Microsoft Boosts .Net Speech Technologies

.Net speech SDK Beta 2 supports SALT, multimodal speech apps, W3C grammar.

Enterprises and systems integrators looking to add speech capabilities to their applications have more options to consider from Microsoft Corp.

The Redmond, Wash., developer last week made available the second beta release of its .Net Speech SDK (Software Development Kit), as well as a technical preview of the Microsoft .Net Speech platform and a partner development program.

.Net Speech SDK 1.0 Beta 2, announced at the SpeechTek conference in New York, includes a tool set for the creation and testing of SALT (Speech Application Language Tags)-based, voice-only telephony applications. It supports the development of multi- modal speech applications on clients such as desktop PCs or Tablet PCs using Internet Explorer browser software, officials said.

This version of the SDK offers new enhancements to grammar, prompt creation and editing tools, as well as debugging tools for telephony and multimodal applications.

Other new features include World Wide Web Consortium standards-compatible formats for grammar authoring, a pre-built library of reusable speech telephony and application controls, and grammar libraries.

The .Net Speech platform—a set of technologies running on top of Windows Server—includes a Microsoft speech recognition engine, a text-to-speech engine, the SALT interpreter, a SALT browser and a telephony interface, among other elements.

Microsofts product groups are dabbling with the .Net Speech platform, said James Mastan, director of marketing for .Net speech technologies. Mastan declined to offer specifics but confirmed that business-to-consumer speech applications might be of interest to the MSN team, while business-to-employee, voice-activated, self-service applications might appeal to the Business Solutions unit, which oversees the MSCRM, Great Plains and Navision products.

Microsoft also last week announced a JDP (Joint Development Program) for enterprise customers and partners committed to building and deploying applications built with the .Net Speech SDK on the .Net Speech platform.

The JDP will help provide real-world production- environment-level testing by some of the most demanding Microsoft customers and enterprise partners, Microsoft officials said.

This builds on previously announced strategic alliances with SpeechWorks International Inc., Intervoice Inc. and Intel Corp. that will offer customers solutions and leading hardware and software components that integrate with the .Net Speech platform.

Bernard Elliot, an analyst with Gartner Inc., said the overall .Net Speech platform and Microsofts support for SALT were more important than the SDK itself.

"Theyre using an approach that makes it a lot easier to bring in a range of partners so that when theyre putting together applications, theyre not totally focused on the desktop environment or the Microsoft environment," said Elliot, in Stamford, Conn.

Also at SpeechTek, Rhetorical Systems Ltd. announced the latest release of its text-to-speech engine, Rvoice 4.0. New features of the Edinburgh, Scotland, companys software include modules for the correct pronunciation of U.S. addresses, a modular architecture so that system or application updates can be made without having to redeploy the whole system, and the capability to run multiple languages and voices on one server.

Rvoice 4.0 adds support for multiple text types, including e-mail, news, weather, financial, travel, names, dates and more.

Additional reporting by Mary Jo Foley, editor of Ziff Davis Microsoft Watch newsletter.