Descript, the audio tech company used by podcasters to create and edit their shows, has just completed its first round of funding and its resulted in the raising of $15 million from the venture capital firms Andreessen Horowitz and Repoint. CEO Andrew Mason says they’ll use the funds to grow the company’s team.
Descript has also just made its first acquisition, striking a deal to buy Lyrebird, a Canadian company that uses Artificial Intelligence technology to allow podcasters to generate a realistic sounding text-to-speech audio clone of themselves based on a brief audio clip that’s used as a template. “As soon as we saw it, we became conscious of a Lyrebird-shaped hole in Descript. In some ways, it felt like our missing half,” Mason said in a blog post. “Descriptlets you delete audio by deleting text, but Lyrebird would let you add audio by typing text.” He said in a blog post that he began talking with two-year old Lyrebird a few months ago and the more they spoke, the more the two companies realized they were a good fit. Their aim will be to create a new class of AI-enabled tools for media editing, he said. Lyrebird will now operate as a division out of Descript’s Montreal offices.
Mason, the founder and former CEO of Groupon, said they’re now putting that new technology to use with newly-used podcast production studio software. It includes a beta version of what they’re calling Overdub, which allows podcast hosts and producers to edit in fixes to a podcast using their own voice, simply by converting the text into audio. “Now you aren’t limited to the words you captured in the voice booth — making corrections is as simple as typing them,” Mason explained.
Hoping to put to rest any deep fake fears, Mason said their technology only allows Overdub to be used in a podcaster’s own voice. They do that by requiring a host to speak randomly generated sentences which will prevent others from using pre-existing recordings to create a model of the host’s voice. “We built this feature to save you the tedium of re-recording/splicing time every time you make an editorial change, not as a way make deep fakes,” said Mason. The company has also posted a code of ethics for the use of the technology on its website. He said Descript worked with the teams that work on the NPR podcasts Rough Translation and Planet Money when developing the Overdub feature.
Descript is free for up to three hours of voice content, and then the company charges ten dollars per month.