Daily Archives: December 18, 2018

Researching Hearing Aids with the Raspberry Pi

All around the world, millions of people benefit from wearing hearing aids. Apart from helping them to hear in a better way, hearing aids lower people’s risk of developing dementia, the likelihood for loneliness, and the possibility of their withdrawing from society.

Testing hearing aids outside the laboratory can be a tough task, but researchers have found the highly popular single board computer, the Raspberry Pi (RBPi) a sound investment for testing hearing aid algorithms. Therefore, for hearing aid research, they are using the RBPi boards.

Although researchers have spent a lot of time and energy for developing hearing aids over the years, there is yet room for improvement. According to a signal processing engineer Tobias Herzke at HorTech in Oldenburg, in Germany, this is especially true for situations that are difficult acoustically. However, the RBPi is proving to be a next-generation research tool for the scientists.

To compensate for an individual’s hearing loss, it is necessary to tailor the amplification and compression in the hearing aid. Researchers plug a monitor to the RBPi and fire up the Fitting GUI for the tailoring.

For this, a spin-off company of the University of Oldenburg has developed openMHA. They have designed openMHA as a common, portable software platform, useful for teaching and researching hearing aid. According to Hendrik Kayser, with the openMHA platform, researchers can process signals in real-time with low delays. Hendrik develops algorithms for processing signals for digital hearing devices.

The software platform openMHA offers a set of standard algorithms that form a complete hearing aid. It can process the signal a live microphone generates to perform different activities such as directional filtering, amplification, feedback suppression, and noise reduction. The RBPi helps in testing new algorithms as this can be difficult with hearing aids alone. The RBPi and openMHA help hearing aid researchers with processing audio signals instantly and adapting to the hearing loss of the individual. The main advantage is the delays between incoming and outgoing audio is below 10 ms. The hearing aid actually has no GUI, except when fitting the amplifier parts.

In the laboratory environment, researchers can execute the openMHA software on Linux computers. According to Tobias, the sound environment will be different within a laboratory from that in an environment that a hearing aid user is likely to encounter in real life. This has often led to wrong results in the past, and did not offer a true reflection of the use of hearing aids. In such situations, the ARM-based single board computer, the RBPi offers a wonderful solution.

By taking advantage of the portable nature of the RBPi, and running openMHA on it, the researchers were able to evaluate newer algorithms in realistic outdoor conditions in real time. In fact, researchers were able to implement a new algorithm running on a mobile device for finding out how the user hears in real time while he is running around wearing a hearing aid.

Using an RBPi means one does not have to carry around a Linux laptop and it is far less expensive. The RBPi offers decent computing capabilities in a small space, while consuming low power.