XMOS launches voice processor for smart home devices

XMOS, a UK-based smart IoT chip company, launches XVF3610 voice processor for smart home devices; launches XVF3615 variant of new platform with Amazon wake-up word, and starts alpha program for “Avona” voice reference design

Bristol, UK, 19 November 2021 — XMOS, the UK chip company, today announced the launch of the XVF3610 and XVF3615 voice processors: the next generation of high-performance dual-microphone voice interfaces for wireless speakers, TVs, set-top boxes, smart home appliances, gateway products and more.

Built on XMOS' xcore.ai chip infrastructure, the XVF3610 and XVF3615 deliver industry-leading voice processing solutions in a cost-effective and easy-to-integrate package. The designs mark an evolution of the XVF3510 voice processor already used in Amazon Alexa systems, providing powerful audio echo cancellation (AEC) and interference cancellation algorithms to suppress noise from primary sources and support cross-room voice barge-in mechanisms. The XVF3610 and XVF3615 also include automatic audio reference delay calibration - a key feature to support voice-over TV systems - expanding the range of devices and environments in which voice interfaces can be deployed. The XVF3610 is available now, and more information and ordering links can be found here.

The XVF3615 adds an embedded Amazon wake word to single-channel systems with Alexa technology, enabling manufacturers to offload wake word processing from their host SoC system. The product will be available starting January 2022. Product demos are available now and can be viewed here. Sign up for XMOS's newsletter to be the first to receive news when products are available to order.

Faced with a challenging supply chain, XMOS is confident that the XVF3610 and XVF3615 will be available in volume on standard delivery schedules.

In an additional development of the range of voice solutions offered by XMOS, the company today launched the alpha phase of its new “Avona” voice reference design for target customers. Avona leverages the flexibility of the xcore.ai platform to provide customizable reference designs for voice, wake word, speech-to-intent and intelligent voice applications.

For smart IoT applications, the Avona voice reference design enables engineers to create a complete voice interface solution by combining keyword, event detection or advanced local dictionary support. Partners and customers can also use the xcore software development kit (SDK) to develop their own voice-based solutions. The xcore SDK is programmed in C or C++ and integrates the FreeRTOS environment for user application code, supporting bare metal and RTOS use, providing customers with full flexibility and enabling them to differentiate their end products. The reference design will be generally available in early 2022. To stay informed, sign up for the XMOS newsletter here.

Mark Lippett, CEO of XMOS, said: “The IoT market is creating new verticals and opportunities at a breakneck pace, and we are seeing huge demand in the smart home space. As more applications become commercially attractive or viable, engineers and designers need to be able to rely on hardware that delivers agility and cost-effectiveness without compromising quality. When it comes to voice control, our new family of voice solutions gives them another tool in their arsenal, enabling designers to continue to push the boundaries of voice innovation in the smart home and beyond.”

“Versatility is key to all three of these new voice solutions, and it’s an inherent property of the xcore.ai chip architecture,” said Rob Fisher, Product Director at XMOS. “Designers are empowered to make choices: What compute balance do we want? How many voice elements are customizable? Do we program on bare metal or in a FreeRTOS environment? This flexibility will help them innovate in ways that would not be possible otherwise.”