ReSpeaker: Modular hardware platform adds voice interfaces to any product you like
I've mentioned before that voice interfaces will provide new opportunities for great innovative products. But while big technology companies like Amazon, Google, Apple, Microsoft and Facebook invest huge amounts of money into developing the natural language interfaces that enable speech recognition, we've so far seen few smaller companies taking advantage of the rich diversity of available hardware and software that is available.
So it's great news that the Chinese hardware company Seeed Studios have announced on KickStarter their latest product, a modular development platform called ReSpeaker, with the objective of adding voice interaction to products in your home or office. And they have chosen XMOS to handle the ReSpeaker microphone array and deliver far-field voice capture with enhanced acoustic DSP performance.
As a KickStarter product there are hurdles to overcome before the product is available but by raising three times the initial goal of $40k in a week from more than 1k backers this is a product that people really want to get their hands on. With WiFi access to online services like Microsoft Cognitive Service Alexa Voice Service, Google Speech API, Wit.ai and Houndify, and on-board SD card slot for storage and the integrated lightweight PocketSphinx speech recognition engine, developers will have a huge range options for adding voice-interfaces to a variety of existing and new products.
Running on the Linux-based OpenWrt system with easy to use SDK, developers can build modules that work with the ReSpeaker module and share their work with other developers. Find out more at the ReSpeaker KickStarter page - this may be the start of a voice interface revolution.