Voice Activity Detection Using Dual Power Estimates
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing voice activity detection (VAD) approaches in microphones suffer from high false detections, increased part counts, and high current consumption, leading to user dissatisfaction in devices like cellular phones.
Innovation Solution
The implementation of efficient VAD methods that utilize dual microphones with voice activity detection modules, filtering sound energy into multiple bands, and determining voice activity based on power estimates, with the ability to vary clock frequencies and enter event detection modes to reduce false triggers and power consumption.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional VAD algorithms are used, then voice detection capability is provided, but false detections increase and reliability decreases
Solution Approach 1:
The patent segments the audio signal into multiple frequency bands using filter banks before processing. This segmentation allows the system to analyze different frequency components separately, improving voice detection accuracy by identifying characteristic voice frequency patterns while filtering out non-voice sounds, thereby reducing false detections
Solution Approach 2:
The patent applies different processing characteristics to different frequency bands. Each filter bank channel processes its specific frequency range with appropriate parameters, allowing the system to optimize detection for voice frequencies while being less sensitive to interfering frequencies, thus improving reliability without increasing false positives
2Reliability
If complex VAD processing is performed, then detection accuracy improves, but computational complexity and part counts increase
Solution Approach 1:
By dividing the audio signal into discrete frequency bands through filter banks, the patent simplifies the processing complexity. Instead of analyzing the entire complex audio spectrum, the system processes multiple simpler frequency channels independently, reducing computational burden while maintaining detection accuracy
Solution Approach 2:
The patent processes only the necessary frequency bands and uses threshold-based decision making rather than exhaustive analysis. This partial action approach processes only relevant portions of the signal and uses simple comparison operations, reducing computational complexity while achieving sufficient detection accuracy
3Speed
If continuous VAD monitoring is performed, then real-time detection is achieved, but current consumption increases
Solution Approach 1:
The patent implements periodic VAD monitoring rather than continuous processing. The system periodically samples the audio signal at appropriate intervals, processes it through the filter banks, and makes detection decisions only when needed, maintaining real-time detection capability while significantly reducing power consumption compared to continuous monitoring
Solution Approach 2:
The patent dynamically adjusts processing intensity based on detection needs. The system can transition between low-power standby mode and active processing mode, processing audio only when voice activity is detected or during predetermined intervals, thereby achieving real-time detection capability while optimizing power consumption
Data Source
AI summary
A microphone assembly includes an acoustic sensor and a voice activity detector on an integrated circuit coupled to an external-device interface. The acoustic sensor produces an electrical signal representative of acoustic energy detected by the sensor. A filter bank separates data representative of the acoustic energy into a plurality of frequency bands. A power tracker obtains a power estimate for at least one band, including a first estimate based on relatively fast changes in a power metric of the data and a second estimate based on relatively slow changes in a power metric of the data. The presence of voice activity in the electrical signal is based upon the power estimate.