Single Microphone Wind Noise Detection via Correlation and Power Ratios
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current wind detection techniques in audio devices require multiple microphones or significant processing resources, making them inefficient for real-time wind detection using a single microphone.
Innovation Solution
A method and system that utilize a single microphone to detect wind noise by analyzing the correlation metric between successive audio frames and power ratio differences across specific frequency ranges, allowing for efficient wind noise identification without requiring multiple microphones or excessive processing resources.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If multiple microphones are used for wind detection, then detection accuracy is improved, but device complexity increases
Solution Approach 1:
The patent extracts and utilizes only the audio signal from a single microphone, eliminating the need for multiple microphones. By focusing on extracting meaningful wind detection information from one microphone signal through correlation metrics and power ratio analysis, the system achieves accurate wind detection without requiring multiple microphones, thus reducing device complexity while maintaining measurement precision.
2Measurement precision
If neural networks are used for single-microphone wind detection, then detection capability is improved, but processing resources required increase
Solution Approach 1:
The patent employs simple correlation metrics and power ratio calculations instead of complex neural networks. These lightweight computational methods require minimal processing resources and can be implemented efficiently with standard signal processing techniques. The approach uses basic mathematical operations (correlation calculations and power ratio comparisons) that are computationally inexpensive compared to training and executing neural network models, thereby reducing energy consumption while maintaining adequate wind detection capability.
3Use of energy by moving object
If simple single-microphone analysis is used, then processing resources are reduced, but detection accuracy deteriorates
Solution Approach 1:
The patent transforms the audio signal into different parameter domains to enhance detection accuracy. Specifically, it calculates correlation metrics between audio frames and computes power ratios at different frequency ranges. By changing the representation of the audio signal into these derived parameters (correlation values and power ratios), the system can effectively distinguish wind noise from other sounds using simple comparisons, achieving both low processing resource requirements and adequate detection accuracy.
Data Source
AI summary
A method for detecting wind noise incident on a single microphone may include receiving an audio signal indicative of sound incident on the single microphone, dividing the audio signal into a plurality of audio frames, and determining whether wind noise is incident on the single microphone based on a combination of a correlation metric between successive audio frames of the plurality of audio frames and a power ratio difference between a first power ratio and a second power ratio. The first power ratio may equal an amount of power present in a first frequency range of the audio signal to a total amount of power present in the audio signal across all frequencies. The second power ratio may equal an amount of power present in a second frequency range of the audio signal to the total amount of power present in the audio signal across all frequencies.