Oscillation ::: Materia Forma

28.04.2024 14:00-18:00

Ahnjili ZhuParris: Voice Cloning

15 participants
Registration required
No pri­or know­ledge neces­sary
English spoken

Register here

Designed for both beginners and those with some knowledge of Python, this 4 hour workshop covers the basics of voice cloning, its applications, and ethical considerations. Participants will learn about various voice cloning methods and algorithms, and have the opportunity to create their own personalized cloned voice using the Retrieval-Based Voice Conversion program.

To fully engage in the training of personalized voice cloning algorithms, participants need access to Python and a GPU, such as an Nvidia graphics card or an M1/M2 chip on Mac. Alternatively, the use of the cloud-based computing platform, Paperspace is recommended for its compatibility but requires a subscription fee (approx 8 usd per month). For the voice cloning inference, only CPU is sufficient.

Participants are encouraged to bring at least 10 minutes of their own voice recordings if they wish to clone their voices during the workshop.

Organised in collaboration with Constant vzw.

Ahnjili ZhuParris is a machine learning engineer, Ph.D. candidate, artist, and science communicator currently working at an AI startup. Ahnjili’s artistic research and science communication efforts are dedicated to raising awareness about A.I. and algorithmic violence, which encompasses the violence that may arise from or be justified by automated decision-making systems. Through her work, Ahnjili aims to promote discussions about the ethical implications of these technologies in our society.