Futureverse, a prominent technology and content company specializing in AI and metaverse innovations, headquartered in Los Angeles, has just made a groundbreaking announcement in the realm of music AI. The company has introduced JEN 1, an unprecedented and universal high-fidelity model designed for text-to-music generation.
Despite the significant progress achieved in various AI domains such as imagery and text, the challenge of generating authentic and high-quality music remains intricate. Futureverse has taken a major step forward by publishing a research paper that outlines an exceptionally efficient approach, showcasing superior outcomes compared to the state-of-the-art models previously introduced by technology giants like Google and Facebook.
JEN 1 has been subjected to thorough analysis against state-of-the-art benchmarks, both through computational measurements and human evaluations facilitated by an esteemed advisory board comprising respected figures from the music industry. The specific details of the advisory board members are set to be disclosed soon. The results of the evaluation are impressive, indicating that JEN 1 produces music with a perceptually higher quality score of 85.7 out of 100, outperforming current leading methodologies with a score of 83.8 out of 100. To provide a glimpse of JEN 1's capabilities, Futureverse has released a selection of early audio demonstrations that highlight the approach taken by the model.
Shara Senderoff and Aaron McDonald, Co-Founders of Futureverse, expressed their pride in the significant progress made by their team in advancing music AI. They emphasized the potential benefits for creators and the overall music industry. Senderoff, a notable figure in the music technology sector, previously co-founded the music tech venture fund, Raised In Space, alongside industry luminary Scooter Braun. She has also been a leading advocate for innovations in music copyrights, royalties, and revenue streams through blockchain and web3 approaches. On the other hand, McDonald brings over 20 years of experience in technology as an engineer, product developer, and business leader, managing portfolios exceeding $1 billion. His focus has been on empowering developers and users through Futureverse to create and engage with interoperable content and applications within the metaverse. Both Senderoff and McDonald have been investing in the web3 and blockchain space for over six years.
The intersection of text and music, known as text-to-music generation, introduces exciting possibilities to connect free-form text prompts with musical compositions. The introduction of JEN 1 marks the initial step in Futureverse's ambitious journey within a rapidly evolving sector. JEN 1 is capable of various generation tasks, including text-guided music generation, music inpainting, and continuation, while consistently delivering high-fidelity 48kHz stereo audio and maintaining computational efficiency.
0 Comments