Narrator — one video, more than a hundred languages

Our Lumenta Labs promo videos needed narration in several languages. We could have paid a studio for each language or subscribed to a TTS service. Instead we built our own tool — in five days.

• Narration on Azure Neural TTS, placed around the parts where the original video already speaks — the tool finds them itself
• Music is automatically ducked under the voice
• Subtitles are exported in a format YouTube and Facebook accept
• Translated versions in one click, with brand names left untouched
• More than a hundred languages available — a new language needs neither a new tool nor a new voice actor

• The same promo video got narration in four languages: English, German, Croatian and Dutch
• Built in five days and in daily use since
• It narrates our own promo videos and the films and videos we make for clients

The expensive part of localisation is not the voice but everything around it: timing, music, subtitles, brand names. Automate that and a new language stops being a project and becomes a click. It is what our video localisation service runs on — you get a version for every market, not a new project for every language.