Software-Defined Hearing (SDH)

Software-Defined Hearing (SDH) is a revolutionary technological framework that decouples hearing aid functions from specific hardware, abstracting them into a dedicated software layer. In this architecture, core hearing capabilities—such as Wide Dynamic Range Compression (WDRC), noise reduction, and feedback management—are implemented through programmable code rather than being hard-coded into the silicon of a chip，or IC (integrated circuit).

Much like the transition from feature phones to smartphones, SDH allows a single hardware platform to be dynamically updated and customized. By utilizing high-performance, general-purpose System-on-Chip (SoC) platforms, SDH creates an agile environment where hearing aids can evolve through software updates (OTA, Over the Air), providing users with continuous improvements in sound quality and personalized features without needing to replace the physical device.

Software-Defined Hearing (SDH) enhances traditional hearing aids by shifting hearing functions from hardware to software. This approach offers several benefits:

• Faster Product Updates and Improved User Experience: Software updates enable quick feature additions and improvements.
• Open Integration: Easier integration with cloud, smartphone and AIoT devices.
• Cost Savings: Reduced reliance on specialized hardware components lowers manufacturing expenses.

AI-Defined Hearing (ADH) represents a fundamental paradigm shift from rule-based signal processing to dynamic, self-learning auditory systems. Built upon the foundation of Software-Defined Hearing (SDH), ADH is a framework where Artificial Intelligence (AI) algorithms decide how sound is processed and delivered in real-time, creating highly personalized and adaptive experiences. Rather than relying on fixed rules, ADH integrates machine learning and cloud connectivity to transform hearing aids from simple medical devices into intelligent AIoT platforms that evolve with the user’s preferences and environment.

AI-Defined Hearing architecture is built upon three distinct layers:

• Edge AI at the Embedded Hearing Processing Layer: This layer incorporates AI at the device level, enabling real-time processing directly within the hearing aid. This includes functionalities like CNN and DNN for noise reduction and voice abstraction.
• AI Agent at the Application (APP) or PC Layer: An AI agent operates at the application level, whether on a mobile app or PC software. This layer is responsible for fitting and tuning the hearing device, acting as an intelligent interface between the user and the hearing system.
• Big AI Model for Learning and Knowledge (Cloud Layer): At the cloud layer, a large AI model leverages big data and AI learning to continuously improve and expand the system's knowledge base. This layer facilitates deeper analysis, learning, and overall enhancement of hearing capabilities.