Architecture

Hardware

-defined

(HDH)

The traditional model. Hearing processing runs on a purpose-built ASIC — fixed at manufacture time. Improving it means new silicon: a new chip, a new certification cycle, a new product generation. It usually takes about 3+ years for a new chip design and production. Still how most hearing aids are built today.

Software

-defined

(SDH)

SDH, since 2018. Lyratone moved core hearing functions off the fixed ASIC into a software layer on general-purpose processors (SoC) — features and tuning ship as updates, not new chips. Same move networking (Software-Defined Network) and vehicles (Software-Defined Vehicles) already made and proven. SDH brings up the significant advantages of fast product itereation in months, open integration, and nature extension and adaption of AI.

AI

-defined

(AIDH)

AIDH, the evolution. Once hearing is software-defined, AI reaches the actual decision-making core instead of sitting on top of a fixed pipeline as a bolted-on feature. AI brings up the new advantages in both acoustics algorithms and service applications.

Software-Defined Hearing (SDH)

— The Foundation

Decouples core hearing functions from hardware into a dedicated software layer. That's what makes rapid feature iteration, open-system integration, and over-the-air updates possible. Lyratone has built on this architecture since 2018.

AI-Defined Hearing (AIDH)

— The Evolution

Because hearing is already software-defined, AI can reach the core of the system at all layers of embedded, rather than sit on top of it as a feature. Spans edge AI, an app/PC-layer AI agent, and a cloud model that keeps learning.