– Christophe Waelchli, Product Manager, ON Semiconductor
The evolutionary path of the digital hearing aid provides a good guide to how the consumer electronics industry has developed over the last several years. In many ways, the hearing aid represents the ‘perfect’ embedded system. It has always been subject to the overriding requirements of the wider market; small size, ultra-low-power consumption, high performance and constant improvement through new features and better audio quality
Semiconductor technology enables all end markets through its ability to increase integration on a predictable cycle, as observed in Moore’s law. A higher level of integration leads to increased feature density, which supports smaller end products, in size. There is no better example of this than with a device intended to fit inside an ear or positioned unobtrusively behind it.
The first major improvement in hearing aid functionality came with non-linear amplification, which allowed different sounds to be amplified more objectively based on the signal received. The advent of the digital signal processor (DSP) gave hearing aids a new dimension through increased precision, and the granularity to apply selective amplification to different frequencies.
The introduction of wireless personal area networking, typified by Bluetooth® Low Energy connectivity, ushered in the third generation of hearing aid technology. Connected hearing aids can be controlled by phone apps, with some having the ability to take phone calls and even listen to music. On top of this, the industry is now embracing Artificial Intelligence (AI) and Machine Learning to extend functionality even further. This could be considered the fourth age of hearing aids and one that is going to have perhaps the biggest impact of all.
Additionally, the human ear is a particularly good location to place all kinds of sensors. Today’s hearing aids include sensors that will be useful for medical purposes (e.g., heart rate monitoring, blood oxygenation, glucose metering, and fall detection for elderly people). These sensors can also be used to enable consumer electronics features, including step-counting and heart rate monitoring.
How smartphones and Bluetooth connectivity redefined the market
While non-linear amplification and DSP technology had a direct impact on the performance and user experience of hearing aids, the significance of wireless connectivity is perhaps more far-reaching. Adding such connectivity to any device brings it into the IoT era; it is no longer an isolated application, but something that is connected, part of an ecosystem now, a much bigger entity.
In the case of hearing aids, adding Bluetooth wireless technology allowed hearing aids to connect directly to smart phones. This may seem like a simple extension, but is in fact a huge step forward. Prior to wireless connectivity in hearing aids, the wearer had to reach behind their ear to access tiny control buttons on the device. This was particularly challenging for senior citizens – the traditional wearers of hearing aids – who typically also have reduced dexterity.
Indeed, wireless connectivity provides a way for users to control and modify the aid’s settings, but it offers much more than a new way of turning the volume up or down. Once connected to a smart phone application, the hearing aid becomes an extremely versatile peripheral, one that provides a direct and very intimate connection to the user. Hearing aids are equipped with two very important features, a microphone (or several) and a speaker. With the addition of Bluetooth connectivity, this allows them to communicate concurrently with both the user and the wider Internet. It effectively gives wearers direct access to anything the internet has to offer. Edge AI algorithms that allow the recognition of words like “OK Google” or “Alexa™”, offer the user a similar experience as current smart speakers.
Alongside Edge AI, “comfort algorithms” have had a considerable impact on improving the listening experience. Examples include environmental classification, which has allowed hearing aids to adapt to the prevailing environmental conditions, or noise reduction, which filters out background sounds.
Once connected with an app, however, these capabilities can also be set or modified by the user, but require explicit intervention. The addition of AI and Machine Learning aims to make such user intervention unnecessary, enabling hearing aids to adapt to the conditions or user’s personal preferences automatically. With reinforced learning, the performance should improve over time. By allowing data to be shared securely with other users through a smart phone app, all users will benefit from the improvements made as the system learns through Artificial Intelligence.
Hearables are the new wearables
Through the integration of all these features, manufacturers have significantly moved the dial on how well hearing aids operate and the user experience they deliver. While they are primarily intended to help people with hearing loss or other auditory impediment, it reflects a more general trend towards the adoption of wearable technology. There are already examples on the market that integrate advanced features, such as fall detection and body temperature monitoring. We can expect to see more of this in the near future.
This will also be extended to include more advanced features, enabled in large part by the companion smart phone application and the Internet. For example, real-time language translation and transcription are two features already being offered by at least one hearing aid manufacturer. Other convenient functions, such as background noise cancellation, are also becoming popular with consumers. As this trend continues there is likely to be some convergence between hearing aids and earphones more commonly used for listening to music and making calls.
Some hearing aid manufacturers are already integrating seamless interfacing to voice- controlled digital assistants. There could also be unexpected benefits to this trend; the convenience that comes from wearing a hearing aid could encourage more people to explore their use, which may ultimately benefit those who need supplemental auditory support but don’t yet know it.
The enabling technology
This evolution is primarily enabled by advancements in digital signal processing – the creation of superior DSP architectures that offer higher performance processing with fewer clock cycles, along with exceptional audio quality. These DSP advancements are in turn enabled by a semiconductor processing technology that provides the platform for small and more highly integrated solutions to achieve the feature density necessary to realize the next evolutionary step in hearing aids.
An example here is the latest device in the EZAIRO® family of DSP-based audio processor solutions from ON Semiconductor®. Targeting hearing aids and hearables, the Ezairo 8300 features six processing cores, each optimized for performance in this application area. This includes three DSP cores, a microcontroller and two hardware accelerators, one of which is a neural network accelerator designed to execute AI and Machine Learning features in an energy-efficient way and without the need for additional processor intervention.
As Figure 1 shows, the architecture of the Ezairo 8300 includes flexible input and output stages, along with a number of on-chip peripherals and interfaces.
The features offered by the Ezairo 8300 have been developed to enable a wide number of advanced and emerging products, including wireless-enabled hearing aids that are medically prescribed or available over-the-counter. The new high quality ADC and DAC of the Ezairo 8300 allows the chip to handle HiFi audio, which is instrumental in enabling a new generation of hearable and wearable technology with added convenience and consumer features. The middle ground here also includes the more conventional headphones and headsets.
The existence of hearable and wearable technology is enabled by advancements in digital signal processing, coupled with wireless connectivity and the integration of AI or Machine Learning.
With ease-of-use and long battery life being major considerations to the success of this technology, the Ezairo 8300 is perfectly suited to enabling these new markets. It offers unprecedented feature integration with ultra-low power consumption and is built on cutting-edge 22 nm process technology. This puts it at the forefront of this emerging trend and ON Semiconductor plans to accelerate this through an expanding Ezairo product portfolio.
All trademarks and logos are the property of their respective owners. ON Semiconductor and the ON Semiconductor logo are trademarks of Semiconductor Components Industries, LLC dba ON Semiconductor or its subsidiaries in the United States and/or other countries. The Bluetooth word mark and logo are registered trademarks owned by Bluetooth SIG, Inc. Alexa is a trademark of Amazon.com, Inc. or its affiliates. Google is a trademark of Google LLC.