Supertone technology gives "Mask Girl" a unique voice

In Netflix's original series "Mask Girl," three actors were cast to play the protagonist, Kim Mo-mi, and the highlight is when she turns herself into an internet live streamer at night while disguising herself behind a mask. When Mo-mi masquerades...

A scene from Netflix original series ″Mask Girl″ [NETFLIX]

In Netflix's original series "Mask Girl," three actors were cast to play the protagonist, Kim Mo-mi, and the highlight is when she turns herself into an internet live streamer at night while disguising herself behind a mask.

When Mo-mi masquerades as a streamer, her voice also changes into a synthesized voice blending those of actors Lee Han-byeol and Nana. Behind the gimmick is AI audio startup Supertone, a company acquired by K-pop powerhouse HYBE.

Supertone utilizes its multi-speaker voice morphing technology to mix and match each voice and tone of two actors to create an entirely new voice.

“Our technology is unique in the sense that it utilizes voice synthesis technology Nansy, short for Neural Analysis and Synthesis,” a Supertone spokesperson said. “Nansy can freely break down, blend or manipulate what constitutes a voice, such as its tone, intonation and pitch to create a very realistic, one-of-a-kind voice. We analyzed each actor’s voice to create a new voice to fit what the director wanted to portray for Mask Girl and we combined the two voices in different proportions for each scene.

“That is why some viewers thought the streamer’s voice was Lee's, and others thought it was Nana's,” the spokesperson added.

Out of the various voice models created by Supertone, director Kim Yong-hun singled out the best fit for the series. A sound studio dedicated to films, Livetone, was in charge of sound design to minutely adjust the details and create the perfect voice.

“Mask Girl,” features the star-studded lineup of Lee, Nana and Ko Hyun-jung, who all portray one character at different stages of her life. Based on the popular Naver Webtoon of the same title which ran on the platform from 2015-2018, global viewers were equally mesmerized by the plot: a woman with an inferiority complex about her looks is a mundane office worker by daylight and a streamer who dances her night away after sundown. Using the online moniker “Mask Girl,” she hides her identity, and more precisely her face, behind the mask.

The series topped the list of Netflix’s weekly non-English language TV shows, recording 7.4 million views from Aug. 21 to 27.

“Supertone’s voice morphing technology aligned with what the production staff wanted, to portray a character with a completely different persona as a BJ streamer,” Supertone’s CEO Lee Kyo-gu said in a statement. “The staff wanted a unique, third voice to portray that side of the protagonist.”

The AI audio company provides its technologies to various content-related projects. Singers utilize AI to release their songs in different languages using the fluent accents and intonations of native speakers or switch the gender of their voice, which is what singer Lee Hyun did in his single “Masquerade” released in March. In the Disney+ original series “Big Bet,” the company transformed 61-year-old actor Choi Min-sik’s voice to that of a man in his 30s to vividly portray a younger version of his character.

BY LEE JAE-LIM [lee.jaelim@joongang.co.kr]