Compose & Embellish: Well-Structured Piano Performance Generation via A Two-Stage Approach

TL;DR

This paper introduces a two-stage Transformer framework for generating expressive piano performances by first composing a lead sheet and then embellishing it, significantly improving musical structure and expressiveness.

Contribution

The novel two-stage approach enables better long-range structure in piano generation and allows pretraining on non-piano data, enhancing performance quality.

Findings

Reduces the gap in musical structure compared to real performances by half.

Improves richness and coherence of generated music.

Effective two-stage framework for expressive piano performance generation.

Abstract

Even with strong sequence models like Transformers, generating expressive piano performances with long-range musical structures remains challenging. Meanwhile, methods to compose well-structured melodies or lead sheets (melody + chords), i.e., simpler forms of music, gained more success. Observing the above, we devise a two-stage Transformer-based framework that Composes a lead sheet first, and then Embellishes it with accompaniment and expressive touches. Such a factorization also enables pretraining on non-piano data. Our objective and subjective experiments show that Compose & Embellish shrinks the gap in structureness between a current state of the art and real performances by half, and improves other musical aspects such as richness and coherence as well.