Chip giant Motorola has announced a new processor with built-in speech recognition that can understand its user's voice even in noisy environments and may help people start talking to consumer electronics products like car stereos.
Motorola's digital signal processors (DSP), which are often used for digital audio information, feature built-in speech recognition software developed by the company's Lexicus Division. Lexicus's speech recognizer software achieves accuracy rates of 99.8 percent in quiet environments and 96 percent in high-noise environments, the company said.
The biggest challenge for speech recognition software is how to deal with the babble of background noise that typically surrounds a user's voice. That problem has limited the use of the technology for consumer products such as speaker phones, set-top boxes, car stereos, and toys. But Motorola expects its new processors to appear in a variety of consumer electronics devices.
The recognizer will be built into two DSP chips, the 56166 and the 56800.