Earlier this year, a startup called Futureverse raised a $54m Series A funding round, promising to “introduce AI to the metaverse”. It turns out music is going to play a part in those plans.
Futureverse has just published a research paper on JEN-1, which it describes as “a universal high-fidelity model for text-to-music generation”.
This is an area we’ve been writing about regularly: Google’s MusicLM, Meta’s MusicGen and the work of startups including Mubert, Splash and Cassette all involve text-to-music interfaces. Futureverse’s paper argues that its system is the best so far.
“JEN-1 produces music of perceptually higher quality (85.7/100) compared to the current best methods (83.8/100),” as it puts it. “More importantly, human judges confirm JEN-1 generates music highly aligned with text prompts in a melodically and harmonically pleasing fashion.”
What is Futureverse going to use JEN-1 for? It’s keeping that to itself for now beyond promising “more announcements to come”.
The company was co-founded by Shara Senderoff and Aaron McDonald. Senderoff was previously the co-founder of music-tech fund Raised In Space.