June product updates and announcements
Voice recording, voice samples, improved model submissions and more!
Summer is in full swing here at Uberduck, and you know what that means: writing code, shipping features, and even touching grass on special occasions.
We've been hard at work on a lot of cool stuff recently. Here's the run-down.
New features
- We launched a new in-browser recorder for custom voice clones. Now when you purchase a custom voice clone, you can record and submit audio without leaving your web browser.
- We expanded contributor model submissions to Hifi GAN models. Hifi GAN is a state-of-the-art neural vocoder created by Jungil Kong, Jaehyeon Kim, and Jaekyoung Bae. To learn about the role of a vocoder in a neural text to speech system, join our Discord and ask in the #machine-learning channel.
- We added voice samples for every single voice on Uberduck. Now you can sample the Uberduck voices without even logging in! You can even access the voice samples via our API.
- We published a guide to finding the best and most popular voices on Uberduck. Spoiler alert: use the voice leaderboard
Top new voices
A few favorite voices, as recommended by the Uberduck community:
- Bog, created by scruffygamer
- SpongeBob SquarePants (regular), created by Gosmokeless28
- Max, created by finger-player
- pishifat, created by hecko
- Morshu (Cd-i), created by Dawd-de-Era and Gosmokeless28
June 30, 2022