EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

How Far Can a U-Net Go? An Empirical Analysis of Music Source Separation Performance

Daniel Kostrzewa, Mikolaj Kondziolka, Robert Brzeski, Jeremiah Abimbola, Pawel Benecki

Applied Sciences · 25.02.2026 · 16 (5) · 2195

DOI: 10.3390/app16052195 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Applied Sciences, online erschienen am 25.02.2026. Verfasst von Daniel Kostrzewa und 4 weiteren Autor:innen. Die Arbeit wurde bisher 3-mal zitiert (Crossref) und nennt 42 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: statutory research funds of Department of Applied Informatics, Silesian University of Technology, Gliwice, Poland.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Music source separation (MSS) focuses on decomposing a mixed audio signal into individual instrumental components and is increasingly relevant for music production, restoration, remixing, education, and music information retrieval. Deep learning methods, particularly U-Net architectures operating on time–frequency representations, have recently advanced the state of the art beyond traditional signal-processing techniques. This work presents an optimized multi-source U-Net model for separating selected musical instruments from stereo mixtures. The system uses magnitude spectrograms generated by the short-time Fourier transform and is trained and evaluated on the MUSDB18 dataset. We systematically examine architectural and training-related factors, including normalization strategies, dropout placement, optimizer selection, loss weighting, data augmentation, and spectrogram-domain modifications. Separation quality is measured using BSS Eval metrics, assessing artifacts, interference, and distortion. Experimental results show that the proposed configuration achieves competitive performance relative to established convolutional and U-Net-based open-source systems, especially in terms of vocal track separation, offering practical insights into designing efficient models for multi-instrument separation.

Förderung

  • statutory research funds of Department of Applied Informatics, Silesian University of Technology, Gliwice, Poland

Zitieren

APA

Kostrzewa, D., Kondziolka, M., Brzeski, R., Abimbola, J., & Benecki, P. (2026). How Far Can a U-Net Go? An Empirical Analysis of Music Source Separation Performance. Applied Sciences, 16(5), 2195. https://doi.org/10.3390/app16052195

Vancouver

Kostrzewa D, Kondziolka M, Brzeski R, Abimbola J, Benecki P. How Far Can a U-Net Go? An Empirical Analysis of Music Source Separation Performance. Applied Sciences. 2026;16(5):2195. doi:10.3390/app16052195

BibTeX
@article{Kostrzewa2026,
  author = {Kostrzewa, Daniel and Kondziolka, Mikolaj and Brzeski, Robert and Abimbola, Jeremiah and Benecki, Pawel},
  title = {How Far Can a U-Net Go? An Empirical Analysis of Music Source Separation Performance},
  journal = {Applied Sciences},
  year = {2026},
  volume = {16},
  number = {5},
  pages = {2195},
  doi = {10.3390/app16052195}
}
Kostenlos starten →