Pittsburgh, PA — (press release excerpt) — Cepstral, LLC today announced the release of Theta, a state-of-the-art software engine for deploying the highest quality spoken language output possible in a small footprint.

Theta is a speech synthesis engine that uses unit selection technology that can produce the style, character, and delivery that fits an application. Theta's footprint allows voice delivery on small devices, as well as multi- port installations on servers.

According to Kevin Lenzo, CEO and co-founder of Cepstral, “Theta . . . can put real customer quality voice output into devices, PDAs, games, cars, services and speech applications, and meet the increasing demand . . . for characteristic, human voices. Even robots need to have the right voice.”

High quality voice synthesis is needed in any situation where natural- sounding and intelligible verbal communication is required, but where the delivery of live or pre-recorded human voices is prohibitively expensive, technologically complex, or insufficiently adaptable. Cepstral's Theta enables assistive technologies, telematics applications, games, toys, servers, desktops, and PDAs to speak to their owners, clients, and customers. Game characters can interact with natural sounding voices, and Web sites can engage customers with spoken personalized information.

Text-to-speech speech synthesizers have historically been slow, artificial-sounding and inaccurate. Many such systems still require high capacity computer servers to host a few channels of voice, and take many months to implement. With Theta, Cepstral has reversed this trend and opened possibilities to new clients and partners once wary of the technology.

Theta allows the delivery of scaleable-footprint voices that can be customized to specific devices or applications, yielding the best possible spoken language output for the target platform. Theta enables synthesis on devices and server deployments with the quality and character that speech systems need to deliver a great voice experience and increased productivity. The output can be tailored to an application or domain, resulting in much higher quality output than previously possible with synthetic speech.

Theta is available for most popular computer platforms, including Microsoft Windows, and WinCE/PocketPC, Linux, Apple OSX, and QNX. The total footprint, with voices, varies in size according to each implementation — from 2 to 32 megabytes per voice instance. Custom tailored to clients' unique needs, these voices can be built for general or domain synthesis, and multiple voices can be deployed concurrently on the Theta engine.

 
This article was originally published on LinuxDevices.com and has been donated to the open source community by QuinStreet Inc. Please visit LinuxToday.com for up-to-date news and articles about Linux and open source.