3 papers
eess.AS2026
Weakly Guided and Autoregressive Beamformer Parameterization for Generalizable Moving Speaker Extraction in Higher-Order Ambisonics
Jakob Kienegger, Tal Peer, Sina Khanagha +1
Linear spatial filters (beamformers) enable robust, generalizable and interpretable speech enhancement with performance guarantees under ideal parameterization. Modern beamformers…
eess.AS2026
Your U-Net Dereverberation Model is Secretly an RIR Encoder
Sina Khanagha, Timo Gerkmann
In this work, we analyze the ability of NCSN++ U-Net based audio dereverberation models to capture global room characteristics in their intermediate representations. Through an emp…
eess.AS2026
Bone-conduction Guided Multimodal Speech Enhancement with Conditional Diffusion Models
Sina Khanagha, Bunlong Lay, Timo Gerkmann
Single-channel speech enhancement models face significant performance degradation in extremely noisy environments. While prior work has shown that complementary bone-conducted spee…