Noise Reference Estimation Using Adaptive Sidelobe Cancellation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Modern headset communication devices face challenges in accurately isolating and reducing non-stationary noise sources due to the limitations of existing Voice Activity Detection (VAD) and Noise Reduction (NR) systems, which are ineffective in environments with significant ambient noise, leading to inaccuracies in noise estimation and poor performance in noise reduction.
Innovation Solution
A device and method utilizing multiple microphones with a generalized sidelobe canceller, adaptive noise estimation module, and smoothing filter to produce enhanced noise reference signals, effectively separating speech and noise signals and improving noise estimation by leveraging multi-microphone information and spectral subtraction in the frequency domain.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional Voice Activity Detection and Noise Reduction systems are used, then device complexity is kept low, but noise estimation accuracy deteriorates in non-stationary noise environments
Solution Approach 1:
The system segments the noise estimation task into multiple components: a Generalized Sidelobe Canceller for spatial noise reduction, an adaptive noise estimation module for spectral analysis, and a smoothing filter for temporal refinement. Each module handles a specific aspect of noise characterization, improving overall accuracy without requiring a complete system redesign.
Solution Approach 2:
The patent introduces spatial dimensionality by using multiple microphones arranged in an array. The Generalized Sidelobe Canceller exploits spatial information to separate noise from speech signals, adding a dimensional approach that traditional single-microphone systems lack. This spatial processing significantly improves noise estimation accuracy in non-stationary environments.
2Reliability
If multiple microphones with spatial processing are used, then noise reference signal quality is improved, but device complexity increases
Solution Approach 1:
The system performs preliminary noise reference signal generation through the Generalized Sidelobe Canceller before the main noise estimation process. By pre-processing the microphone signals to create a cleaned noise reference, the system reduces the computational burden on subsequent stages and improves the reliability of the final noise reduction.
Solution Approach 2:
The adaptive noise estimation module continuously refines the noise reference signal by comparing estimated noise with actual noise measurements and adjusting its parameters accordingly. This feedback mechanism ensures that the system adapts to changing noise conditions, maintaining high reliability without requiring manual intervention or system reconfiguration.
3Measurement precision
If adaptive filtering and spectral subtraction are applied, then signal separation accuracy is improved, but processing time increases
Solution Approach 1:
The smoothing filter applies partial filtering by selectively processing only the necessary frequency bands and time frames where noise estimation is most critical. This approach achieves adequate separation accuracy without performing exhaustive processing on all signal components, thereby reducing overall processing time while maintaining effective speech-noise separation.
Data Source
AI summary
A device for noise estimation comprises a first microphone capturing a nominal speech signal, and a second microphone capturing a nominal noise signal. A generalized sidelobe canceller of the device applies spatial noise reduction, and comprises a blocking matrix filter to adaptively process the nominal speech signal to produce a speech cancellation signal, a node for subtracting the speech cancellation signal from the nominal noise signal to produce a noise reference signal, a noise cancellation filter to adaptively filter the noise reference signal to produce a noise cancellation signal; and a node for subtracting the noise cancellation signal from the nominal speech signal to produce a speech reference signal. An adaptive noise estimation module of the device comprises a node for subtracting the noise reference signal from the speech reference signal and outputting a difference signal; a smoothing filter for filtering the difference signal to produce a long term difference signal; and a node for adding the long term difference signal to the noise reference signal to produce an enhanced noise reference signal.