Generative Modelling for Controllable Audio Synthesis of Expressive Piano Performance
This work addresses the problem of generating expressive and controllable piano performances for musicians and creators, offering a tool for inspiring new interpretations of existing music.
The paper tackles controllable audio synthesis of expressive piano performances by developing a GM-VAE model that generates realistic piano audio following temporal conditions for articulation and dynamics, enabling fine-grained style morphing.
We present a controllable neural audio synthesizer based on Gaussian Mixture Variational Autoencoders (GM-VAE), which can generate realistic piano performances in the audio domain that closely follows temporal conditions of two essential style features for piano performances: articulation and dynamics. We demonstrate how the model is able to apply fine-grained style morphing over the course of synthesizing the audio. This is based on conditions which are latent variables that can be sampled from the prior or inferred from other pieces. One of the envisioned use cases is to inspire creative and brand new interpretations for existing pieces of piano music.