VOICE PROCESSING IN HOME AUTOMATION

Once just a feature of high-end luxury residences, home automation is now bridging to mainstream residences and will become a standard feature for new and existing homes.

As the market and technology matures, high-quality voice processing performance becomes increasingly important for home automation and security applications.

Interactive applications operated by voice recognition, for example the ability to control home appliances and door security systems, are key features of home automation networks. Poor acoustic echo cancellation, ambient noise and signal distortion make it increasingly difficult for a home automation system to perform reliably. If impeded by poor voice performance, voice recognition cannot easily detect commands to turn on/off appliances and voice authentication has difficulty verifying the user to allow access to the residence.

Integration of Telephony and Intercom

Home automation systems have traditionally included very basic voice processing techniques that provide half-duplex speakerphone performance. As these terminals integrate speakerphone functionality and home security systems rely on voice verification and recognition technologies, high-performing voice processing solutions become a key element in home automation system design.

DESIGNING A HIGH-PERFORMANCE SYSTEM

Overcoming Physical Limitations of Plastic Enclosures

To achieve high performance in a small plastic enclosure, small speakers are often driven in a non-linear range to meet audio requirements. An algorithm that can cancel non-linear echo and handle gain and distortion in the echo path solves this problem.

Overcoming Excessive Noise

To overcome excessive noise, an algorithm that provides good noise reduction while minimizing distortion is required.

Psychoacoustic noise reduction provides improved noise reduction over traditional methods and allows the system designer to achieve high performance at a lower cost by using a single microphone solution.

Designing a Scalable, Highly Integrated System

A good home automation system is scalable, allowing the end-user to build on the hardware as their needs evolve. It also allows the designer to develop a single platform and spin-off multiple variants to the system without having to redesign the hardware.

High integration, field upgradability and flexibility are also key criteria for a viable voice processing solution. Integration eases the design by reducing the complexity of interfacing multiple components and reduces the bill-of-material cost. Field upgradability and flexibility allow the designer to continuously enhance the feature set and add functionality without changing hardware.

ZARLINK'S HANDS FREE VOICE PROCESSING SOLUTION

Zarlink Semiconductor has developed a family of single-chip devices targeting high-performance hands-free communications, including home automation systems.

The company's newest hands-free solution, now commercially available, is a dedicated voice processor that combines integrated dual-channel wideband codecs with multiple interfaces. The device supports advanced echo cancellation, psychoacoustic noise reduction, full-duplex operation and is flexible and field-upgradeable. The chip's patented software algorithms continuously track changes in the echo path, even during double-talk conditions, and reduce background noise while preserving high voice quality.

The solution is based on a flexible platform with basic and advanced performance features for introductory and high-end systems. The field upgradeable solution allows designers to develop future platforms supporting new features, such as voice recording, verification, recognition, messaging and prompting, without re-designing the hardware

Link to White Paper: "Voice Processing for Home Automation Systems"