SKT가 광복 79주년을 맞아 AI 기술로 1945년 광복 전후의 영상 및 음원을 선명하게 복원하는 디지털 프로젝트 ‘815 리마스터링’을 공개했다.
Restore old video and audio sources with our own AI technology
Application of ML-based sound source separation to national anthem voice recording
To mark the 79th anniversary of liberation, SKT has unveiled a digital project called '815 Remastering' that uses AI technology to clearly restore video and audio from before and after liberation in 1945.
On the 14th, SKT uploaded a newly produced '815 Remastering' video combining restored video and audio sources to its official YouTube channel 'AI help you?'. The content will also be exhibited at the Independence Hall of Korea in the future.
SKT said, “We planned the project with the aim of vividly restoring the scenery immediately after liberation and helping many people indirectly experience the moving moments of that day.”
SKT has improved existing low-quality videos into high-quality ones using its AI media improvement and restoration solution, ‘SUPERNOVA.’ The sound source was also restored to a high-quality voice by removing background noise and other noises using ML-based sound source separation technology.
An SKT official said, “We have improved the quality of the original content by applying AI and deep learning-based image quality restoration and super-resolution technology.”

The original content used in the project is footage of the streets of Seoul immediately after liberation in 1945, and the sound source of the national anthem recorded in 1942. The footage is a 28-second long document of a parade on the streets of Seoul on August 16, 1945, immediately after liberation. It is a scene where many people who did not know about liberation on August 15 poured into the streets the next day, on the 16th.
The sound source is believed to be the oldest existing patriotic vocal recording. It is known that the sound source was included in an album produced and distributed by the Korean National Association on August 29, 1942, a day of national shame, to show their determination for independence. It is currently housed in the Independence Hall of Korea.
An official from the Independence Hall of Korea said, “It is meaningful that the audio of the national anthem that our ancestors sang to strengthen their resolve for independence and the video footage from the time of liberation have been restored as improved content using today’s AI technology,” adding, “Furthermore, with the 80th anniversary of liberation approaching next year, it is very meaningful that we can vividly remember the excitement of liberation through the video footage of our ancestors’ spirit of independence and the joy of liberation contained in the materials.”
An SKT official explained, “We hope that this project will serve as an opportunity to look back on the sacrifices of our independence fighters and share the joy of liberation.” SKT signed an agreement with the Independence Hall of Korea in 2020 to build an MR/AR eco-museum, and has been jointly planning various AI and digital content for the past five years.
Park Gyu-hyun, SKT’s Vice President and Head of Digital Communications, said, “This year, as we celebrate our 40th anniversary, we are striving to create various digital contents using AI as a ‘global AI company’ and build empathy with our customers.” He added, “In the future, we will continue to use SKT’s various technologies to effectively convey social and national meanings in a way that suits the characteristics of digital and communicate externally.”