CAS Quarterly

Winter 2023

Issue link: https://digital.copcomm.com/i/1487666

Contents of this Issue

Navigation

Page 64 of 91

C A S Q U A R T E R L Y I W I N T E R 2 0 2 3 63 Val Kilmer suffered from throat cancer and the chemotherapy and tracheostomy he received resulted in the loss of his voice. Using old recordings of his voice, the company Sonantic used AI to recreate his voice for the new Top Gun movie. You can hear how he sounds synthesized in a demo issued by Sonantic, available by searching "Sonantic Val Kilmer" on YouTube. You can also find out more on the company's website at sonantic.io/blog. Sonantic had 10 times fewer data than a normal project. As a result, they developed new algorithms to produce higher quality voice models from the existing data. The more data that these AI programs have—and the fact that they can learn from even more input—means that the reproduction of the human voice can get closer and closer to sounding like the genuine article. In fact, Sonantic released a demo showing a synthesized voice that is flirting and teasing. They do this by incorporating non-speech sounds such as breaths and laughs into their voices. Searching "Sonantic what's her secret?" on YouTube will return a video showing the AI using a flirting voice. An informative article about the company and tech, with embedded audio examples, is available by searching "Sonantic" at TheVerge.com. Sonantic, which was acquired by Spotify this summer, is not the only company with AI-based software creating very lifelike sounding voices. Another company that seems to be doing well is Synthesia (synthesia.io). There are a lot of examples on that page of different voices and video avatars. You can choose an avatar, type in text, and have an avatar speak your text in a really impressive manner. You can even create a free example video from your own input using one of their voices, all of which have been learned from real voice actors. Synthesia boasts 12,000+ companies of all sizes as customers. Yet another company doing this is WellSaid (WellSaidLabs .com). Searching YouTube for "WellSaid AI" will bring up a video that explains how they create their voices. Searching "WellSaid DJ" will bring up another of a DJ who uses it to help with his radio show programming. Along with these three, there are many, many companies using AI to create voices for text-to-speech applications, including Speechify (speechify.com)—which has Gwyneth Paltrow as one of its voices—Murf (murf.ai) and Speechelo (speechelo.com). Sonantic's What's Her Secret video.

Articles in this issue

Links on this page

Archives of this issue

view archives of CAS Quarterly - Winter 2023