Perceptual Frequency Weighting for Accurate Howling Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing howling detection methods fail to accurately identify howling phenomena due to their inability to consider the varying sensitivity of human ears to different frequencies, leading to inaccurate suppression of feedback noise.
Innovation Solution
A howling detection method and apparatus that performs window separation processing on audio signals, calculates signal energy indicator values for each frequency, and uses perceptual coefficients to determine a howling indicator value, which is then compared to a threshold, incorporating a counting mechanism to accurately detect howling based on human ear sensitivity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If existing howling detection methods use simple energy detection of output signal, then the detection process is simple and fast, but the detection accuracy is low due to inability to consider human ear sensitivity to different frequencies
Solution Approach 1:
The audio signal is segmented into multiple frequency bins through Fourier transformation, allowing independent analysis of energy at different frequencies. This segmentation enables the detection method to consider human ear sensitivity by applying different weights to different frequency components, thereby improving detection accuracy without excessive complexity increase.
Solution Approach 2:
The patent applies local quality by using frequency-dependent weighting factors that reflect human ear sensitivity characteristics. Different frequency regions are assigned different weights according to the perceptual model, allowing the detection to focus on frequency regions where howling is more perceptible to human ears, thus improving accuracy while maintaining computational efficiency.
2Measurement precision
If spectral entropy method is used for howling detection, then frequency analysis is performed, but the method still lacks consideration of human ear sensitivity leading to inaccurate detection results
Solution Approach 1:
The patent changes the detection parameter from raw spectral energy to perceptual energy by introducing frequency-dependent weighting factors. This parameter transformation aligns the detection metric with human perceptual characteristics, improving detection accuracy. The computational complexity increase is manageable because the weighting factors can be pre-calculated and stored in lookup tables.
Data Source
Figure 1~2
Figure 3
Figure 4
AI summary
A howling detection method is provided. A window separation processing is processed on an audio signal to obtain a plurality of analysis windows. A signal energy indicator value of each preset frequency in at least one analysis window is obtained by using a preset perceptual coefficient corresponding to each frequency, to obtain a perceptual energy indicator value of each frequency, the preset perceptual coefficient corresponding to each frequency indicating a sensitivity of a human ear to a sound of each frequency. It is determined whether howling occurs according to the perceptual energy indicator value of each frequency in the at least one analysis window.