Rome, October 12, 2023 — Translated, a leading provider of AI-powered language solutions, has unveiled a first-of-its-kind podcast and video localization service that enables companies and users to translate and recreate their own voice in target languages. By combining transcription, adaptive machine translation (MT), linguists, and AI voice cloning technologies, Translated enables anyone to leverage the hallmark of their voice in the most widely spoken languages.
Until now, companies and creators have faced significant cost and time barriers breaking into the global scene. While generative AI voice technologies can help quickly render voices in another language, they don’t ensure that the generated audio conveys the original message properly across cultures. Purely automated systems lack the sophistication to deliver the authenticity and emotional impact that brands need to communicate their vision and intent. In addition, the latest examples of generative AI voice technologies are not ready to solve this problem, as they are typically limited to one person speaking and no audio content other than the voice.
Translated’s new AI-powered voice translation service is powered by Matedub, a first-of-its-kind dubbing tool that automatically transcribes an audio or video source, translates it into the target language, and allows for editing of the translation. Users can tailor the text spoken by the AI voice, helping them resonate more with audiences in the target language. Translated’s linguists, carefully selected and specifically trained through a network of 300,000 native-speaking professional translators, edit the translations to help the AI voice mimic the feeling and intonation of the original voice.
Translated’s Audiovisual Director Fabio Minazzi commented on the news: “Similar to the impact of movable type on book printing, recent innovations in voice cloning and multilingual AI speech are breaking down communication barriers in unprecedented ways. At Translated, we’ve been investing in AI technologies and a new breed of professional linguists called “transvoicers,” whose translation and language skills are used to create text and synthetic voice performances in a reliable and engaging way. Today, we are launching a new service to help companies and individuals maintain their distinctive voice when speaking to a global audience.”
In Matedub, translators work with a custom adaptive MT model based on ModernMT, Translated’s leading adaptive MT system (IDC, 2022). The MT system learns from previous translations and improves in real-time through corrections, adapting to context. The reliability of the AI voice cloning process is reflected in Translated’s experience developing AI voices for clients such as Airbnb, providing subtitles to creators within the YouTube platform, and partnering on projects such as Voice For Purpose, which has been endorsed by the World Health Organization.
Translated’s high-quality AI dubbing comes at a fraction of the cost of traditional dubbing, between 10 to 30 percent of the typical cost, depending on the use case. It is already available in the 20 most spoken languages in the world, with more to be added soon.
The new podcast and video localization service is provided as a stand-alone service, as well as through connectors to content management systems (CMS) and Version Control Systems (VCS) for automated input/output of content. Companies can contact Translated’s sales department or their own account manager directly to start using the new service. Creators can upload the audio or video they wish to have localized at video.translated.com to get a quote or contact Translated at translated.com/contact-us for a larger partnership.