Open Source System for Audio Processing

点击翻译按钮获取中文摘要

This proposal responds to RFA–DC-16-002, in which the NIDCD is seeking to support the development of acoustic signal processing tools that have the potential to accelerate research studies and to facilitate the translation of novel algorithms for hearing aids. Approximately 15% of American adults report some degree of hearing trouble, and 12.7% of Americans aged 12 and over experience bilateral hearing loss (HL). The 2008 hearing aid industry survey MarkeTrack VIII reports that less than 1 in 10 people with mild hearing loss uses amplification, while only 4 in 10 people with moderate-to-severe hearing loss use amplification, leaving still a large swath of the hearing impaired population without rehabilitation. While the reasons for poor hearing aid adoption are diverse and complex, out-of-pocket costs rank high as an issue to be addressed. And in fact, the Senate Report Language for FY2013 appropriations “strongly urges NIDCD to support research grants that could lead to less expensive hearing aids, so such aids could become accessible and affordable to more people." As the lead federal agency promoting the nation’s hearing healthcare, NIDCD is actively seeking to address accessibility from the public health perspective. Aside from cost, one of the primary barriers to hearing aid adoption remains sound quality, and the sense that hearing aids are still not quite able to meet all the needs of hearing impaired consumers. Creare and its collaborators at Boys Town National Research Hospital (BTNRH) plan to develop and disseminate an open-source audio processing platform to spur innovation in Hearing Aid (HA) research. The result of this effort will be a user-friendly, portable, upgradable and wearable “master hearing aid” (MHA) that will allow collaborative development and the open exchange of new processing algorithms within the broader hearing research community, including both academia and industry. The hardware used in the MHA has an open, published design architecture, using components currently available on the electronics market. We envision three levels of users: (1) “expert” users implement and test new algorithms directly in firmware through an interface that gives them access to the basic features of the hardware (audio input and output, power management, housekeeping, etc.); (2) “developers” users interact at the software application programming interface (API) level to modify parameters of the algorithms already implemented in firmware and evaluate their performance in a variety of conditions; and (3) “professional” users have access to “pre-programmed” algorithms with some degree of parameter control through a simple user interface, to test the relative benefits of amplification variants available with the wearable master hearing aid.