Company Brings AI Voice Recognition to Raspberry Pi

Matrix Labs is releasing the Voice for the Raspberry Pi after a successful crowdfunding campaign. With the accessory, the mini-computer gains access to a set of microphones, with which a connection with, for example, Google’s and Amazon’s voice control platforms is possible.

The Voice is a 3.14 inch diameter disk-shaped accessory that fits directly onto a Raspberry Pi. The devboard contains seven mems microphones that are connected to a Spartan6 fpga from Xilinx, supplemented with 64Mbit sdram. On top are eighteen RGBW LEDs. Optional is the ESP32 module, which provides WiFi and Bluetooth support.

Matrix Labs speaks of an open source platform and developers can use JavaScript, Python and C++ to develop for it, among other things. With apps, they can tap into the voice recognition services Microsoft Cognitive Service, Amazon Alexa Voice Service, Google Speech API, Wit.ai and Houndify. The Voice includes algorithms for noise suppression, remote voice capture and dereverberation.

The makers hope to deliver their device in May, the starting price on Indiegogo is $55. Matrix Labs previously came with a more extensive dev board: the Creator. It also contains a temperature, ultraviolet and pressure sensor and an accelerometer, gyroscope and magnetometer. The Creator is available for $99.