Adaptive Pulse Noise Suppression Using Linear Prediction Residuals

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Conventional noise suppression techniques fail to adaptively detect and remove pulse noise based on changes in signal levels, leading to inefficient noise suppression, especially in dynamic environments like car radio receivers and video conference systems.

Innovation Solution

A noise suppression apparatus that uses linear prediction analysis to derive coefficients, calculates a prediction residual signal, and performs limit control based on a dynamically determined threshold value, effectively suppressing pulse noise only when its level exceeds the threshold, thereby reducing adverse effects on voice/sound output.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Device complexity

If a fixed threshold value is used for noise detection, then the detection method is simple, but the noise detection accuracy deteriorates when signal levels change

Engineering Contradiction:
Improvedetection method complexityVSAvoidnoise detection accuracy
Core Design Contradiction:
Device complexityVSMeasurement precision

Solution Approach 1:

The patent applies dynamics by making the threshold value adaptive rather than fixed. The threshold is dynamically adjusted based on the signal level of the input signal, allowing the noise detection system to adapt to changing signal conditions. This resolves the contradiction by sacrificing some system complexity to achieve accurate noise detection across varying signal levels.

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The patent changes the parameter of the threshold value from a constant to a variable that depends on the signal level. By making the threshold parameter change according to the input signal characteristics, the system achieves accurate noise detection without requiring overly complex fixed-threshold methods.

Inventive Principle:
Principle #35Parameter changes

2Productivity

If noise suppression is performed without adaptive threshold adjustment, then the processing is fast and simple, but the noise suppression effectiveness deteriorates in dynamic signal environments

Engineering Contradiction:
Improvenoise suppression speedVSAvoidnoise suppression effectiveness
Core Design Contradiction:
ProductivityVSReliability

Solution Approach 1:

The system dynamically adjusts the threshold based on signal level while maintaining real-time processing capability. This allows the noise suppression to remain both fast and effective by adapting to changing conditions without requiring complex reprocessing.

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The system uses feedback from the signal level to adjust the threshold value, creating a closed-loop control that ensures effective noise suppression. The threshold is continuously adjusted based on the current signal characteristics, maintaining reliability while preserving processing speed.

Inventive Principle:
Principle #23Feedback

3Measurement precision

If a dynamic threshold value is calculated based on signal level, then the noise detection accuracy improves, but the device complexity increases

Engineering Contradiction:
Improvenoise detection accuracyVSAvoidthreshold calculation complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent implements a relatively simple parameter change approach where the threshold is adjusted based on signal level using a straightforward calculation method. This balances the need for accurate noise detection with the desire to avoid excessive system complexity.

Inventive Principle:
Principle #35Parameter changes

Data Source

PatentEP3404657B1Noise suppression apparatus, noise suppression method, and computer program
Publication Date: 2019.09.25 PANASONIC INTELLECTUAL PROPERTY CORP OF AMERICA
  • EP3404657B1 patent drawingFigure 1
  • EP3404657B1 patent drawingFigure 2
  • EP3404657B1 patent drawingFigure 3

AI summary

A noise suppression apparatus disclosed is capable of suppressing pulse noise in an input signal even in a situation in which the level of the input signal changes. The pulse noise mixed in the input signal is suppressed, and linear prediction coefficients for the input signal is derived by linear prediction analysis. A prediction residual signal is then calculated from the input signal using the linear prediction coefficient. A threshold value is calculated based on the signal level of the input signal and the signal level of the prediction residual signal is compared with the threshold value. A limit control is performed on the prediction residual signal depending on a result of the comparison, and an output signal is generated based on the prediction residual signal having been subjected to the limit control using the linear prediction coefficient.