Expressive MIDI-format Piano Performance Generation

TL;DR

This paper introduces a neural network model capable of generating expressive MIDI piano performances with micro-timing, dynamics, and pedal effects, aiming to produce more human-like and emotionally expressive music.

Contribution

The work presents a novel symbolic music generation model that enhances expressivity in MIDI piano performance, addressing limitations of previous models in musical expressiveness.

Findings

Generated performances exhibit micro-timing and dynamics comparable to human performances.

The model demonstrates strong potential for expressive music generation despite limited training.

Expressive qualities are reflected in rich polyphony and pedal effects.

Abstract

This work presents a generative neural network that's able to generate expressive piano performance in MIDI format. The musical expressivity is reflected by vivid micro-timing, rich polyphonic texture, varied dynamics, and the sustain pedal effects. This model is innovative from many aspects of data processing to neural network design. We claim that this symbolic music generation model overcame the common critics of symbolic music and is able to generate expressive music flows as good as, if not better than generations with raw audio. One drawback is that, due to the limited time for submission, the model is not fine-tuned and sufficiently trained, thus the generation may sound incoherent and random at certain points. Despite that, this model shows its powerful generative ability to generate expressive piano pieces.