Artificial intelligence (AI) machine translation and voice recognition specialist Elsolu has announced the launch of the beta service for ‘deepSAME’, a next-generation AI-based video translation service, starting from the 5th of this month.
This new service clones the speaker’s voice within the video to naturally dub it into multiple languages and precisely lip-syncs it according to the pronunciation of the target language, marking a complete departure from the traditional subtitle-based translation method and signaling a paradigm shift toward immersive video translation.
‘DeepSAME’ has been developed based on over 20 years of accumulated language data by Elsolu and their proprietary AI machine translation and voice processing technology. While maintaining the automatic translation capability and accuracy provided by the previous subtitle-based solution ‘SubEditor’, it evolves towards dubbing to alleviate the fatigue and decreased engagement associated with watching subtitles. Notably, it integrates world-class voice cloning and lip-sync technology to replicate the speaker’s emotion, tone, and atmosphere as if performed by a native speaker.
The use of ‘deepSAME’ is simple. Users can upload a video or simply input a YouTube video link, and the system automatically separates the background music from the speaker’s voice. It then performs AI multilingual dubbing while preserving the speaker’s tone and timbre. Furthermore, the lip-sync technology is applied to automatically adjust the speaker’s lip movements to match the pronunciation of the target language, allowing multilingual viewing while preserving the original content’s immersive experience and emotional conveyance.
Significantly, ‘deepSAME’ provides overwhelming time and cost efficiency compared to traditional labor-intensive dubbing and lip-sync methods. Where previous methods typically cost millions of won and required more than a day for production, ‘deepSAME’ can automatically process dubbing and lip-sync for entire videos in just a few minutes. It currently supports 33 languages, including Korean, English, Japanese, Chinese, and Vietnamese, and is designed to suit the global content production environment.
Additionally, it includes a post-editing editor that allows users to manually adjust subtitles, voice, timing, etc., after automatic translation and dubbing, facilitating the creation of high-quality outputs. In tests with major overseas content producers, it demonstrated a reduction in production time by more than 90% compared to existing methods.
CEO Jongwook Mun stated, “DeepSAME is not just a simple video translation tool; it is a next-generation platform that enables AI to replicate the speaker’s voice and lip movements to create immersive multilingual content.” He added, “We will set a new standard in the global video production market with a service that satisfies production efficiency, content quality, and user immersion.”
The ‘deepSAME’ beta service is available from August 5 to November 30, allowing anyone to freely experience AI dubbing and lip-sync features for up to ten videos with a length of 5 minutes in 33 languages upon registration.
Meanwhile, Elsolu, formerly known as ‘Systran International’, is a leading AI technology company in South Korea, providing solutions in the fields of machine translation, voice recognition, and voice synthesis to various sectors including domestic and international public, defense, finance, manufacturing, and distribution industries. Their main products include ezNTS (machine translator), ezDAS (voice recognizer), ezTTS (voice synthesizer), AI Transcribe (E2E voice recognizer), and AI SubEditor (video subtitle translator).