Earphone Voice Activity Detection Using Dual-Microphone Gain Adaptation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Robust voice activity detection in noisy environments is challenging for small devices mounted in the ear due to issues with false positives and negatives using fixed thresholds.
Innovation Solution
A method utilizing two microphones, one ambient and one ear-canal, with band pass filtering and time-weighted power estimation, followed by calculating and comparing fast and slow time-weighted ratios to adjust gain based on voice activity, optionally using SVD or neural nets for determination.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If fixed thresholds are used for voice activity detection, then the system is simple to implement, but false positives and false negatives increase in noisy environments
Solution Approach 1:
The patent implements dynamic threshold adjustment by continuously adapting the detection threshold based on the statistical characteristics of the ambient noise. Instead of using a fixed threshold, the system calculates the mean and standard deviation of the noise signal and sets the threshold as a function of these parameters, allowing the threshold to automatically adapt to changing noise conditions and maintain reliable detection accuracy.
Solution Approach 2:
The system changes the detection parameter from a fixed threshold value to a dynamically calculated threshold based on noise statistics. By computing the mean and standard deviation of the noise signal and using these to determine the threshold, the system transforms the detection criterion to match the actual acoustic environment, thereby reducing false positives and negatives.
2Reliability
If multiple microphones and complex processing are used to improve detection accuracy, then voice activity detection reliability improves, but device complexity and power consumption increase
Solution Approach 1:
The patent extracts only the essential statistical features (mean and standard deviation) from the noise signal that are sufficient for threshold determination. By focusing on these two key parameters rather than performing full spectral analysis or complex signal processing, the system achieves reliable detection while minimizing computational load and power consumption.
Solution Approach 2:
The system uses the ambient noise signal itself to generate the detection threshold, rather than requiring external calibration or complex reference signals. The noise signal provides its own statistical characteristics (mean and standard deviation) that directly determine the threshold, allowing the system to adapt automatically without additional processing overhead.
3Reliability
If multiple microphones and complex processing are used to improve detection accuracy, then voice activity detection reliability improves, but device complexity increases
Solution Approach 1:
The patent segments the signal processing into distinct functional stages: noise estimation, threshold calculation, and voice activity decision. By separating these functions and processing them in sequence, the system reduces the complexity of each individual stage while maintaining overall detection accuracy. The noise estimation stage independently computes statistical parameters, which then feed into the simple threshold comparison stage.
Data Source
AI summary
An electronic device or method for adjusting a gain on a voice operated control system can include one or more processors and a memory having computer instructions. The instructions, when executed by the one or more processors causes the one or more processors to perform the operations of receiving a first microphone signal, receiving a second microphone signal, updating a slow time weighted ratio of the filtered first and second signals, and updating a fast time weighted ratio of the filtered first and second signals. The one or more processors can further perform the operations of calculating an absolute difference between the fast time weighted ratio and the slow time weighted ratio, comparing the absolute difference with a threshold, and increasing the gain when the absolute difference is greater than the threshold. Other embodiments are disclosed.