The XK-VF3500-L33-AVS dev kit from XMOS enables developers of smart TVs, soundbars, set-top boxes, digital media adapters, and flat devices with stereo loudspeakers to evaluate the XMOS XVF3500 voice processor, and prototype far-field voice interfaces using the Alexa Voice Service. The VocalFusion Stereo Dev Kit for AVS uses a compact linear array of four digital microphones to capture voice commands and conversations in noisy environments and pass the captured speech to the Alexa cloud-based speech recognition system using an applications processor such as a Raspberry Pi.

Enabling voice control of devices located across the room from the user requires mitigating the challenges of distance, isolation, noise, and directivity, while also enhancing voice clarity. The XVF3500 achieves this using full-duplex stereo acoustic echo cancellation (AEC) with barge-in and programmable AEC latency, in combination with an adaptive beamformer and noise suppression, to locate the desired speech source and effectively isolate voice commands from the stereo audio the device is playing, while suppressing background noise and room echoes.

The kit contains an XMOS XVF3500 processor base board, linear microphone array (4x Infineon IM69D130 MEMS microphones), Raspberry Pi header (I²S audio and I²C control connectivity), micro USB connector for power (and optional USB2.0 device connectivity), low jitter audio quality clock, stereo DAC with integrated headphone amplifier, four user input buttons, and 13 LEDs providing direction-of-arrival indication. The XMOS solution packs DSP, I/O processing, and control software into a single chip, enabling customers to significantly reduce their bills of materials, flexibly tailor their designs, and reduce costs.