Press Press2022.07.05

CJ OliveNetworks applies “AI Voice Cloning” to tvN’s descriptive video service

공유하기 Facebook Twitter Kakao Copy Url

- The Korean TV channel, tvN, has become the first of its kind to introduce AI Voice Cloning to its programs to automatically convert text into voice. 

- An AI algorithm allows any text to be converted into voice with thousands of voice variations.

- Actress Jeon Mi Do narrates a descriptive video of tvN’s weekend drama, “Alchemy of Souls,” to offer a special experience for visually impaired consumers.


CJ OliveNetworks, led by CEO Cha In Hyok, has introduced AI technology to tvN’s descriptive video service to offer a special experience to consumers. 


The descriptive video service refers to a form of audio-visual translation of television content for visually impaired consumers, where the overall description of the scene, such as characters’ facial expressions, gestures, or change of scene, is delivered through the voice of a narrator. 


Most descriptive video services record a separate audio track that describes the scene, mix it with the primary audio track to create a secondary track, and broadcast it as a bilingual broadcast. This forces another round of sound mixing after final editing, resulting in a longer lead time for production. 


To tackle the issue, CJ OliveNetworks has become the first of its kind to use AI Voice Cloning technology as a descriptive video service to help visually impaired consumers, who depend solely on the sound when watching television, better enjoy the media content through the voice of a famous figure that they are familiar with. 


AI Voice Cloning is a voice synthesis technology powered by an AI algorithm to analyze and automatically generate individual voices. Once the algorithm is trained with a voice, it can read texts in the same voice through automatic voice generation. 


This technology removes the need for a separate audio track for each descriptive video content as the AI can generate the narrations of various programs at once based on the voices it learned, reducing the time until the content is delivered to visually impaired consumers. 


tvN has collaborated with actress Jeon Mi Do to narrate the descriptive video applied with AI Voice Cloning, offering greater joy to the audience. Jeon recorded various sample sentences for AI training to allow the AI to narrate tvN’s weekend drama, “Alchemy of Souls”, in her voice. 


“It was a very rewarding experience for us to offer a special descriptive video experience to visually impaired consumers through AI Voice Cloning,” said Lee Hyun Ki, the team lead for the Virtual Human Lab at CJ OliveNetworks. He said he will “accelerate the development of media AI technologies, such as AI Voice Cloning or Virtual Human, to enable all to enjoy a differentiated experience and deeper immersion into media content.”