AI and the future of localization

Posted By Marc Loftus on July 25, 2019 09:24 am | Permalink
Yesterday, I had the chance to sit down with Chris Carey, chief revenue officer/managing director of the Americas for BTI Studios ( BTI specilizes in localization, including dubbing and subtitling, with 25 locations around the world. Just last month, the company opened a new facility in Burbank, CA, after having an office there for several years.

Carey was in NYC for the Media & Entertainment Services Alliance's (MESA) Media & Entertainment Day event, where pros gathered to share the latest news, tech and trends impacting the smart content side of the M&E industry. He told me he was going to be a part of a panel discussion and planned to talk about the impact of AI on his side of the business, as well as make some predictions for the future of localization.

Photo (L-R): BTI Studios' Chris Carey and Post's Marc Loftus

Localization can range from just subtitles, all the way to a lip sync-acurate audio dub in any number of languages. There is a greater cost for the latter, obviously, but Carey says content owners recognize that a well-done dub will receive higher engagement from consumers than just subtitles, and that often justifies the cost. 

For content to be "localized", it goes through a several-step process. For example, in a case where English-langauge content is being prepared for foreign markets ("outgoing" is how Carey describes it), the process begins with the creation of a word-for-word script derived from the dialogue track. This is used for direct translation into another language. A second "creative dub script" is also created. This version takes into consideration how another language may sync up with an actor's on-screen performance. Words may be revised slightly to make for a clearer interpretation or better fit. Next, an actor is brought into one of BTI's global studios for a recording session. The actor will record their lines to picture, working with an engineer and director, in a similar fashion to a traditional ADR session.

Since BTI has so many international locations - France, Germany, Hungary, Indonesia, Asia and Belgium to name a few - they have access to a large talent base that's capable of delivering the regional dialogue needed for each job.

And the process works both ways. Carey cites a scenario where a French media company is looking to deliver its content to an English- or Spanish-speaking audience. The transcription would take place at BTI's Paris studio, and then they would tap other locations, like the new studio in Burbank, for local voice talent to record the translated script. Media owners benefit by opening up their content to entirely new markets.

During our hour-long chat, Carey touched on AI's role in improving localization services. Right now, there are AI engines that can efficiently transcribe 'speech-to-text'. 'Text-to-text' - one language to another - is another step in the process that dubbing facilities hope to take advantage of using AI technology. And in the future, Carey believes AI will be capable of performing 'text-to-synthesis', whereby a new dialogue track could be created automatically in the language desired for the project. At the MESA event, he planned on sharing his insight on AI's future role.