Sensor Array Post-Filter for Directional Noise Suppression
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional beamformer post-filtering techniques fail to effectively improve the signal-to-noise ratio (SNR) for highly correlated point noise sources, as they assume noise is either incoherent or diffuse, and do not account for directional noise sources.
Innovation Solution
A Sensor Array Post-Filter that uses adaptive post-filtering based on a probabilistic model to accurately model and suppress both diffuse and directional noise sources, employing expectation-maximization (EM) based incremental Bayesian learning to adapt model parameters over time, allowing for improved SNR by accurately tracking signal and noise sources.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional post-filtering techniques are used, then processing simplicity is maintained, but SNR improvement fails for highly correlated point noise sources
Solution Approach 1:
The patent segments the noise field into distinct components: diffuse noise and directional point noise sources. This segmentation allows the post-filter to handle each noise type separately using appropriate processing techniques, improving SNR for correlated point noise while maintaining manageable complexity through targeted processing strategies.
Solution Approach 2:
The patent employs adaptive parameter adjustment by estimating the number of point noise sources and their characteristics dynamically. The post-filter adapts its parameters based on the statistical properties of the noise field, allowing it to effectively suppress correlated noise sources while maintaining flexibility and avoiding excessive complexity.
2Measurement precision
If conventional post-filtering assuming incoherent or diffuse noise is applied, then computational load is reduced, but tracking precision of directional noise sources deteriorates
Solution Approach 1:
The patent performs preliminary estimation of the number of point noise sources and their statistical characteristics before applying the full post-filtering process. This preliminary action prepares the system to accurately track directional noise sources while optimizing computational resources, balancing tracking precision with processing efficiency.
Solution Approach 2:
The patent replaces conventional mechanical filtering approaches with probabilistic modeling and statistical estimation methods. By using probability theory to model noise sources and their correlations, the system achieves high tracking precision for directional noise while maintaining computational efficiency through mathematical rather than brute-force processing.
3Reliability
If adaptive probabilistic modeling is implemented, then noise suppression accuracy improves, but computational complexity increases
Solution Approach 1:
The patent develops a universal probabilistic framework that can handle multiple noise types (diffuse and directional) and varying numbers of noise sources through a single cohesive model. This multi-functional approach improves noise suppression effectiveness across different scenarios while avoiding the need for multiple separate complex models, thereby managing overall system complexity.
Data Source
AI summary
A “Sensor Array Post-Filter” provides an adaptive post-filter that accurately models and suppresses both diffuse and directional noise sources as well as interfering speech sources. The post-filter is applied to an output signal produced by a beamformer used to process signals produced by a sensor array. As a result, the Sensor Array Post-Filter operates to improve the signal-to-noise ratio (SNR) of beamformer output signals by providing adaptive post-filtering of the output signals. The post-filter is generated based on a generative statistical model for modeling signal and noise sources at distinct regions in a signal field that considers prior distributions trained to model an instantaneous direction of arrival for signals captured by sensors in the array.