Microphone Array Noise Suppression via Phase Difference Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing noise suppression methods using phase differences in microphone arrays often result in false detection of vocalization intervals due to residual noise, leading to deteriorated sound quality and reduced speech recognition accuracy, especially when time-smoothing is applied.
Innovation Solution
A noise suppressing apparatus that calculates phase differences between sound signals from a microphone array, determines sound arrival rates from distinct phase difference areas, and applies a suppression coefficient based on the dissimilarity between these rates to accurately identify vocalization intervals and reduce noise.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If time-smoothing is applied to reduce false detection of vocalization intervals, then false detection due to residual noise is reduced, but vocalization intervals may be interrupted and speech recognition accuracy deteriorates
Solution Approach 1:
The patent divides the sound signal analysis into multiple frequency bands and calculates sound arrival rates for each band separately. By segmenting the frequency spectrum and analyzing each band independently, the system can detect vocalization intervals more accurately without being misled by residual noise in individual bands, thus resolving the contradiction between reducing false detection and maintaining detection accuracy.
Solution Approach 2:
The patent applies sound arrival rate calculation across multiple frequency bands rather than relying on a single band or excessive smoothing. This partial action approach - using multiple bands with moderate smoothing - achieves better balance between false detection reduction and vocalization interval accuracy compared to aggressive time-smoothing of the entire signal.
2Ease of operation
If a single sound arrival rate threshold is used to determine vocalization intervals, then the determination process is simple, but false detection occurs due to residual noise
Solution Approach 1:
The patent segments the frequency spectrum into multiple bands and calculates sound arrival rates for each band. Instead of using a single threshold for the entire spectrum, the system evaluates arrival rates across multiple segmented bands, which reduces false detection from residual noise while maintaining operational simplicity through standardized processing of each band.
Solution Approach 2:
The patent combines the results from multiple frequency bands to make the final vocalization interval determination. By merging the information from multiple bands, the system achieves more reliable detection accuracy while keeping the overall process relatively simple through systematic combination of band results.
3Object-affected harmful factors
If noise suppression is applied to improve sound quality, then residual noise is reduced, but vocalization intervals may be falsely detected due to over-suppression
Solution Approach 1:
The patent applies different noise suppression levels to different frequency bands based on their local characteristics. By adjusting the suppression strength locally in each frequency band rather than uniformly across the entire spectrum, the system reduces residual noise effectively while avoiding over-suppression that could cause false vocalization detection.
Solution Approach 2:
The patent dynamically adjusts noise suppression parameters based on the detected sound arrival rates in each frequency band. This dynamic approach allows the system to suppress noise adaptively - applying stronger suppression where residual noise is problematic while maintaining weaker suppression in bands where vocalization signals are present, thus preventing false detection.
Data Source
AI summary
A noise suppressing apparatus calculates a phase difference on the basis of a first and second sound signal obtained by a microphone array; calculates a first sound arrival rate on the basis of a first phase difference area and the phase difference and a second sound arrival rate on the basis of a second phase difference area and the phase difference; calculates a dissimilarity that represents a level of difference between the first sound arrival rate and the second sound arrival rate; determines whether the pickup target sound is included in the first sound signal on the basis of the dissimilarity; and determines a suppression coefficient to be applied to the frequency spectrum of the first sound signal, on the basis of a result of the determination of whether the pickup target sound is included and on the basis of the phase difference.