Voice cloning is making inroads in an entertainment-adjacent industry: the music world. In a milestone agreement, SoundLabs AI, an AI music startup, will provide Universal Music Group’s extensive roster of musicians with a range of new technologies — including a “language transposition” feature.

SoundLabs’ forthcoming feature, MicDrop, allows artists to use their own voices to create voice models. These models, also known as voice clones, will not be made available to the public. 

This stipulation is a clear nod to the concerns expressed by many creative professionals, such as actors, who fear that their likeness may be used without their permission.

Once a voice clone is created, an artist (or producer, with the artist’s consent) can modify and adapt it for different uses. For example, the MicDrop feature includes functions such as voice-to-instrument, which allows a voice to imitate different instruments, as well as speech-to-singing.

MicDrop, described as “a cutting-edge AI vocal plug-in,” also offers what SoundLabs calls “language transposition.” According to a June 18, 2024 press release by UMG, this function “could help artists release songs around the world without a language barrier.” 

Neither SoundLabs nor UMG has elaborated on the tech behind MicDrop, but it stands to reason that it will be based on past research and recent advancements in speech translation. This task has seen marked improvements since 2022, a flurry of investment activity (particularly for automated dubbing), and ongoing competition among tech goliaths such as Google, Meta, and Alibaba.

“Thrilled to be in partnership with Universal Music Group to bring ethically trained voice AI models to their artists! “ exclaimed SoundLabs on LinkedIn.

A Tight-Knit, and Tight-Lipped, Team

SoundLabs keeps a low profile, except for naming its co-founders. The CEO, BT (a.k.a., Brian Transeau), is a Grammy-nominated electronic producer, composer, and songwriter — and software developer — with a résumé that includes work with the likes of Madonna, Sting, and David Bowie. 

Lead Developer Michael Hetrick, PhD, and Chief Innovation Officer Joshua Dickinson are similarly celebrated software developers and co-founders of Unfiltered Audio, which develops creative audio tools. Rounding out the team is COO (and BT’s wife) Lacy Transeau, who also serves as an artist liaison.

Speaking in a May 2024 interview with MusicTech, BT shared that SoundLabs had closed a pre-seed round, and teased, without naming, MicDrop, saying that the company has “been diligently building some wildly new both traditional DSP and machine learning/AI technologies into a suite of plugins.”

He added, “They are totally game-changing technologies, none of which are public-facing technologies available to producers currently. […] To say I’m excited about them is the understatement of the century.”

Rolling Stone reported that UMG artists will gain access to MicDrop “later this summer.” For the time being, a beta version is not currently available, although would-be users can sign up for a waitlist.