Firmware improves voice quality in speakerphones

Zarlink Semiconductor has brought out new firmware to enable its hardware platforms support high-quality voice performance in hands-free speakerphones and conference units.

Zarlink Semiconductor has brought out new firmware to enable its hardware platforms support high-quality voice performance in hands-free speakerphones and conference units.

The ZLS38502 firmware uses novel noise-reduction and echo-cancellation techniques to allow designers to overcome distortion, echo and double-talk issues that plague speakerphones.

The new firmware works with Zarlink's previously released ZL38004 hardware platform to provide a complete voice-processing system that simplifies design and speeds time to market of hands-free communication equipment.

"Speakerphone sound quality depends on a combination of voice processing algorithms and plastic enclosure design", says Andre Coucopoulos, Marketing Manager with Zarlink's voice processing product line.

"Zarlink's voice processing firmware supports full-duplex operation and integrates psychoacoustic noise reduction, echo cancellation and equalisation".

"As a result, our solution can better cope with lower quality plastic designs and support more natural voice conversation when compared to competing solutions".

Zarlink will be cohosting a European seminars series with Avnet-Memec to discuss hands-free communication design and introduce new technologies that ensure high-quality voice performance.

These free half-day seminars begin later this month.

Zarlink's firmware solution includes ZLAcoustics, a complete package of noise-reduction and echo-cancellation techniques solving speakerphone voice performance challenges.

Zarlink's ZLS38502 firmware integrates psychoacoustic noise reduction to reduce audible noise without significantly distorting voice.

Traditional noise-reduction methods remove noise across the entire frequency spectrum, resulting in lower signal integrity.

In comparison, psychoacoustic noise reduction cancels only the noise elements of the signal that are more noticeable to the human ear.

Due to limited space in a speakerphone's plastic enclosure, system designers must use small speakers but still meet high speakerphone volume requirements.

As a result, speakers are often driven to a level with high distortion.

Zarlink's ZLS38502 firmware cancels both linear and non-linear echo, allowing the designer to drive the speaker volume higher while minimising the effects of distortion on speakerphone performance.

An integrated 16/22 band equaliser ensures compliance to industry standards governing frequency response of speakers and microphones.

The equaliser can also amplify/attenuate specific frequency bands to improve voice quality for cost-effective, smaller speakers.

In double-talk situations, when two callers are speaking at the same time, poorly designed voice processing solutions will stop converging and resume convergence when double-talk ceases.

Callers will hear an audible burst of echo while the algorithm re-adapts to the new echo environment.

In comparison, Zarlink's voice-processing solution converges and actively tracks movement in a conference room and adapts to the changing echo path.

Zarlink's voice processing technologies are fully supported by a complete range of design tools, including evaluation boards, diagnostic tool kits and reference designs.

Back to top