A Study on Synthesizing Expressive Violin Performances: Approaches and Comparisons

TL;DR

This paper compares two approaches for expressive violin performance synthesis, analyzing their effectiveness, diversity, and controllability through experiments, to address challenges in modeling expressive musical interpretation.

Contribution

It introduces and evaluates two distinct EMS approaches for violin, highlighting their strengths and limitations in a comparative study.

Findings

End-to-end approach achieves high naturalness.

Parameter-controlled approach offers better controllability.

Trade-offs exist between diversity and model effectiveness.

Abstract

Expressive music synthesis (EMS) for violin performance is a challenging task due to the disagreement among music performers in the interpretation of expressive musical terms (EMTs), scarcity of labeled recordings, and limited generalization ability of the synthesis model. These challenges create trade-offs between model effectiveness, diversity of generated results, and controllability of the synthesis system, making it essential to conduct a comparative study on EMS model design. This paper explores two violin EMS approaches. The end-to-end approach is a modification of a state-of-the-art text-to-speech generator. The parameter-controlled approach is based on a simple parameter sampling process that can render note lengths and other parameters compatible with MIDI-DDSP. We study these two approaches (in total, three model variants) through objective and subjective experiments and discuss several key issues of EMS based on the results.