Hearing Aid Voice Detection Using Adaptive Filtering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Hearing assistance devices amplify both desired and undesired sounds, including the user's own voice, leading to decreased sound quality and issues like the occlusion effect, and existing voice detection methods struggle to accurately distinguish the user's voice from other sounds due to low signal-to-noise ratios and environmental changes.
Innovation Solution
The use of an adaptive filter in a hearing assistance device with two microphones, one positioned on the ear and another in the ear canal, to detect the user's voice by processing signals and adjusting to changes in voice and environment, allowing for reliable voice detection and reduction of amplification.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the hearing assistance device amplifies all sounds above the audibility threshold, then the user can hear desired sounds better, but the user's own voice is also amplified causing decreased sound quality and the occlusion effect
Solution Approach 1:
The patent segments the sound processing by using separate microphones and signal paths for detecting the user's own voice versus amplifying external sounds. The first microphone detects sounds for amplification while the second microphone specifically detects the user's voice, allowing independent processing of these two sound types to resolve the contradiction between helpful amplification and harmful own-voice amplification.
Solution Approach 2:
The patent introduces an intermediary signal processing system that includes a voice detector and adaptive filter. This intermediary system processes the signal from the second microphone to generate a separate control signal that modulates the amplification of the user's own voice, thereby mediating between the need for sound amplification and the need to reduce own-voice amplification.
2Measurement precision
If a bone conductive microphone is added to detect the user's own voice, then own voice detection capability is improved, but the device complexity increases and signal-to-noise ratio remains low
Solution Approach 1:
The patent uses a second microphone to create an acoustic copy of the user's voice that travels through air from the mouth to the microphone. This acoustic copy is then processed by the voice detector and adaptive filter to generate control signals, providing a simple yet effective method for own voice detection without requiring complex bone conduction sensors.
3Measurement precision
If a directional microphone is added and oriented toward the user's mouth, then own voice detection is improved, but the effectiveness depends on directivity and presence of other sound sources
Solution Approach 1:
The patent implements a dynamic adaptive filtering system that continuously adjusts its parameters based on the detected voice characteristics and environmental conditions. The adaptive filter modifies its transfer function in real-time to maintain optimal own voice detection performance across varying acoustic environments, making the system versatile rather than dependent on fixed directional properties.
Solution Approach 2:
The patent incorporates feedback through the adaptive filter that uses the detected voice signal to continuously refine the detection process. The system monitors the output and adjusts the filter parameters accordingly, creating a closed-loop control system that improves detection accuracy while compensating for environmental variations and other sound sources.
4Measurement precision
If a fixed filter is used to distinguish the user's voice from other sounds, then initial voice detection works, but the filter cannot self-correct for changes in the user's voice and environment
Solution Approach 1:
The patent replaces the fixed filter with a dynamic adaptive filter that automatically adjusts its parameters in response to changes in the user's voice characteristics and environmental conditions. The adaptive nature of the filter allows it to track and accommodate variations over time, maintaining accurate voice distinction without requiring manual recalibration.
Solution Approach 2:
The adaptive filter performs self-correction by automatically adjusting its own parameters based on the incoming signal characteristics. The system monitors the detected voice and environmental conditions, then autonomously modifies the filter settings to maintain optimal performance, eliminating the need for external intervention or fixed parameters.
Data Source
AI summary
A hearing assistance system includes a pair of left and right hearing assistance devices to be worn by a wearer and uses both of the left and right hearing assistance devices to detect the voice of the wearer. The left and right hearing assistance devices each include first and second microphones at different locations. Various embodiments detect the voice of the wearer using signals produced by the first and second microphones of the left hearing assistance device and the first and second microphones of the right hearing assistance device. Various embodiments use outcome of detection of the voice of the wearer performed by the left hearing assistance device and the outcome of detection of the voice of the wearer performed the right hearing assistance device to determine whether to declare a detection of the voice of the wearer.