Analog Voice Activity Detection Circuit for Low Power Always-Listening Applications
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current audio sensor technologies face challenges in integrating voice control capabilities due to high sensor bandwidths and computing loads, leading to inefficiencies in power consumption and form factor requirements, particularly in always-listening applications where traditional digital signal processing is power-intensive and inefficient.
Innovation Solution
Analog voice activity detection systems that convert audio signals into sub-band signals, estimate energy statistic values, and classify them using analog processing to generate a wakeup signal, reducing the need for continuous power and integrating the processing directly into microphones, thus minimizing power consumption and component costs.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If traditional digital signal processing (ADC, DSP, memory) is integrated into a microphone for voice activity detection, then voice control capabilities are achieved, but power consumption increases and form factor requirements become more stringent
Solution Approach 1:
The patent extracts only the essential voice activity detection functionality from the full digital signal processing chain. Instead of integrating ADC, DSP, and memory blocks, the invention uses a simplified analog circuit that detects voice activity without requiring continuous digitization or complex digital processing, thereby achieving voice control capability with minimal power consumption.
Solution Approach 2:
The patent changes the operating domain from digital to analog for the voice activity detection function. By using analog processing instead of digital signal processing, the system achieves the same voice detection capability with significantly reduced power consumption, as analog circuits consume less power when continuously operating for always-listening applications.
2Reliability
If traditional digital signal processing components are continuously powered for always-listening applications, then voice detection accuracy is maintained, but power consumption becomes excessive
Solution Approach 1:
The analog voice activity detection circuit operates autonomously without requiring continuous power to external digital processing components. The circuit self-regulates by detecting voice activity patterns directly in the analog domain and generates wake-up signals only when voice activity is detected, eliminating the need for continuously powered ADC and DSP blocks while maintaining detection accuracy.
3Adaptability or versatility
If ADC, DSP, and memory blocks are integrated into a microphone package, then smart microphone functionality is achieved, but the form factor and component costs increase
Solution Approach 1:
The patent extracts only the essential voice activity detection functionality from the full smart microphone signal chain. Instead of integrating ADC, DSP, and memory blocks, the invention uses a simplified analog circuit that detects voice activity without requiring continuous digitization or complex digital processing, thereby achieving smart microphone functionality with minimal component integration.
Solution Approach 2:
The patent inverts the traditional approach by using analog processing instead of digital processing for voice activity detection. This inversion simplifies the circuit architecture, reducing the number of components needed while maintaining the smart microphone's ability to detect and respond to voice input.
Data Source
AI summary
According to some embodiments, an analog processing portion may receive an audio signal from a microphone. The analog processing portion may then convert the audio signal into sub-band signals and estimate an energy statistic value, such as a Signal-to-Noise Ratio (“SNR”) value, for each sub-band signal. A classification element may classify the estimated energy statistic values with analog processing such that a wakeup signal is generated when voice activity is detected. The wakeup signal may be associated with, for example, a battery-powered, always-listening audio application.