Speech Enhancement Circuit With Adaptive Multi-Stage Noise Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing speech enhancement technologies are inadequate in effectively suppressing noise, which affects the quality of voice calls.
Innovation Solution
A processing circuit and method that employs Fourier transform, multiple noise reduction processes, and inverse Fourier transform to generate a target signal, utilizing both deep learning-based and signal processing-based noise reduction techniques to improve noise suppression.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a single noise reduction processing is applied to the spectral signal, then the processing complexity is low, but the noise suppression effect is insufficient
Solution Approach 1:
The patent divides the noise reduction process into multiple sequential stages: first noise reduction processing followed by second noise reduction processing. Each stage uses different processing methods tailored to different noise characteristics, thereby improving overall noise suppression effectiveness while managing processing complexity through structured segmentation of the enhancement workflow
Solution Approach 2:
The patent implements dynamic selection of noise reduction methods based on real-time noise analysis. The system analyzes noise features in the intermediate signal and adaptively chooses between different second noise reduction processing methods (such as spectral subtraction, Wiener filtering, or deep learning-based methods) according to the detected noise characteristics, optimizing the balance between noise suppression and processing complexity
2Reliability
If multiple different noise reduction processes are applied sequentially, then the noise suppression effect is improved, but the processing time increases
Solution Approach 1:
The patent performs preliminary noise analysis on the intermediate signal before applying the second noise reduction processing. This preliminary action identifies the noise characteristics and pre-determines the most suitable processing method, avoiding trial-and-error approaches and reducing the overall processing time by preparing the optimal processing path in advance
Solution Approach 2:
The system dynamically selects from multiple second noise reduction processing methods based on noise analysis results. By choosing only the necessary processing method rather than applying all possible methods, the system achieves effective noise suppression while minimizing processing time through adaptive method selection
Data Source
AI summary
A processing circuit performing a speech enhancement method processes a to-be-processed signal to generate a target signal and executes a plurality of program codes or program instructions to perform the following steps: performing Fourier transform on the to-be-processed signal to generate a spectral signal of the to-be-processed signal; performing a first noise reduction processing on the spectral signal to obtain a first intermediate signal; performing a noise analysis on the first intermediate signal to obtain a noise feature; performing a second noise reduction processing on the first intermediate signal to generate a second intermediate signal when the noise feature does not satisfy a target condition; and performing inverse Fourier transform on the second intermediate signal to generate the target signal. The first noise reduction processing is different from the second noise reduction processing.