06/08/2024 | 06:50 CEST | Group Announcements
Understanding speech in background noise (‘speech-in-noise’ problem) is the most pressing need to solve for people with hearing loss. Sonova is now addressing this challenge with the launch of Phonak Audéo Sphere™ Infinio, a unique hearing aid using real-time Artificial Intelligence (AI)1. Its proprietary AI chip instantly separates clear speech from background noise and allowed users in a clinical study to more than double their speech understanding in noisy environments, compared to current products in the market2. In parallel, Sonova is also launching the Phonak Audéo Infinio platform3. This new family of hearing aids marks the Group’s biggest release to date. According to the WHO, hearing loss affects 430 million people worldwide, a number that is expected to rise to 700 million – 1 in every 10 people – by 20504. Unaddressed hearing loss can adversely affect participation in social life5 as well as contribute to cognitive decline6.
Staefa (Switzerland), August 6, 2024 – Sonova Holding AG (SWX: SOON) today announces the launch of two new platforms under the company’s primary brand, Phonak: Audéo Infinio and Audéo Sphere Infinio, marking a new era of hearing aids built on a real-time AI based technology platform. Audéo Sphere Infinio builds on Sonova’s new and proprietary dual-chip technology. Of the two new chips, one applies real-time AI to sound processing. With this development, Sonova sets new standards for sound quality, connectivity and power management – all critical needs for people with hearing loss – along with helping to improve speech understanding and conversation quality. The new flagship instrument builds on the company’s expertise as well as existing technologies, while now integrating real-time AI sound processing on the hearing device to significantly increase speech intelligibility2. Customers can expect the new hearing aids to be available from August this year7.
Arnd Kaldowski, Sonova’s Group CEO says: “Solving the speech-in-noise problem is the long-standing goal of modern hearing aid development and essential to improving the quality of life of millions of people with hearing loss. While every user would state we have come a long way in the last 10 years, they will say they still can’t understand conversations as well as they would like to, when background noise is present, like in many crowded places, such as public transport or restaurants. We are convinced that with our new technology, we can make a significant difference for people in need.”
Andi Vonlanthen, Sonova’s Global Head of Research and Development for Hearing Instruments says: “During Sonova’s 77-year history, we have always been focused on addressing the most pressing challenges of people with hearing loss. With the development of our new real-time AI technology, Sonova has made a huge step forward – also in terms of technological ingenuity. To achieve these significant results, you must implement a powerful Deep Neural Network (DNN) in a hearing aid. This required us to develop a special chip – DEEPSONIC™, our DNN unit, as well as a specific training and testing environment, so we could optimize the chip’s performance. The entire process took many years and a significant investment in technology development. We are excited to see this innovation going to market now.”
About the Phonak Audéo Infinio platform and portfolio
Next to the new hearing instrument, Phonak is also launching the broad Phonak Infinio platform, which is powered by the new ultra-responsive Phonak ERA™ chip. This chip redefines standards for sound quality, power management and connectivity, resulting in reduced listening effort and fatigue8 for users, while significantly extending Bluetooth connection reach. The technology includes:
About the Phonak Audéo Sphere Infinio platform
Phonak Audéo Sphere Infinio is the first hearing aid powered by dual-chip AI technology. In addition to the ERA chip, the device also includes DEEPSONIC, a dedicated real-time AI / Deep Neural Network (AI / DNN) chip from Sonova's proprietary research and development. This has enabled a breakthrough in the long standing goal to provide optimal speech understanding in background noise, addressing the ‘speech-in-noise’ problem.
The dedicated real-time AI / Deep Neural Network (AI/DNN) DEEPSONIC chip provides the computational architecture optimized for DNNs (Deep Neural Networks) to address the ‘speech-in-noise’ problem. It has 53 times more processing power compared to current hearing aid industry chip technology11,12, which allows it to instantly separate clear speech from background noise in a fundamentally different way from the AI used in existing hearing instruments1. The user experiences significantly improved speech understanding and ease of understanding from any direction, allowing people to engage in dynamic group conversations2 without needing to constantly move their head.
About the value of technology development in the hearing aid industry
The hearing instruments industry has undergone important technological developments in the last 30 years that have continuously improved user benefit, while also addressing unmet hearing needs. As the global leader, Sonova has played a crucial role in these technological breakthroughs, including: the first directional microphone technology, Bluetooth connectivity, rechargeable instruments and the use of AI for personalized fitting. Real-time AI now promises an even more profound improvement for people affected by hearing loss, as it directly deals with the key challenge: understanding speech in background noise. Its use will pave the way for the next generation of technological advancements in the hearing aids industry. Addressing hearing loss early is important - according to the WHO, unaddressed hearing loss is estimated to pose an annual global cost13 of USD 980 billion14.
For more details on Phonak Audéo Sphere Infinio and the Infinio portfolio please refer to the global Phonak media release: https://www.sonova.com/en/media/phonak-elevates-sound-quality-new-infinio-portfolio-introducing-paradigm-shift-real-time-ai
**ENDS**
1 Real-time AI means that the AI DNN (Deep Neural Network) has been trained on sound samples to separate speech from background noise by processing the audio signal in real-time. This approach is fundamentally different from the use of AI for sound or acoustic scene classification, which is used in current hearing aids in the market. The real-time AI signal processing is performed by the DEEPSONIC™ chip, which is included in the Phonak Audéo Sphere Infinio platform together with the Phonak ERA™ chip.
2 Wright, A., et al. (2024). Spheric Speech Clarity applies DNN signal processing to significantly improve speech understanding from any direction and reduce the listening effort. Phonak Field Study News retrieved from https://www.phonak.com/evidence
3 The Phonak Audéo Infinio platform includes a single microchip: the Phonak ERA chip.
4 The World Health Organization (WHO) (2024, Feb 2). Deafness and hearing loss: https://www.who.int/news-room/fact-sheets/detail/deafness-and-hearing-loss
5 Why hearing health matters: https://www.phonak.com/en-int/well-hearing/hearing-health
6 Lin, F., et al. (2023, July 17). Hearing intervention versus health education control to reduce cognitive decline in older adults with hearing loss in the USA (ACHIEVE): a multicentre, randomised controlled trial. DOI: https://doi.org/10.1016/S0140-6736(23)01406-X
7 Planned launch dates: US and Canada early August, Europe and Asia-Pacific countries from late August onwards.
8 Latzel, M., et al. (2024) “Speech Enhancer reduces listening effort and fatigue.” Phonak Field Study News retrieved from www.phonak.com/evidence
9 Stewart, E., et al. "Adaptive Phonak Digital (APD) 3.0 is the preferred first fit compared to a leading competitor device". Phonak Field Study News in preparation expected August 2024.
10 The Bluetooth® word mark and logos are registered trademarks owned by Bluetooth SIG, Inc. and any use of such marks by Sonova AG is under license.
11 Åleskog, C., et al. (2022, Nov 8). Recent Developments in Low-Power AI Accelerators: A Survey: https://www.mdpi.com/1999-4893/15/11/419
12 Reuther, A., et al. (2019, Aug 29). Survey and Benchmarking of Machine Learning Accelerators: https://arxiv.org/pdf/1908.11348.pdf
13 Cost includes health sector costs (excluding the cost of hearing devices), costs of educational support, loss of productivity and societal costs. Of these costs, 57% are attributed to low- and middle-income countries.
14 The World Health Organization (WHO) (2024, Feb 2). Deafness and hearing loss: https://www.who.int/news-room/fact-sheets/detail/deafness-and-hearing-loss