DisSR: Disentangling Speech Representation for Degradation-Prior Guided Cross-Domain Speech Restoration
Abstract
Previous speech restoration (SR) primarily focuses on single-task speech restoration (SSR), which cannot address general speech restoration problems. Training specific SSR models for different distortions is time-consuming and lacks generality. In addition, most studies ignore the problem of model generalization across unseen domains. To overcome those limitations, we propose DisSR, a Disentangling Speech Representation based general speech restoration model with two properties: 1) Degradation-prior guidance, which extracts speaker-invariant degradation representation to guide the diffusion-based speech restoration model. 2) Domain adaptation, where we design cross-domain alignment training to enhance the model's adaptability and generalization on cross-domain data, respectively. Experimental results demonstrate that our method can produce high-quality restored speech under various distortion conditions. Audio samples can be found at https://itspsp.github.io/DisSR.
Cite
@article{arxiv.2602.12701,
title = {DisSR: Disentangling Speech Representation for Degradation-Prior Guided Cross-Domain Speech Restoration},
author = {Ziqi Liang and Zhijun Jia and Chang Liu and Minghui Yang and Zhihong Lu and Jian Wang},
journal= {arXiv preprint arXiv:2602.12701},
year = {2026}
}
Comments
Accepted to 2026 IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP 2026)