Network Microphone Voice Detection Using Sound Metadata Feedback

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing voice-controllable media playback systems face issues with impaired microphone performance due to hardware problems or environmental factors, leading to poor voice command detection and downstream processing, which affects the effectiveness of the voice detection and downstream processing, which can lead to ineffective voice detection and ineffective voice detection and downstream processing, which can result in poor voice detection and ineffective voice detection and downstream processing.

Innovation Solution

The system employs network microphone devices (NMDs) that analyze sound metadata, such as spectral data and signal levels, to identify performance issues without revealing the original audio content, and adjust parameters like gain, spatial processing, and noise reduction to improve voice detection.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If voice detection is performed in out-loud settings with multiple networked playback devices, then media accessibility and listening experience are improved, but microphone performance is impaired due to hardware problems or environmental factors

Engineering Contradiction:
Improvemedia accessibilityVSAvoidmicrophone performance
Core Design Contradiction:
Adaptability or versatilityVSReliability

Solution Approach 1:

The system continuously monitors microphone performance using sound metadata (spectral data, signal levels) and provides feedback to adjust processing parameters. When degradation is detected, the system automatically modifies gain, spatial processing, or noise reduction settings to compensate for hardware issues or environmental factors, maintaining reliable voice detection in out-loud settings.

Inventive Principle:
Principle #23Feedback

Solution Approach 2:

The patent dynamically changes audio processing parameters such as gain, spatial processing weights, and noise reduction thresholds based on monitored sound metadata. By adjusting these parameters in response to detected performance degradation, the system maintains effective voice detection despite hardware problems or environmental noise in multi-device playback systems.

Inventive Principle:
Principle #35Parameter changes

2Measurement precision

If sound metadata is analyzed to identify performance issues, then voice detection accuracy is improved, but system complexity increases

Engineering Contradiction:
Improvevoice detection accuracyVSAvoidsystem complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The system extracts only the necessary sound metadata features (spectral data, signal levels) from the audio signal for performance monitoring, rather than processing the complete audio stream. This selective extraction approach improves voice detection accuracy through targeted analysis while minimizing the computational complexity and processing overhead required for continuous monitoring and adjustment.

Inventive Principle:
Principle #2Taking out (Extraction)

Data Source

PatentEP3853848B1Voice detection optimization using sound metadata
Publication Date: 2026.01.07 SONOS INC
  • EP3853848B1 patent drawingFigure 1A
  • EP3853848B1 patent drawingFigure 1B
  • EP3853848B1 patent drawingFigure 2A~2B

AI summary

Systems and methods for optimizing voice detection via a network microphone device, NMD, are disclosed. In one example, individual microphones of a network microphone device detect sound. The sound data is captured in a first buffer and analyzed to detect a trigger event. Metadata associated with the sound data is captured in a second buffer and provided to at least one network device to determine at least one characteristic of the detected sound based on the metadata. The network device provides a response that includes an instruction, based on the determined characteristic, to modify at least one performance parameter of the NMD. The NMD then modifies the at least one performance parameter based on the instruction.