KT releases multilingual AI voice synthesis service capable of emotional expression

By Kim Joo-heon Posted : July 19, 2022, 14:42 Updated : June 23, 2023, 00:58

[Courtesy of KT]

SEOUL -- KT, a top telecom company in South Korea, released a multilingual artificial intelligence-based voice studio where users can create video and audio content capable of delivering human emotions using 100 synthesized voices. The service is available in Korean, English, Chinese and Japanese.
 
Using an AI voice function, users can express five different emotions -- joy, calm, neutrality, sadness and anger. Emotional dubbing technology was applied to enable dubbing emotions into artificial voices as they are read by users.
 
KT has created an audiobook through collaboration with Millie's Library, a top domestic online book subscription service, using the voice of Yoon Do-hyun, a 50-year-old singer known for his rock band YB. A book titled "I quit my job due to difficulty in relationships" was recreated into an audiobook using Yoon's artificial voice.
 
"We will work with many different partners including our affiliated companies to release more innovative AI voice studio services," KT's AI and big data business department head Choi Joon-ki said in a statement on July 19. The company said various giveaway events will be provided until August 7.
 
The audiobook market rapidly grew after South Korea was hit by the COVID-19 pandemic in 2020. In September 2021, Genie Music, the online and smartphone music subscription service wing of KT, acquired a controlling stake in Millie's Library targeting the audiobook market estimated at about 30 billion won ($22.8 million).

Millie's Library provides unlimited access to different types of e-books including texts and audio to users with a monthly subscription of 9,900 won. As of December 2021, Millie has secured some 4 million users.
기사 이미지 확대 보기
닫기