HYBE-owned Supertone’s new AI ‘voice changer’ lets artists change their vocals… in real-time

Supertone, as long-time MBW readers will know, is the AI voice replication software that Korean music giant HYBE fully acquired in a $32 million deal in 2022.

At the end of February, the AI capabilities of Supertone were demonstrated via a voice clone of HYBE’s own CEO, Jiwon Park, ‘speaking’ on a company earnings call.

Earlier today (April 16), Supertone launched the beta version of a new tool called  ‘Supertone Shift’.

The tool is described as a “one-of-a-kind real-time voice changer” which HYBE claims will “empower creators with new possibilities of expressions and content creation”.

Picture the scene: an artist singing live on stage, via multiple different AI-assisted voices, all switched in real time.

In May last year,  HYBE showcased the possibilities of what it can do with Supertone’s technology when it released a new single called Masquerade from an artist called MIDNATT.

MIDNATT, the alter-ego of HYBE-signed artist Lee Hyun, used a prototype version of Supertone Shift to “seamlessly” transition between his own vocals and the vocals of a female singer on the track Masquerade.

That same track was touted by HYBE as the “first-ever multilingual track produced in Korean, English, Japanese, Chinese, Spanish and Vietnamese”.

According to HYBE, MIDNATT (aka Lee Hyun) sang the track in the six languages, and using Supertone’s AI, “the pronunciation data of native speakers was applied to the track to further refine the artist’s pronunciation and intonation”.

According to Supertone and HYBE, the new tool, Supertone Shift, allows users to “immediately switch to their desired voice” from a library of ten predefined voices in real-time, with their voice analyzed for conversion “within a mere 10 seconds”.

Users can then customize the voice by adjusting a blend ratio as well as voice parameters such as pitch, joy, and reverb with “simple and intuitive sliders”.

According to HYBE and Supertone, this level of control “helps users craft their unique style while maintaining their distinct sound, perfectly aligning with their creative intent”.

You can see a demonstration of how the new tool works in the demo video below. Supertone says in the video that it “can even make a custom voice for you” if you can’t find the exact voice you want from its presets.

HYBE and Supertone claim that Supertone Shift “represents a new frontier of real-time voice conversion technology (RTVC) with latency down to 47 milliseconds and hyper-realistic voice quality, making it an optimal choice for live content and performances”.

Today’s Supertone announcement says: “Whether VTubers, live streamers, and podcasters provide elevated audio experiences for listeners, or gamers and costume players immerse their followers in interactive voice chats, creators can seamlessly shift to their alter ‘identity’ online.”

Within the demo video above, Supertone confirms that a user “can even sing in [their chosen] voice on stage”.

The beta version of Supertone Shift is currently available for free downloads as a desktop application until June 26, 2024 via its official website, with its official launch slated for the latter half of this year.

“Through Supertone Shift beta, our goal is to collect feedback from an array of creators to incorporate a broader range of voice samples and additional features to better serve various content creation environments.”

Kyogu Lee, Supertone

“Through Supertone Shift beta, our goal is to collect feedback from an array of creators to incorporate a broader range of voice samples and additional features to better serve various content creation environments,” said Kyogu Lee, President of Supertone.

“We hope this innovation will elevate the value of voice within the creative process, pushing the boundaries of creativity for creators.”

Another product launched recently by Supertone arrived in November 2023 in the form of Supertone Clear, a real-time AI audio plugin for voice separation and noise reduction.

Supertone says that the tool has been used widely by creators including post-production engineers, video editors, streamers, musicians, and record producers in 198 countries.

The company reports that it garnered over 27,000 monthly active users within just six months of its launch in November, 2023.

Supertone first generated global media attention in January 2021 with its so-called Singing Voice Synthesis (SVS) technology.

The company used this tech to “resurrect” the voice of South Korean folk superstar Kim Kwang-seok, with the subsequent AI-generated voice debuted on Korean television show Competition of the Century: AI vs Human.Music Business Worldwide

Related Posts