Futureverse, a versatile AI company introduced its new text-based music generator in a recent research paper published on August 10. Dubbed Jen-1, this unpublished model aims to overcome the limitations of current music generators, such as Google’s MusicLM. It also promises to offer higher audio quality and longer, more complex compositions than those currently available.Jen-1 will be released in early 2024 and will be capable of creating songs up to three minutes in length. It will also provide support for producers of incomplete songs by offering “continuation” and “filler” options.
The “continuation” feature will allow creators to upload an incomplete song to Jen and get suggestions on how to finish it. Meanwhile, “fill” will fill in damaged or unfinished gaps in a song. Aaron McDonald, co-founder of the company, sees Jen-1 as a way to expand the creativity of human artists.
McDonald notes that musicians throughout history have embraced new technologies to expand their musical horizons. In this sense, he sees Jen-1 as a way to democratize high-quality music, making it accessible not only to those with access to professional recording studios, but to a wider audience.