Zarlink Semiconductor is sampling a single-chip voice processing technology targeted at hands-free communication systems, including high-end car kits and speakerphones. The voice processor is a flexible and programmable platform integrating dual-channel codecs with 16-bit sampling, audio-quality digital-to-analog converters (DACs), and increased MIPS capacity. It helps designers of speakerphones and hands-free car kits to more effectively deal with complex noise conditions, such as high background noise, echo, and double-talk.
Meeting a growing demand for voice conferencing systems delivering superior voice quality, Zarlink’s voice processing engine is designed specifically to handle severe noise conditions. The device’s patented software algorithms continuously track echo path changes and reduce background noise while preserving high voice quality.
The chip also delivers exceptional performance in double-talk situations when two people are talking at the same time. Most hands-free communication systems only deliver half-duplex operation, meaning only the voice signal with the highest power is transmitted. As a result, the conversation becomes stilted.
Many competing full-duplex solutions also will stop tracking the echo environment during double-talk and then have to readapt once double-talk has stopped. This results in an audible burst of echo. In comparison, Zarlink’s hands-free voice technology algorithm can continuously converge and track changes in the echo path to support full-duplex operation.