XCORE-VOICE

xCORE Voice is a series of products produced by XMOS that contain all the algorithms necessary to aggregrate data from multiple microphones, and process the signal to produce an audio signal suitable for the humans of automated speech recognition systems.

The challenge in voice processing is to remove background noise, remove the far-end signal, and enhance the near-end signal to just contain the voice. The far-end signal is typically msuic or speech reproduced through a speaker. The near-end signal is what is picked up by the microphones which is a mix of the voice signal, the far-end signal, and echoes of both. xCORE Voice products contain algorithms to perform, amongst others the following tasks:

xCORE Voice products are based on the XS2 architecture, and include the XVSM-2000 chip.

...
Wikipedia