This repository implements a neural vocoder that reconstructs waveform audio from frozen WavLM-Base+ representations, specifically trained and evaluated on French speech corpora.
Neural vocoding for VoiRS speech synthesis - converts mel spectrograms to high-quality audio. This crate implements state-of-the-art neural vocoders including HiFi-GAN and DiffWave for converting mel ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results