Discover our platforms to differentiate
your next generation of energy efficient SoC

Meet us during the TSMC online technology symposiums and ecosystem forums

TSMC Virtual week

August 24, 2020
Nicolas Gaude

Nicolas Gaude - Business Development & Product Marketing Manager

TSMC OIP Ecosystem Forum NA & Europe, August 25th & 26th
“Innovative Ultra-Low-Power Voice recognition solutions powered by Dolphin SPEED Platforms in TSMC 22ULL”

Ying Zhao

Ying Zhao - Sales Director China

TSMC OIP Ecosystem Forum China, August 26th
“Innovative Ultra-Low-Power Voice recognition solutions powered by Dolphin SPEED Platforms in TSMC 22ULL”

Innovative Ultra-Low-Power Voice recognition solutions powered by Dolphin SPEED Platforms in TSMC 22ULL

The emergence of powerful and ultra-low power voice recognition algorithms drives the growth of connected devices controlled by voice. Hands-free interactions and voice control are becoming the new user interface standard. This is especially true nowadays with the Covid-19 pandemic upsetting our traditional habits of touch control. Highly accurate, flexible and smart voice recognition solutions powered by Ultra Low Power technologies are key requirements to serve leading-edge audio applications. It is a crucial step towards ultimate user experience and energy-efficient ubiquitous sensor networks.

To meet the future market requirements, Dolphin Design has developed in TSMC 22ULL a disruptive energy efficient audio system platform, called SPEED, relying on its design expertise in audio, power management and AI solutions. This audio platform aims at enabling a voice interface to almost any devices while adding an insignificant power cost and shortening the design Time-To-Market. Our SPEED platform merges together a digital SoC architecture with a comprehensive set of high-fidelity audio IP with state-of-the-art audio performances (SNR, THD, etc) and disruptive ultra-efficient voice activity detectors. Ultra-low-power consumption figures are achieved thanks to the use of innovative power management IP leveraging the outstanding performance of TSMC 22ULL technology and Dolphin Design historical expertise. A neural processing unit designed for Artificial Intelligence and Machine Learning audio applications is attached to the platform offering highly accurate wake word, command word and event detection with near-zero power consumption. The smart combination of Dolphin’s versatile expertises paves the way to unrivaled energy-efficiency levels to maximize battery lifetime of modern connected devices

In this presentation, we will share the content of the new 22ULL SPEED platforms and will demonstrate how Dolphin Design offering can enable breakthrough energy efficient audio product keeping BOM and integration costs and design cycles under control. We will show the full consistency and completeness of the audio platform offering differentiation to TSMC customers who wish to “voice-control more with less energy”.

Any question regarding our presentation?

Would you like to be updated on Dolphin Design’s news, events and offers (I understand I can opt-out at any time – read our privacy policy)