Suno Launches AI Tool to Blend Speech and Music

AI music generation platform Suno has introduced a new feature called Speech, designed to generate spoken-word audio tracks complete with matching background music. Users can input a text prompt or a general concept, then describe their desired vocal characteristics and musical style. The underlying AI model then synthesizes both the spoken voice and the accompanying soundtrack into a single, cohesive audio file.

According to Suno product chief Jack Brody, the company refined the Speech feature during a month-long testing phase with a select group of early users. Suno envisions the tool being utilized for a variety of creative formats, including spoken-word poetry, guided meditation tracks, and children's bedtime stories. However, the feature remains in beta and still exhibits some technical quirks, such as occasionally rendering a requested British accent with an Australian inflection.

For content creators and developers, this update streamlines the production of multimedia audio. Instead of separately recording voiceovers, licensing background tracks, and manually mixing the elements in digital audio workstations, creators can generate fully produced spoken-word content in one step. This could significantly lower the barrier to entry for producing podcasts, audiobooks, and localized voice content.

Despite the technological expansion, Suno continues to face scrutiny regarding its training methodologies, as the company has not disclosed the datasets used to build the new model. This launch occurs amid intense legal pressure for the startup. Major record labels have filed copyright infringement lawsuits against Suno, and a Munich court recently ruled against the company, dismissing the argument that training AI models on copyrighted material constitutes fair use.