EUVIMEDEUVIMED entdecken
Crossref

Instrument-wise music similarity representation learning with source separation and human preference

Takehiro Imamura, Yuka Hashizume, Wen-Chin Huang, Tomoki Toda

The Journal of the Acoustical Society of America · 01.10.2025 · 158 (4_Supplement) · A434-A434

DOI: 10.1121/10.0041415 ↗

Kurz gefasst

Fachartikel in der Zeitschrift The Journal of the Acoustical Society of America, erschienen am 01.10.2025. Verfasst von Takehiro Imamura und 3 weiteren Autor:innen.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

This study presents an instrument-wise music-similarity representation learning (InMSRL) framework that leverages music source separation and human preference data without requiring clean stems at inference. Conventional methods follow either the Cascade architecture, which separates sources prior to per-instrument similarity extraction and thus propagates separation errors, or a Direct model that attempts single-stage disentanglement but often fails to isolate target instrument features. We enhance both paradigms. First, we introduce end-to-end fine-tuning (E2E-FT) for the Cascade architecture, jointly optimizing the separator and similarity extractors with an auxiliary separation loss to mitigate the error propagation. Second, we propose a multi-task Direct model that reconstructs individual stems from its own embeddings, thereby sharpening instrument-level disentanglement. Finally, we incorporate perception-aware fine-tuning (PAFT) using human preference data to improve perceptual similarity representation. Experimental evaluations show that (1) Cascade with E2E-FT significantly boosts objective InMSRL metrics, (2) the multi-task Direct variant further enhances disentanglement performance, (3) PAFT markedly elevates perceptual InMSRL performance, and (4) the combined Cascade with E2E-FT and PAFT consistently outperforms all other models. [Work supported by JST AIP Acceleration Research JPMJCR25U5, Japan.]

Zitieren

APA

Imamura, T., Hashizume, Y., Huang, W., & Toda, T. (2025). Instrument-wise music similarity representation learning with source separation and human preference. The Journal of the Acoustical Society of America, 158(4_Supplement), A434-A434. https://doi.org/10.1121/10.0041415

Vancouver

Imamura T, Hashizume Y, Huang W, Toda T. Instrument-wise music similarity representation learning with source separation and human preference. The Journal of the Acoustical Society of America. 2025;158(4_Supplement):A434-A434. doi:10.1121/10.0041415

BibTeX
@article{Imamura2025,
  author = {Imamura, Takehiro and Hashizume, Yuka and Huang, Wen-Chin and Toda, Tomoki},
  title = {Instrument-wise music similarity representation learning with source separation and human preference},
  journal = {The Journal of the Acoustical Society of America},
  year = {2025},
  volume = {158},
  number = {4_Supplement},
  pages = {A434--A434},
  doi = {10.1121/10.0041415}
}
Kostenlos starten →