Dr Johan Pauwels

Research

machine learning for audio, audio signal processing, music information retrieval, binaural audio

Interests

Johan’s aim is to make machines understand audio to the level of a trained professional, such that new tools can be developed to assist musicians performing live or in the studio, listeners navigating large music collections and learners studying music. To that end, he uses a combination of machine learning, signal processing, data science and music theory. In recent years, he has been working on narrowing the gap between academic research and user-centric applications, web-based music services and the personalisation of spatial and immersive audio.

Publications of specific relevance to the Centre for Fundamentals of AI and Computational Theory

Grants

Research Group

PhD Students

• Kasia Adamska
  Predicting Hit Songs: Multimodal and Data-Driven Approach
• James Bolt
  Intelligent Audio and Music Editing With Deep Learning
• Carlos De La Vega Martin
  Neural Drum Synthesis
• Harnick Khera
  Informed Source Separation For Multi-Mic Production
• Gregor Meehan
  Representation Learning For Musical Audio Using Graph Neural Network-Based Recommender Engines.
• Jeffrey Miller
  A Geometric Tonal Model For Harmonic Description and Sequence Analysis of Music
• Christos Plachouras
  Deep Learning For Low-Resource Music
• Shi
  C4DM
• Yannis Vasilakis
  Enhancing Compositionality Properties in Joint Music-Language Models

News