Model-Informed Speech Enhancement Using Virtual Room Acoustics and Acoustic Descriptor Optimization
Sensors · 06.06.2026 · 26 (12) · 3630
DOI: 10.3390/s26123630 ↗Kurz gefasst
Fachartikel in der Zeitschrift Sensors, online erschienen am 06.06.2026. Verfasst von Samuel Yaw Mensah und 3 weiteren Autor:innen. Beteiligte Einrichtung: School of Electrical & Information Engineering, Tianjin University, 92 Weijin Road, Nankai District, Tianjin 300072, China. Die Arbeit nennt 45 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
Reverberation and background noise remain persistent obstacles to achieving clear and intelligible speech in enclosed environments. Conventional data-driven or purely empirical dereverberation systems often perform well only under training conditions but lack robustness and physical interpretability when exposed to new acoustic spaces. To address these limitations, this paper proposes a physics-informed speech enhancement algorithm that integrates analytical room acoustics modeling with a descriptor-guided optimization framework. The method employs virtual field simulations based on the Helmholtz equation to estimate key acoustic descriptors, reverberation time (RT60), direct-to-reverberant ratio (DRR), and clarity index (C50), which are then used to adaptively control a model-informed dereverberation filter. This hybrid formulation bridges physical modeling and signal processing, allowing the algorithm to minimize late reverberation energy while maintaining spectral fidelity. Experimental results across multiple simulated and real-room conditions demonstrate measurable improvements over baseline methods, achieving average gains of +6.4 dB in SNR, +1.2 in PESQ, and +0.13 in STOI, along with reduced RT60 and enhanced clarity. The proposed approach offers both computational efficiency and interpretability, making it suitable for real-time deployment in teleconferencing, hearing-assistive, and smart audio applications.
Schlagwörter
Stichwörter der Autor:innen
Autor:innen und Institutionen
- School of Electrical & Information Engineering, Tianjin University, 92 Weijin Road, Nankai District, Tianjin 300072, China.
- Digital Signal Processing Laboratory, Tianjin University, 92 Weijin Road, Nankai District, Tianjin 300072, China.
ORCID-Profile: Samuel Yaw Mensah, Tao Zhang, Xin Zhao, Nahid-Al Mahmud
Zitieren
APA
Mensah, S. Y., Zhang, T., Zhao, X., & Mahmud, N. A. (2026). Model-Informed Speech Enhancement Using Virtual Room Acoustics and Acoustic Descriptor Optimization. Sensors, 26(12), 3630. https://doi.org/10.3390/s26123630
Vancouver
Mensah SY, Zhang T, Zhao X, Mahmud NA. Model-Informed Speech Enhancement Using Virtual Room Acoustics and Acoustic Descriptor Optimization. Sensors. 2026;26(12):3630. doi:10.3390/s26123630
BibTeX
@article{Mensah2026,
author = {Mensah, Samuel Yaw and Zhang, Tao and Zhao, Xin and Mahmud, Nahid-Al},
title = {Model-Informed Speech Enhancement Using Virtual Room Acoustics and Acoustic Descriptor Optimization},
journal = {Sensors},
year = {2026},
volume = {26},
number = {12},
pages = {3630},
doi = {10.3390/s26123630}
}
EUVIMED