Descript 3.1 is now available. Here’s what’s new:

I’m Andy, Descript’s Head of Content — and a total radio novice. I’m chronicling my journey to get better at podcasting. In this episode, I share how I’ve transformed my on-mic voice from a weird android robot into a living breathing human being, with some sage advice from the professionals for learning how to sound natural on the microphone.Many of the things that helped me along the way are from radio veteran Marianne McCune who you might know from NPR’s Rough Translation. So, thank you, Marianne.

Descript’s Overdub technology allows anyone to create a text to speech version of their own voice by providing about ten minutes of recorded audio. We don’t allow people to clone voices other than their own, both via our terms of use, and the voice identification protocol we have built into our training data collection process.

Last week we made a few big announcements, including the latest Descript release: a full multitrack podcast production studio. (Relaunch the app to update, or download here.)

‍‍Babak Nivi agreed to let us peek behind the scenes at the process for turning one free-flowing conversation into many short episodes of the Naval Podcast. Here’s how they do it:

In this post, we introduce MelGAN, a new generative model of raw audio waveforms created by the Lyrebird team that is capable of generating natural sounding speech at a rate of more than 2,500,000 audio samples per second — more than 100x faster than real time, and 10x faster than alternative methods on similar hardware.We believe that MelGAN paves the way for taking many real-time speech applications onto smaller devices. Imagine, for example, in the not too distant future, having real-time text-to-speech translation on your mobile device without the internet. And, it’s application to music translation brings us one step closer to AI-assisted music composing.We’ve open sourced MelGAN and we encourage interested machine learning developers and researchers to check out our code base.