SafeSpeech

This is the source code of "SafeSpeech: Robust and Universal Voice Protection Against Malicious Speech Synthesis".

In this repository, we provide three files as we claimed in "Open Science":

source code.zip: The source code of SafeSpeech;

datasets.zip: The LibriTTS and CMU ARCTIC datasets utilized in our paper;

download_models.py: The code for downloading pre-trained checkpoints of the surrogate model and evaluation metrics.