Projects
Groove Transfer System
My research develops a beat-centric, machine learning–assisted system that automatically infers a recording’s tempo, meter, swing, and expressive micro-timing directly from audio. The system generates a “groove map”- a bar-level representation of human timing; which can transfer the rhythmic feel of one performance to another. By combining rule-based signal analysis with machine learning, this work bridges computational rhythm analysis and expressive performance modelling, contributing both to artistic workflows and Music Information Retrieval (MIR).
Hybrid Granulator
Created a granular synthesis plugin with delay-line feedback, probabilistic grain control, and stereo modulation, adding depth and movement to the sound. Built real-time audio buffering and MIDI-based pitch modulation for a more expressive and hands-on performance experience.
Dynamic Sampling
Used FluCoMa tools to analyse the pitch, loudness, and spectral centroid of monophonic samples, combining these audio descriptors with procedural composition frameworks to create dynamic, generative electronic performances.
