On the Potential of Simple Framewise Approaches to Piano Transcription

TL;DR

This paper demonstrates that simple framewise neural network approaches can outperform complex methods in piano transcription, establishing a new baseline on the MAPS dataset without complex post-processing.

Contribution

It shows that straightforward neural network models with proper training and regularization can surpass state-of-the-art results in piano transcription.

Findings

Simple framewise approach outperforms previous methods

Effective input representations identified for neural networks

Achieved state-of-the-art results on MAPS dataset

Abstract

In an attempt at exploring the limitations of simple approaches to the task of piano transcription (as usually defined in MIR), we conduct an in-depth analysis of neural network-based framewise transcription. We systematically compare different popular input representations for transcription systems to determine the ones most suitable for use with neural networks. Exploiting recent advances in training techniques and new regularizers, and taking into account hyper-parameter tuning, we show that it is possible, by simple bottom-up frame-wise processing, to obtain a piano transcriber that outperforms the current published state of the art on the publicly available MAPS dataset -- without any complex post-processing steps. Thus, we propose this simple approach as a new baseline for this dataset, for future transcription research to build on and improve.