Prediction Residual Limit Control for Adaptive Pulse Noise Suppression
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional noise suppression methods fail to adaptively detect and remove pulse noise based on changes in signal levels, leading to inefficient noise suppression in situations where the input signal level changes.
Innovation Solution
A noise suppression apparatus that includes an analyzer for deriving linear prediction coefficients, a residual calculator for calculating prediction residual signals, a threshold value calculator for determining dynamic threshold values based on input signal levels, a judger for comparing signal levels, and a limit controller for performing limit control on the prediction residual signals, ensuring effective noise suppression only when pulse noise is present.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If conventional fixed-threshold noise detection is used, then the detection method is simple, but it cannot adaptively detect pulse noise when signal level changes
Solution Approach 1:
The patent applies dynamics by making the threshold value adaptive rather than fixed. The threshold calculation unit dynamically adjusts the threshold based on the signal level of the input signal, allowing the noise detection system to adapt to varying signal conditions. This resolves the contradiction by introducing dynamic adjustment mechanisms that maintain simplicity while achieving adaptability.
Solution Approach 2:
The patent changes the parameter of threshold value from a constant to a variable that depends on signal level. By calculating the threshold as a function of the input signal's statistical properties (mean and standard deviation), the system achieves adaptability to different signal levels without requiring complex external control mechanisms.
2Measurement precision
If adaptive threshold calculation based on signal level is implemented, then noise detection accuracy improves, but processing complexity increases
Solution Approach 1:
The system performs self-service by using its own input signal to generate the threshold criterion. The threshold calculation unit computes the mean and standard deviation from the input signal itself, eliminating the need for external reference signals or manual threshold setting. This self-adaptive approach improves detection accuracy while keeping the system relatively simple.
Solution Approach 2:
The patent implements feedback by using the calculated signal statistics (mean and standard deviation) to continuously adjust the threshold value. This closed-loop approach ensures that the threshold always reflects the current signal conditions, improving detection accuracy without requiring complex open-loop control systems.
3Object-affected harmful factors
If limit control is applied to prediction residual signal, then pulse noise suppression effectiveness improves, but adverse effects on voice/sound output increase
Solution Approach 1:
The patent applies local quality by selectively applying limit control only to portions of the prediction residual signal that exceed the adaptive threshold. Rather than uniformly suppressing all residual signals, the system identifies and suppresses only those components that are likely to be pulse noise, thereby maintaining voice/sound quality while achieving effective noise suppression.
Solution Approach 2:
The system dynamically adjusts the suppression threshold based on signal level parameters (mean and standard deviation). By changing the threshold parameter adaptively, the system can effectively suppress pulse noise while minimizing adverse effects on legitimate signal components, as the threshold automatically adjusts to preserve normal signal characteristics.
Data Source
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.