Reconfigurable ADC Input Path for Low-Power Voice Activation

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Conventional voice-activated systems consume excessive power due to high signal-to-noise ratio (SNR) requirements, leading to shortened battery life in mobile devices, as they operate at high performance levels even when monitoring for voice activity rather than performing speech processing.

Innovation Solution

A reconfigurable input path for analog-to-digital converters (ADCs) that dynamically adjusts operational performance based on audio signal quality, switching to a high-performance state only when necessary for speech processing, thereby reducing power consumption by operating in a low-power mode during voice activity detection.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If a single high-performance ADC is used to process all speech at high SNR requirements, then speech processing quality is improved, but power consumption increases

Engineering Contradiction:
Improvesignal-to-noise ratioVSAvoidpower consumption
Core Design Contradiction:
Measurement precisionVSUse of energy by moving object

Solution Approach 1:

The ADC dynamically switches between two operational states: a low-power state during voice activity detection and a high-performance state during speech processing. This dynamic reconfiguration allows the system to adapt its performance characteristics to the current operational requirements, consuming minimal power during monitoring while providing high SNR when speech processing is needed.

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The system changes the operational parameters of the ADC based on the detected voice activity. When voice activity is detected, the ADC transitions from a low-power configuration to a high-performance configuration with adjusted SNR parameters. This parameter change enables the ADC to meet high speech processing requirements only when necessary, rather than operating at peak performance continuously.

Inventive Principle:
Principle #35Parameter changes

2Measurement precision

If the ADC operates at high SNR continuously, then speech processing quality is maintained, but battery life is shortened

Engineering Contradiction:
Improvesignal-to-noise ratioVSAvoidbattery life
Core Design Contradiction:
Measurement precisionVSDuration of action of stationary object

Solution Approach 1:

The ADC operates in periodic cycles, alternating between low-power monitoring mode and high-performance processing mode. During most of the time (when no speech is present), the ADC operates in a low-power state. When speech activity is detected, it briefly switches to high-performance mode for processing, then returns to low-power mode. This periodic operation pattern maintains speech processing quality when needed while extending battery life through extended low-power operation.

Inventive Principle:
Principle #19Periodic action

3Reliability

If high-performance ADC operation is used for voice activity monitoring, then detection accuracy is improved, but power efficiency deteriorates

Engineering Contradiction:
Improvevoice activity detection accuracyVSAvoidpower efficiency
Core Design Contradiction:
ReliabilityVSLoss of energy

Solution Approach 1:

The audio processing function is segmented into two distinct operational phases: voice activity detection and speech processing. During the voice activity detection phase, the ADC operates in a low-power state with reduced SNR requirements, sufficient for detecting the presence of speech. Only when speech is detected does the system transition to the speech processing phase with high-performance ADC operation. This segmentation allows accurate voice activity detection with minimal power consumption.

Inventive Principle:
Principle #1Segmentation

Data Source

PatentEP3286837B1Analog-to-digital converter (ADC) dynamic range enhancement for voice-activated systems
Publication Date: 2022.06.08 CIRRUS LOGIC INT SEMICON LTD
  • EP3286837B1 patent drawingFigure 1~2
  • EP3286837B1 patent drawingFigure 3
  • EP3286837B1 patent drawingFigure 4

AI summary

The dynamic range and power efficiency of a voice-activated system may be improved by dynamically adjusting the configuration of the voice-activated system's input path. In one embodiment, a first portion of audio may be received through an input path of the voice- activated system having a first configuration. A characteristic of the first portion of audio may be determined and the input path may be adjusted to a second configuration based on the determined characteristic. A second portion of audio may then be received through the input path having the second configuration, and speech analysis may be performed on the second portion of audio.