xCORE-VOICE Processors | XMOS

DIRECTIONAL VOICE CAPTURE PROCESSORS

 

xCORE-VOICE processors combine high quality directional voice capture technologies (including beamforming, echo cancellation, noise suppression, direction of arrival and voice activity detection) with the high performance multichannel audio capabilities of xCORE devices, and control/sensor interfaces to deliver industry leading voice capture from PDM MEMS microphone arrays. The individual PCM streams can be backhauled over USB or TDM interfaces to application processors, DSPs and PCs, or sent over a secure wireless connection to cloud speech recognition services.

xCORE-VOICE far-field voice recognition solutions provide affordable voice interfaces for a wide range of product categories including: smart TVs and gaming consoles, IoT home hubs and smart home appliances, automotive and industrial.

XVSM-2000 functional block diagram

xCORE-VOICE XVSM-2000 functional block diagram

 

xCORE-VOICE CAPABILITIES

  • Programmable multicore microcontroller with integrated microphone and voice DSP
    • Beamforming with configurable beam width and non-stationary noise suppression (up to 20dB attenuation)
    • Echo cancellation (up to 80dB att.)
    • Stationary noise suppression (up to 20dB att.)
    • De-reverb (up to 20dB att.)
  • Voice capture
    • Up to 16 element PDM microphone array
    • Up to 100dB SNR
  • High speed USB2.0 or TDM (time division multiplexed) interface backhaul
  • Secure wireless connectivity to Cloud-based speech recognition services
  • Audio output options
    • I2S mono/stereo output
    • DSP options: compressor, limiter
  • System control options
    • Industry’s most flexible GPIO
    • Extensive general purpose libraries
Beamformer microphone

xCORE-VOICE Beamformer Microphone

 

Beamformer field response for various direct sound power (DTR)

xCORE-VOICE Beamformer Field Response