Component
École Nationale Supérieure d'Électrotechnique d'Électronique d'Informatique d'Hydraulique et des Télécommunications
Objectives
Discover tools for processing and generating audio signals, particularly for restoration (noise reduction, click removal, remixing) and music synthesis (physical models, generative AI).
Description
The course is structured as follows:
Lecture 1: Audio signal denoising (click removal, background noise reduction)
Lecture 2: Non-negative matrix factorisation (NMF) and its application to audio source separation
CM3: Physical models for music synthesis
CM4: Overview of AI methods for audio signal processing
Practical work: Interpolation of audio signals for click removal
Project: Writing a presentation document on Shazam, Audionamix or Spotify.
Pre-requisites
Statistical signal processing, time-frequency analysis, machine learning.