Inverse Filter Peak Suppression for Stable Sound Localization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing sound localization technologies face challenges in generating accurate inverse filters for out-of-head localization due to steep peaks or dips in ear canal transfer characteristics, which can lead to unintended sound quality issues and require frequent recalibration when headphones are re-worn.
Innovation Solution
A processing device and method that acquire frequency characteristics, extract extreme values, calculate kurtosis, and determine whether to suppress peaks or dips based on a threshold value, effectively smoothing the frequency response to stabilize sound localization across different headphone wear configurations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If inverse filter is generated using algorithms like least squares method, then filter generation is achieved, but steep peaks or dips occur in frequency characteristics leading to unintended sound quality issues
Solution Approach 1:
The patent applies parameter changes by modifying the frequency characteristics of the inverse filter through peak suppression processing. When steep peaks or dips are detected in the frequency characteristics, the system adjusts the filter parameters by suppressing these extreme values, thereby changing the frequency response parameters to eliminate harmful peaks while maintaining the overall filter generation accuracy.
2Reliability
If ECTF is measured each time user rewears headphones, then accurate filter generation is maintained, but user burden increases
Solution Approach 1:
The patent applies preliminary action by performing peak suppression processing on the frequency characteristics before finalizing the inverse filter. This preliminary processing of the ECTF data removes steep peaks and dips that would cause instability, allowing the system to maintain reliable filter accuracy without requiring remeasurement when users rewear headphones. The preliminary action stabilizes the data in advance, preventing the need for frequent user measurements.
3Adaptability or versatility
If inverse filter is generated at different control points positions, then out-of-head localization is achieved, but unintended peaks occur affecting sound quality
Solution Approach 1:
The patent applies local quality by implementing selective peak suppression at specific frequency regions. Instead of uniformly processing the entire frequency response, the system identifies and suppresses only the harmful steep peaks and dips at specific control points where unintended peaks occur. This localized processing maintains the localization reproduction capability at effective control points while eliminating harmful frequency characteristics at problematic regions.
Data Source
AI summary
A processing device according to this embodiment includes: a frequency characteristics acquisition unit configured to acquire frequency characteristics of an input signal; an extreme value extraction unit configured to extract an extreme value of spectral data; a kurtosis calculation unit configured to: calculate an evaluation value from spectral data; and calculate a kurtosis of a peak or a dip based on a plurality of evaluation values calculated by changing a calculation width, the evaluation value being used for evaluating the peak or the dip corresponding to the extreme value; a determination unit configured to determine whether to suppress the peak or the dip according to a comparison result between the kurtosis and a threshold value; and a suppression unit configured to suppress the peak or the dip with the extreme value that is determined to be suppressed.


