Semi-Adaptive Beamformer With Fixed and Instantaneous RTFs
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Adaptive beamformers require time to converge and may distort or converge in the wrong direction, especially in low-SNR environments, affecting speech quality in microphone arrays.
Innovation Solution
A semi-adaptive beamformer that combines instantaneous and fixed relative transfer functions (RTFs) to quickly determine an MVDR beamforming filter, using a weighted average adjusted by SNR conditions, to enhance speech and suppress noise without distorting the signal.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If adaptive beamforming is used to optimize speech quality, then signal-to-noise ratio improves, but convergence time increases and transient distortion occurs
Solution Approach 1:
The patent applies preliminary action by pre-computing and storing RTF lookup tables during an offline training phase. During operation, the system queries these pre-computed tables rather than computing RTFs in real-time, eliminating convergence delay while maintaining adaptive beamforming capabilities. This resolves the contradiction by preparing necessary data structures beforehand.
Solution Approach 2:
The patent implements dynamics by switching between different beamforming modes (adaptive vs. fixed) based on convergence state detection. When the beamformer has converged, it transitions to fixed mode for low-latency operation; when not converged, it operates in adaptive mode to optimize SNR. This dynamic mode switching resolves the contradiction by adapting the system behavior to current operational conditions.
2Speed
If adaptive beamforming converges quickly, then response time improves, but accuracy deteriorates in low-SNR environments
Solution Approach 1:
The patent applies feedback by continuously monitoring the convergence state of the beamformer and using this information to switch between adaptive and fixed modes. The feedback mechanism ensures that the system only uses adaptive mode when necessary and switches to fixed mode when convergence is detected or SNR is insufficient, preventing inaccurate quick convergence in low-SNR environments.
Solution Approach 2:
The patent changes the operating parameters by switching between adaptive beamforming (for high SNR conditions) and fixed beamforming (for low SNR or converged states). This parameter change allows the system to achieve fast response when conditions permit while maintaining accuracy when conditions require it, resolving the contradiction between speed and precision.
3Object-generated harmful factors
If adaptive beamforming is used to suppress noise, then noise power reduces, but speech distortion increases during convergence
Solution Approach 1:
The patent implements dynamics by switching between adaptive and fixed beamforming modes based on convergence state. During the convergence transition period, the system can switch to fixed mode to prevent speech distortion while maintaining noise suppression capability, then switch back to adaptive mode when convergence is achieved. This dynamic switching resolves the contradiction between noise suppression and speech fidelity.
4Device complexity
If fixed RTFs are used for fast processing, then computational complexity reduces, but adaptability to environmental changes deteriorates
Solution Approach 1:
The patent implements dynamics by switching between fixed RTF mode (for low complexity) and adaptive RTF mode (for high adaptability) based on environmental conditions and convergence state. This allows the system to use simple fixed RTFs during stable conditions while switching to adaptive RTFs when environmental changes occur, resolving the contradiction between complexity and adaptability.
Solution Approach 2:
The patent changes the RTF parameter from fixed to adaptive based on operational conditions. When the environment is stable and the beamformer has converged, the system uses fixed RTFs for low complexity. When environmental changes are detected or convergence is needed, the system switches to adaptive RTF computation, resolving the contradiction by adapting the parameter choice to current needs.
Data Source
AI summary
This disclosure provides methods, devices, and systems for beamforming. The present implementations more specifically relate to semi-adaptive beamforming techniques. In some aspects, a semi-adaptive beamformer may determine an RTF vector based on an audio signal received via a microphone array (also referred to as an “instantaneous” RTF vector) and may further determine an MVDR beamforming filter for the microphone array based on a combination of the instantaneous RTF vector and a “fixed” RTF vector. The fixed RTF vector may include a set of RTFs that are known to produce a relatively accurate MVDR beamforming filter for any users of the microphone array. In some implementations, the semi-adaptive beamformer may determine the MVDR beamforming filter based on a weighted average of the instantaneous RTF vector and the fixed RTF vector, where the weighting can be dynamically adjusted based on the quality of the received audio signal or various other conditions.