Noise-Synchronized Voice Registration for Lombard Effect Robustness

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

The accuracy of speaker verification systems decreases due to the Lombard effect, where the shape of the vocal organ changes in noisy environments, causing significant variation in voice features during the verification phase compared to the registered voice.

Innovation Solution

A voice registration device that reproduces noise during voice input, synchronizes noise reproduction with voice input, and registers voice data or derived data for verification, optionally using echo cancellation to improve signal-to-noise ratio.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If voice registration is performed under silent environment, then voice features are clear and stable, but verification accuracy decreases when actual usage occurs in noisy environments due to Lombard effect

Engineering Contradiction:

Improvevoice feature stabilityVSAvoidspeaker verification accuracy

Core Design Contradiction:

Measurement precisionVSReliability

Solution Approach 1:

The system performs preliminary voice registration under noisy conditions by reproducing noise during the registration phase. This prepares the speaker verification system in advance to handle noisy environments by training the verification algorithm with voice samples that include Lombard effect characteristics, thereby improving verification accuracy when actual usage occurs in similar noisy conditions

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The system changes the acoustic environment parameter during registration by introducing reproduced noise. This parameter change allows the registration phase to capture voice features under conditions that simulate actual usage environments, making the registered voice data more representative and improving verification accuracy across varying noise conditions

Inventive Principle:

Principle #35Parameter changes

2Adaptability or versatility

If noise is reproduced during voice input, then voice features under noisy conditions are captured, but signal-to-noise ratio of registered voice data decreases

Engineering Contradiction:

Improveenvironmental adaptabilityVSAvoidsignal-to-noise ratio

Core Design Contradiction:

Adaptability or versatilityVSLoss of information

Solution Approach 1:

The system uses an intermediary noise reproduction mechanism that introduces controlled noise during registration. This intermediary approach allows the system to capture Lombard effect characteristics while maintaining the ability to separate and process the actual voice signal from the reproduced noise, preventing complete loss of voice information

Inventive Principle:

Principle #24Intermediary (Mediator)

Solution Approach 2:

The system applies partial noise reproduction rather than complete noise immersion. By reproducing noise at controlled levels during registration, the system captures sufficient Lombard effect characteristics to improve environmental adaptability while avoiding excessive noise that would completely overwhelm the voice signal and cause irreversible information loss

Inventive Principle:

Principle #16Partial or excessive action

Data Source

PatentUS12462809B2Voice registration device, control method, program, and storage medium

Publication Date: 2025.11.04 NEC CORP

AI summary

The voice registration device 1X mainly includes a noise reproduction means 220X, a voice data acquisition means 200X, and a voice registration means 210X. The noise reproduction means 220X is configured to reproduce noise data during a time period in which voice input from a user is performed. The voice data acquisition means 200X is configured to acquire the voice data based on the voice input. The voice registration means 210X is configured to register the voice data or data generated based on the voice data as data to be used for verification relating to a voice of the user.