DiffRoll: Diffusion-based Generative Music Transcription with Unsupervised Pretraining Capability

TL;DR

DiffRoll introduces a novel diffusion-based generative model for automatic music transcription that outperforms traditional discriminative methods and can utilize unpaired datasets for training.

Contribution

The paper presents DiffRoll, a diffusion-based generative approach for AMT that enables transcription, inpainting, and training on unpaired data, outperforming existing methods.

Findings

DiffRoll outperforms discriminative models by 19 percentage points.

DiffRoll surpasses similar existing methods by 4.8 percentage points.

The approach enables training on unpaired datasets.

Abstract

In this paper we propose a novel generative approach, DiffRoll, to tackle automatic music transcription (AMT). Instead of treating AMT as a discriminative task in which the model is trained to convert spectrograms into piano rolls, we think of it as a conditional generative task where we train our model to generate realistic looking piano rolls from pure Gaussian noise conditioned on spectrograms. This new AMT formulation enables DiffRoll to transcribe, generate and even inpaint music. Due to the classifier-free nature, DiffRoll is also able to be trained on unpaired datasets where only piano rolls are available. Our experiments show that DiffRoll outperforms its discriminative counterpart by 19 percentage points (ppt.) and our ablation studies also indicate that it outperforms similar existing methods by 4.8 ppt. Source code and demonstration are available https://sony.github.io/DiffRoll/.