Arvind Sanjeev shared this intro video for SPIN, an AI music synthesizer that lets you synthesize compositions in collaboration with a language model, MusicGen.
SPIN looks a bit like a cross between a turntable and an MPC, but it really provides an interface for controlling the parameters for music generation and playback. A DVS (Digital Vinyl System) lets you slow down, zoom in, scratch and listen between the notes.
You can use it to create new compositions, as a simple sound synthesizer, as a playful scratch tool, or to play generative background music.
How It Works
Under the hood, SPIN takes the input prompts in the form of button presses through an Arduino Mega. This is sent via serial to a Raspberry Pi which prompts the MusicGen API.
MusicGen is a single Language Model (LM), designed to create high-quality music based on cues from text descriptions or melodies.
An MP3 file is received as the output of MusicGen, which is loaded onto a Digital Vinyl System (DVS). A transmuted Numark PT-01 and a timecoded control vinyl record serve as the turntable. The Xwax DVS package for Raspberry Pi reads the vinyl timecode through a Behringer audio driver, and the output is played via stereo speakers.
“There are a lot of amazing generative music experiments, from Dadabots’s relentless death metal streaming AI on YouTube to Holly Herndon’s experiments around voice transplantations,” notes Sanjeev. “But I realized we hit a tipping point when I stumbled upon the Riffusion music model; I was taken aback by its depth and realism, including its new update that adds lyrical voices to the output. Inspired by this, I wanted to build a platform to let me further explore and combine never-before-heard combinations of music and sounds. This laid the seed for building SPIN.”
More information on the project is available at Sanjeev’s site.
Check out SPIN, and share your thoughts on it in the comments!
This badly needs a crossfader, and it would also be nice if you could turn off the turntable’s motor and use your hand to change the speed. Otherwise it looks fun and original.
Not enthused about prompting an AI. AI advances in generative music are quite clever, but when you get down to it you’re (currently) just giving a description of the music to a black box, which pop out ready baked segments. You’re not interacting directly with the sequence, harmony, or timbre. It’s more akin to a jukebox than an instrument. None of the AI teams in this space have yet presented anything like a generative MIDI or similar sequencing tool, which is a good indicator that few to none of the people working in field are actually musicians.
I assume that’s why they chose the anachronistic record player form factor. It’s not so much meant as a tool for musicians, but as an entertainment device for music listeners that generates any kind of music they want to hear. It’s a vision of how AI might change the listening experience and the music industry.
Art project on the banality of contemporary music?
so eloquently correct Wilson !
You are clearly not an AI
“Compositions” ?
Nah, just buy two and put a mixer in-between. Treat it as a turntable that always plays original tunes.
Damn, why is everyone salty about this. This is dope, and if you made it yourself you’d be excited too.
Trash
So it’s a laptop in a fancy wooden box and the rest is just for show
yeah this sucks, none of these loops sound interesting or particularly useful as even sample material much less general listening, as I think it’s intended. Gimmicky silicon valley hypebeast simulcrum for emotionally stunted zombies without a creative bone in their body.
“ SPIN is an artifact from a future where music will be hyper-tailored to people’s tastes and preferences.”
That’s the present. Those “people” are called “musicians”
Clever and attractive interface design. Can’t figure out if the concept is genius or garbage though. When you use generative AI — whether it’s text, image, sound, or video — you are simply taking a huge shortcut by mushing together the work of other people and arriving at a blatantly derivative final product that incorporates none of your own creativity.
If you enjoy being creative, rather than rapidly producing “content” for mass consumption, this approach doesn’t seem very appealing.
No, i think it’s a step back from traditional sampling, where you have to get more involved with the creation process
And sampling is already “cheating” to begin with in some sense
Whatever skill and talent is derived from sampling is definitely short circuited by typing in a prompt of exactly what music you want
Not anti-AI; I look forward to being able to tell a generative engine “I need a brass section in C#, 3/4 time, 2 bars” and get what I need. That day is coming, and probably pretty soon – and that’s where the problem lay with this.
This time next year the technology used in this will be incredibly outdated. I like the idea, but the technology is moving too fast to justify dropping huge coin on a piece caught in the middle of it’s transitional period. I’d probably get one if it were cheap, which these clearly will be anything but!
Make it small and reasonably affordable, and you got a sure winner.