Mixtape Collective Picture
Mixtape Collective Picture is an artistic project that blends music culture and generative AI, aiming to spotlight the diverse artists that shape a mixtape and represent them in a group shot.
personal work
roles
Concept
ML engineering
year
mid 2023
NTS
GenAI
music
representation
NTS
GenAI
music
representation
NTS
GenAI
music
representation
NTS
GenAI
music
representation
Mixtape Collective Picture
Mixtape Collective Picture is an artistic project that blends music culture and generative AI, aiming to spotlight the diverse artists that shape a mixtape and represent them in a group shot.
personal work
roles
Concept
ML engineering
year
mid 2023
NTS
GenAI
music
representation
NTS
GenAI
music
representation
NTS
GenAI
music
representation
NTS
GenAI
music
representation

For this experiment, we are using broadcasts from the community radio station NTS as our base material.

NTS Radio is an independent online radio station based in London, broadcasting a wide range of musical genres and cultural content. The platform is known for its eclectic programming and for giving a voice to both emerging and established artists and DJs.

I have been listening to this radio station almost every morning for years now. It played a significant role in shaping my imagination, and its energy has been partly inspiring for this project. I remember during my last vacation in Scotland, I had rented a car to explore the Highlands. One morning, as I was driving the rental car, I was listening to The Early Bird Show with Maria Somerville while taking in the russet and brown autumnal hues of the landscape. It was a wonderful way to start the day, and it's a memory I will hold onto for a long time.

Mixtape Collective Picture is an artistic project that blends music culture and generative AI, aiming to spotlight the diverse artists that shape a mixtape and represent them in a group shot.

The artwork follows a simple concept. We start by harnessing the two NTS's live streams: Live 1 and Live 2. A dedicated virtual space is established for each live stream, updated automatically and in real-time as tracks flow seamlessly one after the other.

As the broadcast kicks off, the interface is a blank canvas. Track by track, the list unveils itself and each new addition comes with a twist. Upon the track's appearance, its artist is identified, and a full-body photograph is generated, isolated, and seamlessly added to the interface. By the time the radio show concludes, the interface morphs into a group picture – a collage of all the artists who contributed to the playlist.

While the assembly of artists on the photo could be reminiscent of collage art, AI generation takes it a step further. There’s a maintained coherence, where the array of artists appear as if they are posing side by side, interacting and collectively partaking in this transient artistic endeavor.

We could even imagine studying "sequences of artists" just like we study language today by observing sequences of words. This could reveal a form of grammar inherent to audio mixtapes.

I found it interesting to place them geographically in the photo based on the order in which their music appears in the show. This adds a temporal aspect to the image's interpretation, which could evoke Egyptian hieroglyphs. The photo would be like a page where the artists are strong symbols with meaning. We could even imagine studying "sequences of artists" just like we study language today by observing sequences of words. By studying all the NTS shows, we could realize that some artist sequences are recurring, and understand that there is a sort of grammar in audio mixtapes.