Ribbit Logo

Objective

About only 20 percent of the US population who could benefit from a hearing aid can afford to wear one. The demand for hearing aids is high, and so is the price. We strive to create a smartphone (i.e., iOS) based application that possesses similar functionalities of a conventional hearing aid at a small fraction of the price. Using advanced techniques such as Fast Fourier Transform, we analyze and process the sound samples in the frequency domain, and manipulate these samples in a way that a hearing impaired individual can hear the sound. In particular, we are filtering out noises and amplifying the desirable sound needed for speech comprehension into ranges that are suitable for the audiology profile of the hearing impaired individual.

Project Background

Hearing aids process and modify sounds into the most desirable forms for hearing impaired people to receive. With hearing aids, hearing impaired population can to some extent regain the ability to hear normally and integrate back into everyday life with minimum inconvenience. However, due to multiple reasons including unaffordable prices, inconveniences, and limited quantities, only 20 percent of the people in the US who could benefit from a hearing aid wear one. This figure is likely to be much less in less developed countries. As a result, a better and more widespread alternative is needed.

Recently, there have been great advances made in handheld devices such as smartphones. These advances bring new features such as powerful computation capability without sacrificing mobility, and wireless connection. We have developed a smartphone based intelligent sound processing and fusion system. Our application is be able to collect all the sounds in the vicinity and process them on the spot according to the customized needs and prescriptions of a specific user. More specifically, the application is be able to selectively amplify and shift sounds to be within the frequency range that is required by the user, as well as filter out extraneous frequencies not required for speech.