Single Microphone Wind Noise Suppression via Adaptive Filtering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing telephony systems struggle to effectively suppress non-stationary wind noise in audio signals received by single microphones, leading to degraded perceptual quality and intelligibility of speech signals, especially in outdoor environments where wind noise is prevalent and difficult to distinguish from speech using conventional noise suppression algorithms.
Innovation Solution
A method that analyzes audio signal frames to detect non-stationary noise, applying a first filter for frames identified as noise and a second filter, such as a high-pass filter, for frames containing speech and noise, using a combination of tests including frequency sub-band analysis and energy level comparisons to differentiate between wind noise and speech.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional noise suppression algorithms are used, then system complexity is reduced, but wind noise suppression effectiveness deteriorates because wind noise is non-stationary and resembles speech
Solution Approach 1:
The patent applies dynamics by making the noise suppression characteristics adaptive rather than fixed. The system dynamically adjusts suppression parameters based on real-time detection of wind noise conditions, transitioning between different suppression strategies (e.g., spectral subtraction vs. simple attenuation) depending on the detected noise type and severity, thereby maintaining effectiveness across varying environmental conditions without requiring overly complex unified algorithms
Solution Approach 2:
The patent utilizes parameter changes by modifying key parameters of the noise suppression algorithm based on detected wind conditions. Specifically, the system changes suppression thresholds, frequency band selections, and attenuation factors according to the detected wind speed and noise characteristics, allowing the same algorithm framework to effectively handle diverse wind noise scenarios through parameter adaptation rather than structural complexity
2Object-affected harmful factors
If aggressive wind noise suppression is applied, then wind noise reduction is improved, but speech quality and intelligibility deteriorate due to over-suppression or distortion
Solution Approach 1:
The patent applies local quality by implementing frequency-selective suppression rather than uniform attenuation across all frequencies. The system identifies specific frequency bands where wind noise predominates (typically lower frequencies) and applies targeted suppression only to those bands, while preserving higher frequency speech components. This localized approach reduces wind noise impact on specific spectral regions without degrading overall speech quality
Solution Approach 2:
The patent utilizes feedback mechanisms where the system continuously monitors the processed audio signal and adjusts suppression parameters based on the resulting speech quality metrics. The feedback loop detects when suppression is too aggressive and automatically reduces attenuation levels or modifies frequency band selection to preserve speech intelligibility, creating a self-regulating system that balances noise reduction with speech quality preservation
3Reliability
If multiple microphones are used for wind noise detection, then wind noise suppression effectiveness is improved through correlation, but device complexity and power consumption increase
Solution Approach 1:
The patent applies the intermediary principle by introducing a virtual or synthesized reference signal that mediates the wind noise detection process. Instead of requiring multiple physical microphones, the system uses a single microphone signal processed through a virtual filter bank to create multiple virtual channels, where the reference signal is derived from the processed audio itself. This intermediary approach enables correlation-based wind noise detection without the physical complexity of multiple microphones
Solution Approach 2:
The patent utilizes copying by creating virtual representations of the audio signal through digital signal processing. The system copies the single microphone input signal through multiple virtual filter banks and processing paths, generating synthetic multi-channel signals that replicate the functionality of physical microphone arrays. These copied signals are then used for wind noise correlation analysis, achieving multi-microphone functionality with a single physical sensor
Data Source
AI summary
A technique for suppressing non-stationary noise, such as wind noise, in an audio signal is described. In accordance with the technique, a series of frames of the audio signal is analyzed to detect whether the audio signal comprises non-stationary noise. If it is detected that the audio signal comprises non-stationary noise, a number of steps are performed. In accordance with these steps, a determination is made as to whether a frame of the audio signal comprises non-stationary noise or speech and non-stationary noise. If it is determined that the frame comprises non-stationary noise, a first filter is applied to the frame and if it is determined that the frame comprises speech and non-stationary noise, a second filter is applied to the frame.