Visit the National Instruments web site
Click on the advert above to visit the company web site

Product category: Embedded Software and Operating Systems
News Release from: ARC International | Subject: Voice codecs for ARCtangent-A5
Edited by the Electronicstalk Editorial Team on 28 May 2003

Voice codecs run on customisable core

Request your FREE weekly copy of the Electronicstalk email newsletter. News about Embedded Software and Operating Systems and more every issue. Click here for details.

A new voice coding library for the customisable ARCtangent-A5 RISC/DSP core achieves performance levels comparable to a pure-play DSP - without the need for a second device to handle system functions.

ARC International has released a voice coding library for the customisable ARCtangent-A5 RISC/DSP core The library forms the first set of voice codecs available for a single-chip design that achieve performance levels comparable to a pure-play DSP - without the need for a second device to handle system functions

By implementing its voice codecs on its ARCtangent-A5 RISC/DSP core, ARC continues to deliver on its mission of simplifying the design process by reducing development time, costs and risk.

The codec library is optimised to run on the ARCtangent-A5, taking advantage of its customisable DSP extensions to reduce MIPS requirements.

As a result, the library delivers some of the most MIPS-efficient voice coding software available on a scalar RISC or DSP, in a unified RISC/DSP solution.

"Our highly customisable ARCtangent processor cores enable developers to easily modify and extend their designs for specific applications", said Mike Gulett, President and CEO of ARC International.

"With our ARC voice coding library, developers will be able to obtain the highest performance voice processing on a single processor system for their VoIP applications in a low- cost solution from a single vendor".

The unified solution eliminates the need for a multiprocessor system to handle applications such as voice over Internet protocol (VoIP), which have traditionally used a DSP for the voice processing and a RISC processor to handle the networking protocols.

As a result of eliminating the need for a second processor, the solution speeds a product to market by simplifying design while reducing silicon cost.

In addition, designers can work in a single software environment for both RISC and DSP development, both lowering cost and reducing time to market by significantly reducing development time.

The single processor uses less power than a multiprocessor design.

ARC's optimised voice codecs further lower a product's power consumption by significantly reducing the CPU loading for voice compression, allowing designs to meet performance goals at lower clock speeds.

In addition, because ARC's voice codecs use only a small portion of the total CPU capacity, developers have enough bandwidth to add differentiating features such as text-to-speech software and multi-caller conferencing to their designs without needing to increase the clock frequency.

Of special interest, is the fact that reduced power demand at the chip level means that expensive system components such as power supplies, batteries and thermal management hardware can be made smaller, perhaps even eliminated, thereby reducing system cost.

This allows a design based on the ARCtangent-A5 and its voice codec library to specifically meet the needs of VoIP applications, such as SOHO gateways and VoIP handsets, which require a low-cost, low-power solution.

ARC's voice coding library includes a full set of codec types, including A- law/mu-law, ADPCM, and CELP compression schemes.

In addition, the library includes algorithms such as line and acoustic echo cancellation (LEC and AEC) for suppressing analogue line and background noise effects.

ARC's voice codecs also allow designers to improve reliability in their VoIP applications by including packet loss concealment software, which protects applications against corrupt or lost data during the transmission of voice packets.

Designers can also reduce the network bandwidth demands of their designs with the library's voice activity generation (VAG) and comfort noise generation (CNG) components by not encoding lulls in voice conversations but still generating sound at the receiving end to keep users from experiencing dead silence.

The ARC voice codecs are fully compliant with ITU (International Telecommunication Union) standards, ensuring VoIP designs will provide high quality voice reproduction.

The library works with ARC's Precise/MQX operating system and Precise/RTCS TCP/IP stack, so developers can obtain all the elements of a voice over IP (VoIP) design from a single vendor, eliminating the need to integrate elements and ensuring a single source of development support.

As a result, product development can proceed quickly with minimal risk allowing developers to concentrate on adding their unique design value rather than spending time solving integration issues.

ARC International: contact details and other news
Email this article to a colleague
Register for the free Electronicstalk email newsletter
Electronicstalk Home Page

Search the Pro-Talk network of sites

Visit the National Instruments web site