Method and apparatus for gradual speech enhancement and restoration: leveraging codec modules for improved denoising and dereverberation
A two-stage pipeline with a lightweight SE module and generative codec module effectively addresses noise and reverberation, enhancing speech quality in challenging environments.
WO2026117263A1PCT designated stage Publication Date: 2026-06-04TENCENT AMERICA LLC
Patent Information
- Authority / Receiving Office
- WO · WO
- Patent Type
- Applications
- Current Assignee / Owner
- TENCENT AMERICA LLC
- Filing Date
- 2025-04-22
- Publication Date
- 2026-06-04
AI Technical Summary
Technical Problem
Existing speech enhancement methods struggle to effectively address both noise and reverberation in real-world environments, leading to degraded speech quality in applications like telecommunications and hearing assistance devices.
Method used
A two-stage pipeline approach is employed, where a lightweight SE module first suppresses environmental noise, followed by a generative codec module that removes reverberation and restores speech naturalness, utilizing quantization techniques and a weighted loss function to preserve critical speech characteristics.
Benefits of technology
The method significantly improves speech quality in noisy and reverberant environments by minimizing distortions and ensuring clear, natural-sounding speech output.
✦ Generated by Eureka AI based on patent content.
Smart Images
Figure US2025025723_04062026_PF_FP_ABST
Abstract
A method includes inputting a mixture signal comprising speech, background noise, and reverberation into a denoising stage that generates a denoised mixture signal with reduced noise compared to the mixture signal; fusing the denoised mixture signal with the mixture signal to generate a fused signal; and inputting the fused signal into a dereverberation and restoration (DR) stage to generate an output signal with reduced reverberation compared to the mixture signal.
Need to check novelty before this filing date? Find Prior Art