Transformer-Based Unpaired Piano Accompaniment Style Transfer
Journal
2025 Asia Pacific Signal and Information Processing Association Annual Summit and Conference, APSIPA ASC 2025
Start Page
240
End Page
245
ISBN (of the container)
979-833157206-8
Date Issued
2025-11-28
Author(s)
Ai, Hsin
Abstract
Arranger-specific style transfer for pop piano covers requires effective content-style disentanglement. To address this, we propose a framework that uses a lead sheet (namely, melody and chords) as a style-agnostic content anchor, enabling precise style manipulation without requiring paired data. We then systematically compare several Transformer-based architectures to evaluate the efficacy of a direct token-based conditioning strategy versus more complicated embedding-based methods. While all approaches effectively capture the target style, our evaluation shows that the simpler token-based model achieves superior performance in both objective and subjective assessments of content preservation and style matching. This finding provides empirical evidences that a robust, explicit content representation (i.e., the lead sheet) is highly effective for this task, offering a practical benchmark for controllable music generation.
Event(s)
17th Asia Pacific Signal and Information Processing Association Annual Summit and Conference, APSIPA ASC 2025
Publisher
IEEE
Type
conference paper