Voice-processing platform for hands-free comms

A new voice-processing device enables high-quality voice in cost-efficient hands-free communication systems, including speakerphones, home automation applications and car kits.

Zarlink Semiconductor has brought out a new voice-processing device that enables high-quality voice in cost-efficient hands-free communication systems, including speakerphones, home-automation applications and car kits.

Available now, the ZL38005 voice-processing platform integrates dual channel narrowband codecs, on-chip memory, multiple interfaces and advanced voice-processing firmware.

Supported by a range of design tools, the voice processor improves voice quality, minimises background noise and reduces system complexity and cost.

The previously introduced ZL38004 platform is an enhanced voice processor integrating advanced firmware and functions for high-end hands-free communication systems.

Zarlink also offers the lightly featured ZL38002 digital echo canceller for basic hands-free communication.

Zarlink's voice-processing family is supported by a range of design tools that remove system-tuning complexity, help speed design time and shorten time to market.

Zarlink's design tools include a complete evaluation board, and a mini-evaluation board that fits directly into a speakerphone plastic enclosure to provide quick and easy performance testing.

Zarlink's diagnostic tool kit supports simple "plug and test" capabilities to easily measure echo profile, frequency response and distortion.

Zarlink also offers a Bluetooth evaluation kit for advanced hands-free car kits.

The ZL38005 hardware platform and integrated ZLS38501 firmware ensure hands-free voice quality in severe noise conditions.

For example, conference phones must cope with background noise, sound reverberation due to poor plastic enclosure design and acoustic coupling between the microphone and speaker.

Car-kit designers must provide systems that support high-quality voice despite severe noise conditions caused by engine, tyre and wind noise.

The ZL38005 platform includes narrowband codecs with input/output sampling of 8kHz, and integrated acoustic and line echo cancellers with programmable echo tail up to 256ms and noise reduction up to 20dB.

The device delivers exceptional performance in double-talk situations.

Most commercial voice-processing products fall back to half-duplex operation during double-talk, meaning voice signals pass in only one direction.

During double-talk only the signal with the highest strength is transmitted.

This results in a choppy and unnatural conversation.

In comparison, Zarlink's ZL38005 voice processor algorithm is able continuously to converge and track changes in the echo path to support full-duplex operation.

Designed in to a hands-free desktop speakerphone, for example, Zarlink's voice-processing technology continuously converges, actively tracks echo and adapts to the changing echo path.

Poorly designed voice-processing products will stop tracking echo during double-talk situations.

A lower-quality algorithm will re-adapt when double-talk ceases, creating an audible burst of echo.

The ZL38005 voice processing platform is available now in a 14 x 14mm, 100-pin lead-free QFP package.

Back to top