The price of high-quality prescription-based hearing aids is thousands of dollars due to complex hardware and lack of economy of scale. We propose bone conduction headphones utilizing low latency and stereo microphones, connected to a regular smartphone, to effectively achieve similar results to that of a standard high quality hearing aid for 1% of the cost. We set up an audio testing lab using off-the-shelf components to test the performance of existing and proposed solutions. Existing low-cost software- based hearing aid applications on smartphones were found to introduce an unacceptable latency of  ~250 ms. The audio testing configuration this lab uses has a total latency of 7.3 ms when the proposed system is modified to use wired connection to the smartphone. However, it is not achievable over current implementation of Bluetooth Low Energy (BLE) signals. The stereo microphone system, with speech and noise played from opposite sides of the dummy head, demonstrates the need for spatial noise suppression. We also tested the effectiveness of occlusion effect and hypothetical ML based noise reduction techniques. Thus, a stereo bone conduction headset connected to a smartphone is a viable solution that mitigates the occlusion effect, yet also allows for spatial noise suppression and possible AI/ML usage in noisy environments. However, more research needs to be done on the latency of BLE headphones.

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

Affordable Bone Conduction Headphone-Based Hearing Aid as an Alternative to Standard Hearing Aid

  • Isabel Ji

摘要

The price of high-quality prescription-based hearing aids is thousands of dollars due to complex hardware and lack of economy of scale. We propose bone conduction headphones utilizing low latency and stereo microphones, connected to a regular smartphone, to effectively achieve similar results to that of a standard high quality hearing aid for 1% of the cost. We set up an audio testing lab using off-the-shelf components to test the performance of existing and proposed solutions. Existing low-cost software- based hearing aid applications on smartphones were found to introduce an unacceptable latency of  ~250 ms. The audio testing configuration this lab uses has a total latency of 7.3 ms when the proposed system is modified to use wired connection to the smartphone. However, it is not achievable over current implementation of Bluetooth Low Energy (BLE) signals. The stereo microphone system, with speech and noise played from opposite sides of the dummy head, demonstrates the need for spatial noise suppression. We also tested the effectiveness of occlusion effect and hypothetical ML based noise reduction techniques. Thus, a stereo bone conduction headset connected to a smartphone is a viable solution that mitigates the occlusion effect, yet also allows for spatial noise suppression and possible AI/ML usage in noisy environments. However, more research needs to be done on the latency of BLE headphones.