XMOS XVF3500-FB167-C

The VocalFusion XVF3500 processor provides smart TV, soundbar, and set-top box developers with a high-performance far-field voice capture front-end solution for prototyping and deploying "across-the-room" voice control products. It uses full-duplex stereo acoustic echo cancellation (AEC) combined with an adaptive beamformer and noise suppression to localize the desired voice source, effectively isolating voice commands while suppressing background noise and room echo. It connects directly to a linear array of four PDM microphones, with USB or I2S audio interfaces for connection to a system processor, making it an ideal platform for developers integrating voice interfaces into "room-edge" devices such as smart panels and soundbars. Application-specific evaluation boards and a range of tools are also available to support customization, debugging, and testing.

• Programmable voice processor with stereo AEC
  • 4-channel digital microphone interface
  • Voice DSP processing
  • Full-duplex stereo acoustic echo cancellation (AEC) with insertion capability and programmable AEC delay
  • Linear array adaptive beamformer
  • De-reverberation, noise suppression, automatic gain control
  • Direction of Arrival (DOA) indication
• Audio interfaces
  • I2S audio interface, 48kHz sample rate
  • High-speed USB2.0 compatible device, 48kHz sample rate
  • Multi-channel USB Audio Class 1.0
• Audio output
  • I2S output to DAC, up to 48kHz PCM
• System control options
  • I²C serial control interface
  • USB control interface
• 167-ball FBGA package, 0.5mm pitch
• XVF3500 processor board
• Linear microphone array
  • 100mm length, 33.33mm microphone spacing
  • 4x Infineon IM69D130 MEMS microphones
• Micro USB connector for USB2.0 device connection (and power supply)
• Interface for I2S audio and I2C control connections
• Stereo DAC with integrated headphone amplifier
• 4 configurable user input buttons and 13 user-controlled LEDs
• xTAG interface for JTAG debugging