Home
🐸 TTS is a deep learning based text-to-speech solution. It favors simplicity over complex and large models and yet, it aims to achieve the state-of-the-art results.
Based on the user study, 🐸 TTS is able to achieve on par or better results compared to other commercial and open-source text-to-speech solutions. It also supports various languages and already applied to more than 13 different languages.
The general architecture we use comprises two separate deep neural networks. The first network computes acoustic features from given text input. The second network produces the voice from the computed acoustic features. We call the first model "text2feat" and the second "vocoder".
🐸 TTS also serves a Speaker Encoder model that can be used for computing speaker embedding vectors for various purposes including speaker verification, speaker identification, multi-speaker text-to-speech models.
Currently, we implemented the following methods and models.