Harmonic-Percussive Source Separation Versus Multi-Condition Training Set for Robust Speaker Verification in Adverse Acoustic Environments: Viability Study and Best Practices
Information · 09.11.2026 · 17 (11) · 1121
DOI: 10.3390/info17111121 ↗Kurz gefasst
Fachartikel in der Zeitschrift Information, online erschienen am 09.11.2026. Verfasst von Daniele Casali und 3 weiteren Autor:innen. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
Speaker Recognition (SR) systems suffer from severe performance degradation when deployed in real-world environments with ambient noise and reverberation. This study investigates a joint harmonic-percussive feature representation, combining RASTA-MFCC descriptors extracted independently from both harmonic and percussive streams that help to manage noise in classical machine learning classifiers (SVM, Random Forest, Naïve Bayes), and compares them with the deep embedding extractor ECAPA-TDNN. Evaluated on the VOiCES dataset across diverse room geometries and background distractors, our results demonstrate that while deep neural models handle adverse acoustics natively, HPPS provides a consistent performance boost for classical ML pipelines with a decrement of the error rate up to 5% on weighted mean error and 8% on specific noises as television, highlighting a viable mechanism for enhancing robustness at the expense of additional computational overhead.
Zitieren
APA
Casali, D., Paolizzo, F., Cesarini, V., & Costantini, G. (2026). Harmonic-Percussive Source Separation Versus Multi-Condition Training Set for Robust Speaker Verification in Adverse Acoustic Environments: Viability Study and Best Practices. Information, 17(11), 1121. https://doi.org/10.3390/info17111121
Vancouver
Casali D, Paolizzo F, Cesarini V, Costantini G. Harmonic-Percussive Source Separation Versus Multi-Condition Training Set for Robust Speaker Verification in Adverse Acoustic Environments: Viability Study and Best Practices. Information. 2026;17(11):1121. doi:10.3390/info17111121
BibTeX
@article{Casali2026,
author = {Casali, Daniele and Paolizzo, Fabio and Cesarini, Valerio and Costantini, Giovanni},
title = {Harmonic-Percussive Source Separation Versus Multi-Condition Training Set for Robust Speaker Verification in Adverse Acoustic Environments: Viability Study and Best Practices},
journal = {Information},
year = {2026},
volume = {17},
number = {11},
pages = {1121},
doi = {10.3390/info17111121}
}
EUVIMED