The latest version of the open Master Hearing Aid (openMHA) software, version 4.18.0, is now available. With significant new algorithms and practical examples, openMHA 4.18.0 once again sets new standards for the research and development of signal processing for hearing aids. These advances will enable universities and industry to develop and test new and innovative hearing aid technologies that will help provide the best possible hearing solutions for the hearing impaired in the long term. “With version 4.18.0 of openMHA, we offer researchers and developers even more powerful tools for the development and optimization of modern hearing aid technologies,” says Dr. Kirsten Wagener, Head of Research at Hörzentrum Oldenburg gGmbH. “The new examples and methods demonstrate the enormous potential of openMHA in hearing research and signal processing. openMHA is used worldwide. The continuous development of this software ensures that researchers and developers can work purposefully on the best hearing solutions of the future without being restricted by commercial interests or proprietary software.”

Important new features in openMHA 4.18.0:

Speech enhancement with the help of deep neural networks

Three new signal processing modules, so-called “plugins”, offer different approaches for real-time, low-latency speech enhancement using Deep Neural Networks (DNN) trained in PyTorch. These methods work with an algorithmic delay of only 5.4 ms and use four microphone input signals. The introduction of these DNN-based algorithms represents a significant advance in hearing aid technology that continues to be intensively researched - also with the support of openMHA.

Vocoder for simulating a cochlear implant

A new set of plugins implement CI vocoders that enable the simulation and auralization of the output of a cochlear implant. Two common signal processing methods have been implemented: the ACE strategy (“Advanced Combination Encoder”, “n-of-m”) and the CIS strategy (“Continuous Interleaved Sampling”). These technologies improve the understanding of how cochlear implants work and support the development of better hearing solutions.

Hearing loss simulator

Various plug-ins and their combination in a new configuration to simulate hearing loss includes level-dependent attenuation in 32 frequency bands and spectral smearing to simulate reduced frequency selectivity. These functions enable a more realistic simulation of different hearing losses and their effects and thus a more precise fitting and optimization of hearing aids in simulations that can also meet the individual needs of users.

For more information, please visit the project website at www.openmha.org or contact us at info@hz-ol.de.

About openMHA:

The project “Open community platform for hearing aid algorithm research”, funded by the National Institutes of Health and the German Research Foundation (DFG) as part of the Excellence Strategy (Cluster of Excellence Hearing4all), aims to conduct sustainable, focused research into the improvement and new development of hearing aids. The openMHA is an established open source software platform for signal processing in hearing aids. The platform is constantly being further developed in a collaboration between the Hörzentrum Oldenburg and the Carl von Ossietzky University of Oldenburg and is used worldwide in research and development to develop and test innovative hearing solutions.