Multi-Microphone Beamforming for Adaptive Spatial Noise Suppression
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for spatial noise suppression in multi-microphone devices struggle to effectively separate desired audio signals from unwanted ambient noise and interfering sounds, particularly in the presence of noise and limited directional capabilities.
Innovation Solution
An apparatus and method that determine audio data with different directivity configurations to capture sound from the same direction, apply beamform weights, and generate a more omnidirectional directivity configuration to suppress noise, using a noise suppression parameter applied to microphone audio signals.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Length of moving object
If beamforming is used to capture sound from a specific direction, then directional audio capture is improved, but the ability to capture omnidirectional sound deteriorates
Solution Approach 1:
The system dynamically switches between different directivity configurations (directional and omnidirectional) based on the audio scene requirements. Multiple beamformers with different directivity patterns are maintained ready, and the appropriate one is selected or blended in real-time to achieve both directional and omnidirectional capture capabilities adaptively
Solution Approach 2:
The audio processing system is designed to perform multiple functions by supporting both directional beamforming and omnidirectional capture modes within the same apparatus. Multiple beamformers with different directivity configurations enable the system to serve both specialized directional recording and general omnidirectional recording needs
2Object-affected harmful factors
If adaptive beamforming weights are used to suppress noise, then noise suppression is improved, but the complexity of processing increases
Solution Approach 1:
The system applies adaptive beamforming selectively to specific frequency bands where noise suppression is most beneficial, rather than uniformly across all frequencies. This partial application reduces the overall processing complexity while maintaining effective noise suppression where it matters most
Solution Approach 2:
The audio processing is divided into multiple frequency bands, each handled by separate beamforming operations. This segmentation allows the complex adaptive processing to be distributed across multiple simpler parallel operations, reducing the complexity burden on any single processing stage
3Object-affected harmful factors
If post-filtering is applied to suppress interfering energy, then noise suppression is improved, but the quality of desired sound may deteriorate
Solution Approach 1:
Different directivity configurations are applied to different spatial regions and frequency bands. Directional beamformers with tailored patterns suppress interfering sounds from specific directions while preserving desired sounds from target directions. The post-filtering is localized to specific spatial and spectral regions rather than applied uniformly, preventing degradation of desired sound quality
Data Source
AI summary
An apparatus including circuitry configured to: obtain at least two microphone audio signals; determine audio data including different directivity configurations that are able to capture sound from substantially a same or similar direction; determine at least one value related to the sound arriving from at least the same or similar direction based on the audio data; determine further audio data including at least one configuration which provides a more omnidirectional directivity configuration than the audio data; determine at least one value related to the sound based on the further audio data; and determine a noise suppression parameter based on the at least one value related to the arriving sound and the value related to the sound. The spatial noise suppression parameter is configured to be applied to the microphone audio signals in the generation of a playback audio signal.