Hearing Aid Speech Detection Using Band-Split Filters
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing hearing aids struggle to effectively enhance speech in environments with varying noise levels, such as cafeterias or parties, due to sensitivity to modulated noise and slow adaptation speeds in noise level changes.
Innovation Solution
A hearing aid with separate detection mechanisms for voiced and unvoiced speech signals, using band-split filters and a selective gain controller to increase gain in frequency bands where voiced speech dominates over noise, while avoiding artifacts by employing a frequency-band-dependent level difference table.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the percentile difference method is used to detect speech, then speech detection works in steady noise or quiet surroundings, but the system becomes sensitive to modulated noise and performs inadequately in varying noise environments
Solution Approach 1:
The patent divides the frequency spectrum into multiple bands and processes each band independently. The band-split filter separates the input signal into frequency bands, and the speech detector analyzes each band separately to detect voiced and unvoiced speech signals. This segmentation allows the system to distinguish speech from modulated noise more effectively by examining spectral characteristics across different frequency regions.
Solution Approach 2:
The patent applies different detection criteria and gain adjustments to different frequency bands based on local speech characteristics. The selective gain controller increases gain by a predetermined amount in specific frequency bands where voiced speech is detected, while leaving other bands unchanged. This localized processing optimizes speech enhancement in specific spectral regions without adversely affecting other areas.
2Measurement precision
If the speech intelligibility index (SII) optimization method is used, then speech enhancement precision is very high, but the adaptation speed is poor in noisy environments
Solution Approach 1:
Instead of continuously optimizing the complex SII value, the patent applies a simplified approach by detecting speech presence and applying predetermined gain increases only in specific frequency bands where speech is detected. This partial action on key frequency bands achieves sufficient speech enhancement without the computational burden of full SII optimization, thereby increasing adaptation speed.
Solution Approach 2:
The patent changes the gain parameter in a predetermined manner based on detected speech characteristics. When voiced speech is detected in a frequency band, the selective gain controller increases the gain by a predetermined amount. This discrete parameter adjustment is computationally simpler and faster than continuous SII optimization, enabling rapid adaptation to changing noise conditions while maintaining enhancement precision.
3Reliability
If gain is increased in frequency bands where speech is detected, then speech intelligibility is improved, but artifacts may be introduced that reduce intelligibility
Solution Approach 1:
The patent segments the frequency spectrum and applies gain adjustments selectively to specific bands. The band-split filter divides the input signal, and the selective gain controller applies predetermined gain increases only in frequency bands where voiced speech is detected. This segmentation prevents artifacts by avoiding unnecessary gain in bands where speech is not present, while still enhancing intelligibility in relevant bands.
Solution Approach 2:
The patent applies different gain characteristics to different frequency bands based on local speech detection. The selective gain controller increases gain by a predetermined amount in specific bands where speech is detected, while maintaining original gain in other bands. This localized approach enhances speech intelligibility in relevant frequency regions without introducing artifacts through excessive or inappropriate gain in other regions.
Data Source
AI summary
A hearing aid (60A) configured to be worn by a hearing-impaired user has a speech detector (10A) and a speech enhancer (40A) for enhancing speech being present in an input signal of the hearing aid (60A). The speech detector (10A) has means (11, 12) for independently detecting the presence of voiced and unvoiced speech in order to allow for the speech enhancer (40A) to increase the gain of speech signals suitably fast to incorporate the speech signals themselves. The hearing aid (60A) has means (49A, 50A) for communicating information regarding the detected speech signals wirelessly to a similar hearing aid (60B) worn contralaterally by the user for the purpose of mutually enhancing speech signals in the two hearing aids (60A, 60B) when speech is detected to be originating from the front of the user, and means (52B) for suppressing speech enhancement in the contralateral hearing aid (60B) when speech is detected to be originating from the ipse-lateral side of the user. The invention further provides a method of enhancing speech in a hearing aid.