Synthetic Wave-Geometric Impulse Responses for Improved Speech Dereverberation
This work addresses speech quality improvement for applications like hearing aids or communication systems, but it is incremental as it builds on existing synthetic data approaches.
The paper tackles speech dereverberation by using a hybrid synthetic dataset (GWA) that combines wave-based and geometric methods to simulate Room Impulse Responses (RIRs), showing that models trained on this dataset outperform those trained on prior geometric methods across four real-world datasets.
We present a novel approach to improve the performance of learning-based speech dereverberation using accurate synthetic datasets. Our approach is designed to recover the reverb-free signal from a reverberant speech signal. We show that accurately simulating the low-frequency components of Room Impulse Responses (RIRs) is important to achieving good dereverberation. We use the GWA dataset that consists of synthetic RIRs generated in a hybrid fashion: an accurate wave-based solver is used to simulate the lower frequencies and geometric ray tracing methods simulate the higher frequencies. We demonstrate that speech dereverberation models trained on hybrid synthetic RIRs outperform models trained on RIRs generated by prior geometric ray tracing methods on four real-world RIR datasets.