VocalFusion stereo-AEC voice processor for Amazon Alexa Voice Service

XMOS high-performance digital signal processing solutions help you build the Alexa Voice Service quickly into your products. Our stereo-AEC voice processor gives you up-close voice capture quality and processing accuracy at far-field range – even in noisy environments. Ideal if you want to get Alexa built-in to a smart TV, soundbar, set-top box or digital media adaptor.

Distance

Accurate voice capture and barge-in from across the room, even when music is playing loudly and the commands are spoken softly.

Clarity

Full duplex stereo-AEC, noise suppression and dereverberation cut unwanted noise and echo for optimised ASR processing.

Direction

Using beamforming to track the voice of interest, you’ll get best in class performance at all angles and over long distances.

Our rich optimisation parameters for the XVF3500 voice processor let you adjust noise attenuation, gain control and residual echo so you can tailor the best voice capture performance for your product. And our engineers are on hand to help you move rapidly from prototype to tuning, and from testing to volume production with a world-class product. See our newsroom for more about our latest customer implementations.

Development kit

XK-VF3500-L33-AVS

  • VocalFusion 4-mic stereo-AEC development kit for Amazon AVS

    XVF3500 processor board

    • XVF3500 voice processor
    • Micro-USB connector for power (optional USB2.0 device connectivity)
    • Low jitter, audio quality clock
    • Stereo DAC with integrated headphone amplifier
    • 4 configurable user input buttons
    • 13 user-controlled LEDs
    • xTAG interface for JTAG debug

    XVF3500 voice processor (programmable for far-field voice capture)

    • Full duplex stereo acoustic echo cancellation (AEC)
    • Programmable AEC latency
    • Barge-in
    • 4 channel digital microphone interface
    • Adaptive beamformer
    • Dereverberation
    • Noise suppression
    • Automatic Gain Control (AGC)
    • Direction of arrival indication (DOA)

    Compact microphone array

    • Linear, 100mm long
    • 4 x Infineon IM69D130 PDM (Pulse Density Modulation) MEMS microphones
    • 33.33mm inter-microphone spacing

    Connectivity

    • Raspberry Pi header
    • Inter-Integrated Sound (I2S) audio interfaces
    • Inter-Integrated Circuit (I2C) serial or USB control interface
    • High speed USB2.0 compliant
    • Multi-channel USB audio class 1.0
    • 167-pin FBGA package 0.5 pitch

Silicon

XVF3500-FB167-C

  • XVF3500 voice processor

    • 4-channel digital microphone interface
    • Full duplex stereo acoustic echo cancellation (AEC)
    • Programmable AEC latency
    • Barge-in
    • Adaptive beamformer
    • Dereverberation
    • Noise suppression
    • Automatic Gain Control (AGC)
    • I2S audio interface or high speed USB2.0 compliant device
    • I2S output to DAC, up to 48kHz PCM
    • I2C serial or USB control interface