{"id":21074,"date":"2023-08-11T19:32:37","date_gmt":"2023-08-11T19:32:37","guid":{"rendered":"https:\/\/nftandcrypto-news.com\/nft\/futureverse-wants-to-democratize-high-music-quality-with-jen-1\/"},"modified":"2023-08-11T19:32:48","modified_gmt":"2023-08-11T19:32:48","slug":"futureverse-wants-to-democratize-high-music-quality-with-jen-1","status":"publish","type":"post","link":"https:\/\/nftandcrypto-news.com\/nft\/futureverse-wants-to-democratize-high-music-quality-with-jen-1\/","title":{"rendered":"Futureverse Wants to Democratize High Music Quality With JEN 1"},"content":{"rendered":"
Futureverse, an AI and metaverse tech company with one of the largest digital collectible communities in the space, published its first research paper introducing a new infrastructure model to democratize high-end music quality by fusing musical compositions with artificial intelligence (AI).\u00a0\u00a0<\/p>\n
The research paper, JEN 1: TEXT-GUIDED UNIVERSAL MUSIC GENERATION WITH OMNIDIRECTIONAL DIFFUSION MODELS,\u201d is authored by members of Futureverse\u2019s Altered State Machine innovation team, and dives into the intricacies of current audio generation.<\/p>\n
JEN 1, according to the press release shared with nft now, is a \u201chigh-fidelity\u201d sound synthesis model for \u201ctext-to-music generation\u201d \u2013 combining musical compositions and free-form text prompts to produce higher quality sound output.\u00a0<\/p>\n
The advancement of AI, as we\u2019ve experienced it so far in 2023, has already made substantial advancements with imagery and text. However, when it comes to producing authentic and realistic music \u2013 that isn\u2019t stealing the image and likeness of an artist\u2019s voice, lyrics, or melodies \u2013 there are still several challenges, which Futureverse\u2019s research paper touches upon.<\/p>\n
The research underscores the complexity of sound generation within musical compositions. Unlike speech, music requires higher sampling rates to capture its nuances, requiring a more refined training process. With AI, enhanced data processing during training requires high-quality audio that also considers the interplay of multiple instruments and intricate melodic structures.<\/p>\n
\u201cJen is spelled J-E-N because she\u2019s designed to be your friend who goes into the studio with you. She\u2019s a tool,\u201d says Shara Senderoff, co-founder of Futureverse in an exclusive interview and first-look with Billboard.\u00a0<\/p>\n
Predicted to release in early 2024, JEN 1 can form up-to three-minute songs and help producers with half-written songs by offering \u2018continuation\u2019 and \u2018in-painting\u2019. It can perform text-guided music generation, music inpainting, and continuation while generating high-fidelity 48kHz stereo audio \u2013 and maintaining computational efficiency.\u00a0<\/p>\n