Microsoft unveils .Net speech platform

Microsoft Corp. unveiled Oct. 30 the first technical preview of its .Net Speech Platform and also announced availability of the second beta release of its .Net Speech Software Developer Kit (SDK) at the Speech TEK conference in New York.

As unveiled, the speech platform contains the Microsoft speech recognition engine, the middleware to connect into a telephony system, the SALT (Speech Application Language Tags) interpreter, a SALT voice browser, and the SpeechWorks text-to-speech engine.

The .Net Speech Platform will give developers and customers a foundation to design a single application that can run a speech-enabled application in a variety of venues, including telephony, desktops, and in a multimodal format on mobile devices. The platform is expected to enter beta testing next summer and ship by the end of 2003.

Although still in pre-beta, this is the first real look at Microsoft’s speech technology. One industry analyst said it is the first proof point that Microsoft is serious about speech.

“This is more than a toe in the water,” said Brian Strachman, senior analyst at In-Stat/MDR in Scottsdale, Ariz.

The second announcement at the show was the release of Beta 2 of the .Net Speech SDK. The .Net in the .Net Speech SDK refers to the fact that the speech developer kit will integrate into Visual Studio.Net, said James Mastan, director of marketing for the .Net Speech Technologies Group in Redmond, Wash.

“Because it fits into the IDE [integrated development environment], a developer can install speech functionality into the rest of the Visual Studio product,” Mastan said.

The integration also allows an application that incorporates speech to run on any Web server running Microsoft ASP.Net (Active Server Pages), according to Mastan.

The kit’s toolset includes the grammar development tool, a prompt creation tool for creating and managing prompts, a debugging tool, the ASP.Net SALT Web controls, and the add-ins for Internet Explorer for multimodal clients.

The SDK will also include a tool for creating and testing SALT-based telephony-only applications, and will use the W3C (World Wide Web Consortium) formats for speech grammar.

“This means all the other applications that conform to the W3C grammar standard can be ported to SALT,” said Mastan.

Although deploying speech applications to a large consumer audience has not kept pace with the hype about speech, Strachman said Microsoft will help make the market.

“Even if Microsoft takes a big piece of the pie, it will be good for its competitors because it will be a much bigger pie,” Strachman said.

Recent Microsoft partnerships with Intervoice, a leading IVR (Interactive Voice Response) designer, and Intel, a supplier of Dialogic speech boards for telephony, also indicate that Microsoft’s speech technology is moving toward mainstream acceptance.