Headset Audio Signal Processing via Two-Stage Beamforming
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing audio signal processing systems using multiple microphones and beamforming techniques face challenges in effectively canceling noise, particularly wind noise and background sounds, due to varying noise levels and spatial configurations of microphones, which affect the signal-to-noise ratio and introduce distortion.
Innovation Solution
A two-stage beamforming approach is implemented, where the first and second beamformers adapt spatial sensitivity to emphasize desired signals and cancel noise locally, and a third beamformer dynamically combines signals from both sides to optimize noise suppression, leveraging the near-field effect and different noise levels on either side of the head to improve signal quality.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a single beamforming stage is used, then the device complexity is low, but the noise cancellation effectiveness is insufficient
Solution Approach 1:
The beamforming process is divided into two distinct stages: a first beamforming stage that processes microphone signals to produce intermediate beamformed signals, and a second beamforming stage that processes these intermediate signals to produce final output. This segmentation allows each stage to perform specific noise cancellation functions, improving overall effectiveness while managing complexity through modular design.
Solution Approach 2:
The second beamforming stage is nested within the overall signal processing chain, taking as input the outputs from the first beamforming stage. This nested structure enables progressive noise cancellation where the second stage refines the results of the first stage, achieving superior noise cancellation effectiveness through hierarchical processing.
2Measurement precision
If microphones are positioned far apart, then the spatial sampling capability is improved, but wind noise amplification occurs
Solution Approach 1:
The first beamforming stage applies local spatial filtering by processing signals from individual microphone pairs with closely spaced microphones. This local processing enhances spatial sampling capability for direction-of-arrival estimation while the close spacing of microphones within each pair prevents wind noise amplification that would occur with widely spaced microphones.
Solution Approach 2:
The intermediate beamformed signals produced by the first beamforming stage serve as a mediator between the raw microphone signals and the final output. This intermediate representation allows the system to capture spatial information effectively while avoiding direct amplification of wind noise, as the first stage processes and conditions the signals before they reach the second beamforming stage.
3Measurement precision
If beamforming is applied aggressively to cancel noise, then the signal-to-noise ratio is improved, but distortion of the desired signal is introduced
Solution Approach 1:
The first beamforming stage applies partial beamforming action to produce intermediate signals with improved signal-to-noise ratio, while the second beamforming stage applies additional processing to further enhance noise cancellation. This progressive, partial action approach allows the system to achieve high signal-to-noise ratio improvement without introducing excessive distortion that would result from a single aggressive beamforming stage.
Solution Approach 2:
The second beamforming stage processes the intermediate beamformed signals with feedback from the first stage's output, allowing adaptive refinement of the noise cancellation. This feedback mechanism enables the system to achieve superior signal-to-noise ratio while minimizing distortion, as the second stage can adjust its processing based on the characteristics of the intermediate signals.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A headset and a method configured to process audio signals from multiple microphones, comprising: a first pair of microphones (101,102) outputting a first pair of microphone signals and a second pair of microphones (103, 104) outputting a second pair of microphone signals; a first near-field beamformer (105) and a second near-field beamformer (106) each configured to receive a pair of microphone signals and adapt the spatial sensitivity of a respective pair of microphones as measured in a respective beamformed signal (XL; XR) output from a respective beamformer (105; 106); wherein the spatial sensitivity is adapted to suppress noise relative to a desired signal; a third beamformer (107) configured to dynamically combine the signals (XL; XR) output from the first beamformer (105) and the second beamformer (106) into a combined signal (Xc); wherein the signals are combined such that signal energy in the combined signal is minimized while a desired signal is preserved; and a noise reduction unit (109) configured to process the combined signal (Xc) from the third beamformer (107) and output the combined signal such that noise is reduced.