The next generation intelligent solution for smart voice applications is now generally available. Our recently announced xcore-voice processor is a complete voice solution built on our third generation xcore architecture, xcore.ai.
Fast, flexible, and economical, the solution enables a wide range of voice and intelligent IoT applications. Armed with xcore-voice, product designers are now fully equipped with all the tools and the software necessary to launch their next revolutionary, voice-powered product, tapping into the demands of modern consumers, who are becoming increasingly accustomed to voice-enabled technologies.
The solution also provides voice pipeline example designs using XMOS industry-proven audio front-end, incorporating far-field processing and support for third party voice algorithms.
Benefiting from xcore.ai’s unique architecture, xcore-voice is designed to deliver voice interfaces cost-effectively without sacrificing audio quality – either standalone, or as a co-processor capable of freeing up the voice processing workload from an existing SoC.
The platform delivers in three key areas:
Designed to minimise power consumption, xcore-voice offers a true voice triggered standby mode. At less than 55mW, xcore-voice provides high integrity ‘always-on’ capability to meet energy saving obligations across a growing number of electronics categories.
The xcore-voice solution comes with an industry proven, high performance, far-field algorithm as standard.
Alongside this, it also enables a flexible voice processing workload in a package that allows for customisation and feature upgrades. And that’s not all. The processor has in-built application specific software components, such as tunable algorithms and core application software components such as FreeRTOS. It is fully equipped with everything a developer would need to create an entire high-end, voice-enabled product.
Lastly, xcore-voice is fully customisable, offering unprecedented flexibility. The processor’s general-purpose, DSP, AI and IO processing capabilities can be adapted to the target application using standard embedded software techniques, drastically reducing the time-to-market when compared to custom silicon or multi-chip solutions. In addition to this, the integration pathways have been significantly simplified for designers, while still offering a complete BOM-optimised solution for turnkey integration.
The inherent power, performance and flexibility of xcore-voice means that the solution can be deployed in a variety of home electronics categories, including smart TVs, smart lighting, fitness machines, toys and home security systems – making it a great option for manufacturers who wish to expand existing portfolios or explore new possibilities.
Additionally, xcore-voice allows for 3rd party model integration. The solution includes an example local command model integration, highlighting the system capabilities to operate and respond with voice commands, without the need for connectivity.
The solution to an urgent issue
With consumers becoming increasingly accustomed to using voice technologies in their everyday lives, limitations of current technology have been put in the spotlight recently.
The need for product designers to deliver ‘across-the-room’ voice interfaces that offer the best possible quality, without compromising on speed-to-market or cost has never been greater. Not only to help designers avoid falling behind rival products, but to allow them to give consumers what they need to live their day-to-day lives.
Purpose-built to tackle that challenge, xcore-voice is delivering a high level of customisation in a cost-effective and power-friendly package. The processor is also enabling the acceleration of development and the ability to easily respond to a rapidly evolving market without cutting corners. It gives designers the capacity to experiment with and reiterate their designs at pace, while maintaining the all-important privacy of the end user.
If you would like to find out more about xcore-voice, or to contact us to discuss your requirements, simply click here, or follow the link below.