Speech Technology Magazine

 

Lattice Semiconductor Releases Voice Detection and Command Solution

Lattice Semiconductor announces low-power, always-listening voice detection and recognition solutions for mobile devices.
Posted Dec 23, 2014
Page1 of 1
Bookmark and Share

Lattice Semiconductor, a provider of ultra-low power, small form factor, customizable solutions, has launched human voice detection and command recognition internal processors for smartphones and other handheld devices. These processors are implemented in the iCE40 family of mobile field-programmable gate array (FPGA) integrated circuits, enabling manufacturers to improve the user experience of their mobile devices with new voice activation capabilities and maximize battery life by minimizing false wake-up triggers to the processor.

The programmable human voice detection and command recognition solutions measure just 2.1 millimeters squared with power consumption about two-thirds less than that of competitors, according to the company.

The technology enables system power savings by ensuring processing sub-systems remain off until a verified human voice is recognized. It improves the user experience by activating exclusively to the owner's voice and providing near zero latency response to specific, high-use commands at much lower power consumption.

"Accurate and reliable voice detection and command recognition is important for today's smartphones and other mobile devices, but sound processing with a power-hungry processor quickly drains batteries when random sounds are misinterpreted as human speech," said Joy Wrigley, product line manager at Lattice Semiconductor, in a statement. "Further, user experience is diminished when devices do not properly capture commands. Lattice's low power, near-zero latency voice solutions alleviate these issues so manufacturers can improve the reliability of their devices' voice-activated functions while extending battery life."


Page1 of 1