Adaptive Equalization System for Speech Intelligibility
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current speech enhancement systems fail to effectively improve speech intelligibility, especially in noisy environments, as they often require voicing decisions or prior knowledge of clean speech and noise levels, which may not be feasible in real-time applications.
Innovation Solution
An adaptive equalization system that automatically adjusts the spectral shape of speech signals using a digital signal processor with modules for subband processing, noise estimation, and intelligibility measurement, allowing for real-time improvement of speech intelligibility without requiring voicing decisions or advanced noise level knowledge.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If speech enhancement algorithms are used to suppress background noise, then speech quality is improved, but speech intelligibility is not significantly improved
Solution Approach 1:
The patent changes the parameter being optimized from general speech quality to speech intelligibility specifically. It uses an intelligibility metric that measures the ratio of speech signal power to noise power in different frequency bands, and adjusts enhancement parameters to maximize this metric rather than just overall signal quality.
Solution Approach 2:
The patent applies different enhancement strategies to different frequency bands based on their importance for intelligibility. It identifies critical frequency regions where speech intelligibility is most affected by noise and applies targeted enhancement to these specific bands rather than uniform enhancement across all frequencies.
2Reliability
If algorithms requiring voicing decisions are used, then speech intelligibility can be improved, but system complexity increases and performance degrades in noisy environments
Solution Approach 1:
The patent extracts and removes the complex voicing decision component from the enhancement system. Instead of requiring separate voicing detection modules and decisions, it uses a continuous intelligibility metric that works directly on the noisy signal without needing to identify speech versus non-speech segments explicitly.
3Reliability
If algorithms requiring additional training or prior knowledge of clean speech and noise levels are used, then speech intelligibility can be improved, but ease of operation decreases and applicability to real-time systems is limited
Solution Approach 1:
The patent implements a self-adjusting system that automatically estimates noise levels and optimizes enhancement parameters in real-time without requiring external training data or prior knowledge of clean speech characteristics. The intelligibility metric continuously adapts to the current acoustic environment and adjusts enhancement accordingly.
Data Source
Figure 1
Figure 2
Figure 3~4
AI summary
An adaptive equalization system that adjusts the spectral shape of a speech signal based on an intelligibility measurement of the speech signal may improve the intelligibility of the output speech signal. Such an adaptive equalization system may include a speech intelligibility measurement module, a spectral shape adjustment module, and an adaptive equalization module. The speech intelligibility measurement module is configured to calculate a speech intelligibility measurement of a speech signal. The spectral shape adjustment module is configured to generate a weighted long-term speech curve based on a first predetermined long-term average speech curve, a second predetermined long-term average speech curve, and the speech intelligibility measurement. The adaptive equalization module is configured to adapt equalization coefficients for the speech signal based on the weighted long-term speech curve.