had not known before. When I got to a booming synth-pop/wave song, I stopped in my tracks: The second or third search result was a Reddit thread in which someone asked: Is this AI-generated?

Now, I have written about my infatuation with genre in this column, but in my understanding genre is an ever-shifting aesthetic and social discourse between real human beings about something they care about deeply—and not the output of a »stochastic parrot« that mimics popular forms, as generative AI models like Suno and Udio, that always (re-)produces the generic. Frankly, I do not see the point in engaging with AI-generated music.

This doesn’t mean that I reject AI in music wholesale—there are plenty of assistive AI tools that can help in the production and even the compositional process, and artists like Jennifer Walshe have made clear that machine learning can be used in powerful and actually creative ways. Nevertheless, when the doubt crept in about a song I had just heard on a mixtape, my heart skipped a beat. Can I still distinguish between what is real and what is not?

These are questions that we have to ask ourselves more and more in our daily lives. Almost a decade after someone coined the phrase »alternative facts,« our social media feeds are now full of deepfakes and so-called AI slop, whether it is music or any other kind of media. This has unsettling consequences because it makes us suspicious of everything, it makes us question whether anything is real—whether there still is a reality that we can all agree on.

Music and art can help us navigate this challenge by providing context rather than mere content. Released on the fourth anniversary of the full-scale invasion of Ukraine, »Perseverantia« does just that. The Odesa-born, New York-based composer-performer Vadim Neselovskyi’s suite for piano and string trio for John Zorn’s Tzadik label speaks of something that cannot be synthesised: The reality of witnessing war from afar.

While Neselovskyi’s occasionally jazz-tinged compositions convey their message through music alone, »Ukrainian Field Notes: Sound, Music & Voices From Ukraine After the Full-Scale Invasion« is a work of words. Journalist Gianmarco del Re had already published a compendium featuring 170 interviews alongside a music compilation through система system in 2023. This vastly expanded edition on Velocity Press now features 300.

With different means and through different media, Neselovskyi and del Re and his interview subjects shed light on a reality that lies outside of music but is inherently connected to it, whether as part of its compositional and production process or even an obstacle. Another example of the latter case is »Roux Ga Roux« by Chem XP for the Berlin-based Noise à Noise—an originally Tehran-based label that is now operating from Berlin.

The album’s release has been postponed »out of respect for the people of Iran.« With its intricate rhythmic patterns and ominous textures, it nevertheless sounds like a befitting soundtrack to the horrors that have unfolded in Iran after mass protests that were squashed as brutally as perhaps never before—all while many Iranians had to watch from afar, waiting for weeks for messages from their friends and families back home.

Diasporic Iranian musicians have always commented on the reality on the ground in their home country in different ways. Kamancheh player and sound artist Saba Alizadeh’s last album »Temple of Hope« dealt with the months-long Woman, Life, Freedom protests in 2022 and 2023. However, his forthcoming album for Berlin’s Karlrecords strikes a more sombre note: The two pieces on »Rituals of the Last Dawn« are long and mournful.

The work of Noise à Noise and Alizadeh is shaped by being removed from the realities of Iran. Also »Santur & Kamancheh for Electronics«, released through Discreet Archive, was not recorded in Tehran. This improvised dialogue between Maryam Rahmani on traditional Persian instruments and David Esser, contributing electronic sounds as well as field recordings, took place at Berlin’s Kühlspot Social Club.

Practices of improvisation, live processing, and simply musicking together in real-time have always served as a means to put us in touch with the integrity of time and space, with other people, and thus reality itself—far beyond the walls of any venue. This is also the case in regard to »Eavesdrop Festival 2024,« a live compilation also released on Karlrecords which was recorded at silent green over the course of two days.

Curator Jasmine Guffond—also represented with a short multi-channel installation piece fittingly called »Approaching Chaos«—had invited a slew of artists such as famed mastering engineer Rashad Becker, turntablist Mariam Rezaei, the duo of vocalist Audrey Chen and noisenik Hugo Equinca, and guitar heroine Nina Garcia. All proceeds from sales are donated to charity organisations providing medical aid and facilitating food sovereignty in Gaza.

The mediation of far-away atrocities is also a theme which Liz Allbee picks up on her »Breath Vessels« album, largely performed on instruments that the Berlin-based trumpeter had designed and built. The first and longest piece, »Elegy for the Lost at Sea,« was inspired by the 2023 Messenia migrant boat disaster during which more than 500 people lost their lives and its juxtaposition with the Titan submersible implosion.

The stark disparity between global reactions to these concurrent events is contrasted further by Allbee’s reflections on urban alienation on two other pieces that were inspired by an artist residency in Paris. John Eckhardt’s big-city reality in Hamburg is a different one, at least according to »Book of Riddims.« The versatile double bassist has teamed up with a slew of feature artists from his neighbourhood for this joyfully eclectic album.

Released under his Fatwires moniker, the self-released »Book of Riddims« draws heavily on all things dub, soundsystem culture, and club music. Vocal artist Mai Linh Dang, artist and actor Dennis Robert, composer-performer Shimo, and musician and comic artist Emol trade places behind the microphone while Eckhardt rolls out ominous textures, heavy bass attacks, and hard-hitting techno grooves to accentuate their performances.

There is further proof that human collaboration trumps simple prompts. »Black Seraphim« by Berlin-based iconoclast Werner Durand and Oakland-based composer-performer John Krausbauer unfolds as a half-hour-long drone, with Durand playing »invented wind instruments« and Krausbauer on violin. Released through Amsterdam’s Moving Furniture, it doesn’t reject reality, but expands it by suspending time.

For now, it seems unthinkable that any generative AI model could synthesise or even replicate records such as these; musics that do not accept the old ways of doing things or even the status quo—that are informed by the concrete conditions under which they were made. The concepts behind them and the context around set them apart from the content that floods our social media feeds and streaming platform playlists.

But what about music that is wholly or at least mostly electronic? Doesn’t the backlash against AI-generated music mirror the ones against the rise of synthesizers and drum machines decades ago? Haven’t we heard the argument that this or that is not real because it misses the human touch already in the 1950s, 1960s, and 1980s? Hasn’t the same kind of alarmism just returned in a different form?

Stefan Goldmann’s »Automation Studies Vol. 1« can be understood as a response to these questions, and not only because of its cheeky title. This massive 3CD compilation of early electro-acoustic compositions released through his own Macro comprises 17 pieces that the Berlin-based composer »generated [!] and recorded anew« with multi-effect processor TC Electronic Fireworx.

Described by the label as »electric rivers of sound flowing continuously through the same bed, while never repeating themselves quite exactly,« these pieces written between 1999 and 2001 used algorithms to create instabilities, uncertainties, and chance at a time when computers were suspected of creating little more than a flow of predictable sameness within music. Goldmann shows that it is not the tool, but the mind that uses it that matters.

Users of Suno and Udio repeatedly argue that the same applies to their prompt work, however that is a dubious argument. »OTONOMA« by Midori Hirano emphasises what separates art from slop. The Berlin-based composer also synthesises her musical pasts—one as a classical pianist, the other as a forward-thinking electronic musician—on a record that not only serves as a platform for personal expression, but also a means of communication.

The title is translated by Hirano as »the space between sounds.« She conceptualised and recorded these ten warm pieces in the post-minimalist tradition in the hope that her listeners »would move through these different spaces of sound as they listen,« as she notes in the press release. This is music as an invitation to not only lean back and consume, but to engage with the artist’s history and intentions—something that is irreplaceable.

It seems to me like we will need more of that, more music that is aware of its own context in the world, that reacts to it conceptually, and that understands itself as a means of communication that is more signal than noise—that faces, processes, and interprets reality and maybe sometimes also tries to change it. If you ask me, therein lies the true purpose of art, which is why AI-generated pieces of music will never qualify as such.

As for that song on the mixtape, I am fairly sure that it wasn’t AI-generated, that there are real people behind this music—though they seem to prefer to remain anonymous and sow some doubt about their reality. In a way, I am almost grateful for that. After all, it led me to do some research, to look and listen closely, to be sceptical, to interrogate a perceived reality. Whether we like it or not, we will need to train skills like these from here on.