The use of deepfakes has generated a lot of noise, particularly when comparing the dangers with the benefits. To help anyone interested in better understanding the nuances of this technology, Linkedin Learning has produced a course called ‘Understanding the Impact of Deepfake Videos.’

Respeecher, as a producer of synthetic media, was invited to provide insights into the use cases of deepfake audio.

Ashley Kennedy, a Managing Staff Instructor at LinkedIn, led the interview and the team of Staff Instructors in the Business and Creative Libraries at LinkedIn Learning. Respeecher’s Chief Research Office, Grant Reaber, speaks about deepfake audio, its process of production, and its current and future use cases in the section ‘What is Deepfake Audio.’

Voice synthesis benefits and use cases

During the interview, Grant Reaber discusses the benefits of speech-to-speech voice conversion technology. Voice cloning allows for changing a person’s voice without sacrificing speech patterns. In this way, deepfake audios can help achieve greater control over the expressed emotions. STS makes it possible to deliver a natural performance, retaining the inflection and the other characteristics of human speech.

“Expressivity and this subtle control over the performance is one reason to use voice conversion. So, actors can express themselves just as much through voice conversion as they could if they were using their [actual] voice. It’s just sort of like an additional creative tool where you can change the voice of the performer.“

Grant Reaber, Chief Research Office, Respeecher

Respeecher works with different motion picture studios and video game producers. One of the company’s latest works includes synthesizing a young Luke Skywalker’s voice for Disney+’s The Mandalorian.

Respeecher creates a synthetic voice that’s indistinguishable from original speech. Some of the most common voice cloning use cases are:

• Making changes to dialogs in movies
• Updating voiceovers for different content
• Improving voices for people with voice disabilities
• Dubbing and localizing video content
• Resurrecting voices from the past

All these use cases become possible only after obtaining permission from the private individual.

In conclusion

In the course, Ashley Kennedy dives into the nuances of cloning human speech and

delves into the impacts of this voice synthesis technology. Unethical deepfakes can lead to several unpleasant outcomes, including the character assassination of public figures.

That’s why part of Respeecher’s mission is to educate individuals and businesses on how to leverage synthetic speech technology the right way. We develop automatic algorithms to detect synthetic speech even if we have not watermarked it.

Enroll in the course to learn more about deepfakes and find out how to leverage them for your needs.

This article was initially published on USA Tribune Media.