HITS

It is getting tougher and tougher to figure out what is real and what isn’t as voice technology continues to advance at a rapid pace, attendees at the Innovation and Transformation Summit (ITS): Localisation event in London were told Feb. 28, during the session “KEYNOTE CASE STUDY: Advancements in Voice Technology.”

The use cases for this type of technology are huge and companies behind it are creating content that is astounding audiences around the world.

During the ITS: Localisation keynote session, a representative from a pioneering start-up discussed how the technology is developing, how it is being used and what the future holds, as the machines continue to rise.

Joining the live event from her computer online was keynote speaker Anna Bulakh, head of ethics and partnerships at Ukraine-based synthetic speech startup Respeecher, which uses proprietary deep learning/artificial intelligence techniques to produce high quality synthetic speech for companies and services that have included Disney Plus so far.

The company already has “quite a big team of brilliant minds working in machine learning, in artificial intelligence, and we can recreate anyone’s voice that we have permission for”, she told attendees

She demonstrated a few interesting use cases for her company’s technology that are applicable in localisation and also discussed the ethics involved in the use of this tech – comparable to the ethics involved with deep fake images – and how her company approaches it.

First, she pointed out that her company “received permissions to recreate” the voices for Star Wars characters including Darth Vader for the Disney Plus TV show The Mandalorian. The audience was “so happy to see those characters be brought back to life,” she said. But her company’s work is not limited to big projects like that, she told attendees.

The company also recreated the voice of Richard Nixon for a documentary back in 2018 and “we have four Emmy statues in our office in Kyiv,” she added.

What the company has achieved with the hyper realistic recreation of voices can also be used in not just the entertainment industry, she said, noting “we are also applying it right now in [the]  healthcare industry and in cybersecurity.”

The company has also been able to recreate voices in multiple languages that can be used for localisation purposes across the world, she said, demonstrating singer Aloe Blacc, who can only speak English, singing in Chinese and Spanish using Respeecher’s technology.

The Innovation and Transformation Summit: Localisation was sponsored by AppTek, Signiant, EIDR, Iyuno, LinQ Media Group, Vubiquity, OOONA, XL8, and Collot Baca, and was produced by MESA, in association with the Content Localisation Council.