Spherical Microphone Array Noise Reduction via Adaptive Wiener Filtering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Spherical microphone arrays on a rigid sphere face challenges in minimizing noise, particularly low-frequency noise, in Ambisonics representations due to the amplification of noise over higher order coefficients, which distorts the sound field representation.
Innovation Solution
The method computes a regularization parameter based on the signal-to-noise ratio (SNR) of the microphone array signals, using a time-variant Wiener filter to adapt the transfer function and reduce noise, thereby optimizing the Ambisonics representation for each frequency sub-band.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the inverse filter response is applied to remove microphone array impact, then the accuracy of Ambisonics representation is improved, but noise is amplified especially at low frequencies and high Ambisonics orders
Solution Approach 1:
The patent applies frequency-dependent regularization to the inverse filter response, where the regularization parameter λ varies with frequency and Ambisonics order. This modifies the filter characteristics to reduce noise amplification at low frequencies while maintaining accuracy at higher frequencies, directly resolving the contradiction between measurement precision and noise amplification.
Solution Approach 2:
The patent selectively applies regularization only to frequency ranges and Ambisonics orders where noise amplification occurs, rather than uniformly across all frequencies and orders. This partial application of the inverse filter response maintains accuracy where needed while suppressing noise where problematic.
2Measurement precision
If high Ambisonics orders are used to improve spatial resolution, then the spatial resolution is improved, but noise amplification increases at low frequencies
Solution Approach 1:
The patent applies different regularization strengths to different Ambisonics orders and frequency ranges. Higher Ambisonics orders receive stronger regularization at low frequencies to suppress noise, while maintaining full resolution at higher frequencies where noise is less problematic, thus achieving local optimization of quality.
3Object-generated harmful factors
If Tikhonov regularization is applied to reduce noise, then noise amplification is reduced, but spatial resolution deteriorates due to excessive smoothing
Solution Approach 1:
The patent uses adaptive regularization where the regularization parameter λ is dynamically adjusted based on frequency and Ambisonics order rather than being a fixed value. This dynamic adaptation allows the system to reduce noise where necessary while preserving spatial resolution where the signal is strong, avoiding excessive smoothing.
Data Source
AI summary
Spherical microphone arrays capture a three-dimensional sound field (P(Ωc, t)) for generating an Ambisonics representation (Anm(t)), where the pressure distribution on the surface of the sphere is sampled by the capsules of the array. The impact of the microphones on the captured sound field is removed using the inverse microphone transfer function. The equalization of the transfer function of the microphone array is a big problem because the reciprocal of the transfer function causes high gains for small values in the transfer function and these small values are affected by transducer noise. The invention minimizes that noise by using a Wiener filter processing in the frequency domain, which processing is automatically controlled per wave number by the signal-to-noise ratio of the microphone array.