Hearing Aid Microphone Array Voice Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current hearing assistance systems face challenges in reliably detecting the presence of a voice source close to the microphone, particularly in noisy environments, as they often rely solely on voice energy estimation without considering the direction of arrival, leading to potential misidentification of noise as speech.
Innovation Solution
A method and system that utilize a directional microphone arrangement with a beam-former algorithm, combining voice energy estimation and direction of arrival analysis to determine the presence of a voice source, allowing for enhanced reliability in voice detection by adjusting gain settings based on both total energy in the voice spectrum and the estimated direction of arrival.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If voice energy estimation alone is used to detect close voice, then the detection process is simple, but the reliability of voice detection deteriorates in noisy environments
Solution Approach 1:
The patent combines voice energy estimation with direction of arrival analysis by processing audio signals through both a beam-former algorithm and voice activity detection. The beam-former calculates direction of arrival using phase differences between multiple microphones, while the voice activity detector analyzes energy characteristics. These two independent analysis paths are merged to make a final voice presence decision, improving reliability without excessive complexity increase.
Solution Approach 2:
The patent adds the dimension of spatial information (direction of arrival) to the traditional temporal-energy-based voice detection. By analyzing the phase differences of audio signals across multiple microphones to determine direction of arrival, the system creates a new detection dimension that complements the energy-based approach, enabling better distinction between speech and noise in noisy environments.
2Device complexity
If only voice energy estimation is used, then the system is simpler, but misidentification of noise as speech increases
Solution Approach 1:
The patent segments the audio signal processing into distinct functional modules: a beam-former module that processes signals from multiple microphones to calculate direction of arrival, and a voice activity detection module that analyzes energy characteristics. This segmentation allows each module to specialize in one aspect of analysis, improving overall detection accuracy while keeping each individual module relatively simple.
Solution Approach 2:
The patent introduces direction of arrival analysis as an intermediary factor between the raw audio signal and the final voice detection decision. Instead of directly equating high energy with voice presence, the system uses direction of arrival information as an intermediate verification step, which mediates between the energy detection and the final classification, preventing misidentification of noise as speech.
3Device complexity
If gain is adjusted based on total energy only, then the control is simpler, but the signal-to-noise ratio deteriorates in varying auditory scenes
Solution Approach 1:
The patent implements dynamic gain control that adapts to changing auditory scenes by continuously monitoring both voice energy levels and direction of arrival. The gain adjustment is dynamic rather than static, responding in real-time to the presence, absence, or movement of close voices. This dynamic adaptation allows the system to maintain optimal signal-to-noise ratio across varying environmental conditions.
Solution Approach 2:
The patent employs feedback mechanisms where the voice activity detection results and direction of arrival information are fed back to the gain control unit. This feedback loop allows the system to continuously adjust the gain based on the actual acoustic environment, ensuring the signal-to-noise ratio remains optimized. The feedback from both energy analysis and spatial analysis works together to control the gain appropriately.
Data Source
AI summary
There is provided a method for providing hearing assistance to a user (101, 301), comprising: capturing audio signals by a microphone arrangement (26) comprising at least two spaced apart microphones (M1, M2); estimating the total energy contained in the voice spectrum of the audio signals captured at least one of the microphones; estimating the value of the direction of arrival of the captured audio signals by comparing the audio signals captured by at least two of the spaced apart microphones; judging whether a voice is present close to microphone arrangement by taking into account the estimated total energy contained in the voice spectrum of the captured audio signals and the estimated value of the direction of arrival of the captured audio signals; outputting a signal representative of said judgement; processing said captured audio signals according to said signal representative of said judgement; and stimulating the user's hearing, by stimulating means worn at or in at least one of the user's ears (39), according to the processed audio signals.