A Universal Music Translation Network

TL;DR

This paper introduces a universal music translation network that can convert music across instruments, genres, and styles using an unsupervised, multi-domain autoencoder capable of translating even unseen musical domains.

Contribution

It presents a novel multi-domain wavenet autoencoder with a shared encoder and disentangled latent space, enabling unsupervised translation across diverse musical domains, including unseen ones.

Findings

Achieves convincing music translation on NSynth and professional musician datasets.

Enables translation from whistling, facilitating instrumental music creation by untrained users.

Operates without supervision or matched sample pairs.

Abstract

We present a method for translating music across musical instruments, genres, and styles. This method is based on a multi-domain wavenet autoencoder, with a shared encoder and a disentangled latent space that is trained end-to-end on waveforms. Employing a diverse training dataset and large net capacity, the domain-independent encoder allows us to translate even from musical domains that were not seen during training. The method is unsupervised and does not rely on supervision in the form of matched samples between domains or musical transcriptions. We evaluate our method on NSynth, as well as on a dataset collected from professional musicians, and achieve convincing translations, even when translating from whistling, potentially enabling the creation of instrumental music by untrained humans.