Audio Signal Processing Device for Babble Noise Detection

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing audio signal processing systems struggle to accurately distinguish babble noise from other noise sources, leading to ineffective noise suppression and degraded sound quality, as they often misjudge sounds with high frequency power fluctuations as babble noise.

Innovation Solution

An audio signal processing system that converts audio signals into the frequency domain, calculates spectral changes between frames, and uses these changes to accurately judge the presence of babble noise, employing a time-frequency conversion unit, spectral change calculation unit, and judgment unit to differentiate between babble noise and other noise types.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If existing noise detection methods are used to identify babble noise, then noise suppression can be performed, but sounds from other sources (e.g., moving vehicles) are mistakenly judged as babble noise, reducing measurement precision

Engineering Contradiction:
Improvenoise suppression effectivenessVSAvoidbabble noise detection accuracy
Core Design Contradiction:
ReliabilityVSMeasurement precision

Solution Approach 1:

The patent applies dynamics by transitioning from static noise detection (based on fixed frequency power thresholds) to dynamic detection (based on temporal spectral change rates). The system continuously monitors how quickly the frequency spectrum changes over time, allowing it to adaptively distinguish babble noise from other sounds based on their different temporal characteristics rather than relying on fixed thresholds that cause misclassification.

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The patent changes the detection parameter from static frequency power levels to dynamic spectral change rates. By calculating the rate of change of frequency spectrum over time (spectral change rate), the system transforms the detection approach into a time-derivative domain, which fundamentally alters how noise characteristics are measured and enables more accurate discrimination between babble noise and other sound sources.

Inventive Principle:
Principle #35Parameter changes

2Loss of energy

If noise suppressor aggressively suppresses detected noise, then noise power is reduced, but legitimate audio signals may be suppressed along with noise, degrading sound quality

Engineering Contradiction:
Improvenoise powerVSAvoidsound quality
Core Design Contradiction:
Loss of energyVSReliability

Solution Approach 1:

The patent applies local quality by implementing frequency-selective noise suppression rather than uniform suppression across all frequencies. The noise suppressor calculates suppression gains for different frequency bands based on the detected spectral change characteristics, allowing aggressive suppression in frequency regions where babble noise is present while preserving frequencies containing legitimate speech or audio signals. This localized approach ensures noise reduction without degrading overall sound quality.

Inventive Principle:
Principle #3Local quality

Data Source

PatentEP2444966B1Audio signal processing device and audio signal processing method
Publication Date: 2019.07.10 FUJITSU LTD
  • EP2444966B1 patent drawingFigure 1
  • EP2444966B1 patent drawingFigure 2A~2B
  • EP2444966B1 patent drawingFigure 3

AI summary

An audio signal processing system including a time-frequency conversion unit which converts an audio signal in time domain into frequency domain in frame units so as to calculate a frequency spectrum of the audio signal, a spectral change calculation unit which calculates an amount of change between a frequency spectrum of a first frame and a frequency spectrum of a second frame before the first frame based on the frequency spectrum of the first frame and the frequency spectrum of the second frame, and a judgment unit which judges the type of the noise which is included in the audio signal of the first frame in accordance with the amount of spectral change.