Signal Enhancement Device Iterative Parameter Estimation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current signal enhancement technologies fail to effectively address noisy reverberant environments, where both noise and reverberation are present, as they require separate processing steps that are not seamlessly integrated, leading to suboptimal noise reduction and reverberation removal.
Innovation Solution
A parameter estimation unit that iteratively updates reverberation, noise, and source parameter estimates using a maximum likelihood approach, employing a variation of the EM algorithm to jointly reduce noise and reverberation in a noisy reverberant signal, enhancing the source signal by considering the interdependence of these parameters.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If separate processing steps are used for noise reduction and reverberation removal, then the processing structure is simple, but the noise reduction and reverberation removal effectiveness is suboptimal
Solution Approach 1:
The patent merges noise reduction and reverberation removal into a unified iterative processing framework where both operations are performed simultaneously through alternating optimization of noise parameters and reverberation parameters, achieving enhanced effectiveness while maintaining reasonable structural complexity
Solution Approach 2:
The unified framework serves multiple functions: it performs both noise reduction and reverberation removal, estimates both noise and reverberation parameters, and handles their interdependence, making the processing structure multi-functional and adaptable to noisy reverberant environments
2Reliability
If noise and reverberation are processed independently, then the processing is computationally efficient, but the interdependence of noise and reverberation parameters is not considered
Solution Approach 1:
The patent employs periodic alternating optimization where noise parameter estimation and reverberation parameter estimation are performed in alternating iterations, allowing the system to consider parameter interdependence while maintaining computational efficiency through structured periodic processing
Solution Approach 2:
The iterative framework maintains continuous improvement of source signal enhancement by repeatedly alternating between noise and reverberation parameter estimations, ensuring that both parameters are continuously refined considering their interdependence until convergence
Data Source
AI summary
The initial values of parameter estimates are set, including reverberation parameter estimates, which includes a regression coefficient used in a linear convolutional operation for calculating an estimated value of reverberation included in an observed signal, source parameter estimates, which includes estimated values of a linear prediction coefficient and a prediction residual power that identify the power spectrum of a source signal, and noise parameter estimates, which include noise power spectrum estimates. Then, the maximum likelihood estimation is used to alternately repeat processing for updating at least one of the reverberation parameter estimates and the noise parameter estimates and processing for updating the source parameter estimates until a predetermined termination condition is satisfied.


