Machine Listening

In the Machine Listening area we concentrate on the automatic analysis and understanding of musical and and other sounds from the world around us.

We use a wide variety of techniques to analyse sounds, including: short-time Fourier transforms (STFTs), wavelets, cosine packets, Mel-frequency cepstral coefficients (MFCCs), hidden Markov models, Bayesian models, sinusoidal analysis, independent component analysis, dynamic Bayesian networks, and sparse representations.

This area is led by Prof Mark Plumbley. Projects in this area include:

• Information Dynamics of Music
• Beat Tracking and Rhythmic Analysis
• Sparse Representations for Audio Source Separation
• Compressed Sensing of Audio Scenes
• Automated Composition
• Machine Listening using Sparse Representations
• Interactive Real-time Musical Systems
• Musical Audio Analysis for Real-Time Interaction
• Real-Time Analysis of Voice for Musical Applications
• Musical Audio Stream Separation
• Sparse Object-Based Coding of Music

PhD Study - interested in joining the team? We are currently accepting PhD applications.