Universal audio synthesizer control with normalizing flows

TL;DR

This paper introduces a novel control method for audio synthesizers using normalizing flows, enabling automatic parameter inference, macro-control learning, and preset exploration within a unified model, improving interpretability and usability.

Contribution

It proposes a new formulation of synthesizer control with invertible mappings using VAEs and normalizing flows, including disentangling flows for better organization of latent spaces.

Findings

Superior parameter inference and audio reconstruction compared to baselines

Disentangles major audio variation factors as latent dimensions

Enables semantic control and smooth parameter mapping

Abstract

The ubiquity of sound synthesizers has reshaped music production and even entirely defined new music genres. However, the increasing complexity and number of parameters in modern synthesizers make them harder to master. Hence, the development of methods allowing to easily create and explore with synthesizers is a crucial need. Here, we introduce a novel formulation of audio synthesizer control. We formalize it as finding an organized latent audio space that represents the capabilities of a synthesizer, while constructing an invertible mapping to the space of its parameters. By using this formulation, we show that we can address simultaneously automatic parameter inference, macro-control learning and audio-based preset exploration within a single model. To solve this new formulation, we rely on Variational Auto-Encoders (VAE) and Normalizing Flows (NF) to organize and map the respective auditory and parameter spaces. We introduce the disentangling flows, which allow to perform the invertible mapping between separate latent spaces, while steering the organization of some latent dimensions to match target variation factors by splitting the objective as partial density evaluation. We evaluate our proposal against a large set of baseline models and show its superiority in both parameter inference and audio reconstruction. We also show that the model disentangles the major factors of audio variations as latent dimensions, that can be directly used as macro-parameters. We also show that our model is able to learn semantic controls of a synthesizer by smoothly mapping to its parameters. Finally, we discuss the use of our model in creative applications and its real-time implementation in Ableton Live