The IBM ViaVoice SDK for Windows, Linux and Macintosh(R) provides programmers with the necessary tools to develop applications that incorporate speech. It includes a robust set of application programming interfaces (APIs) that allows an application to access speech resources. It contains several utilities that enable developers to define and manage what the user can say within an application. There are also several sample programs that can help programmers as they develop their applications for speech. Finally, there are distributable runtime elements that are included with an application that uses IBM ViaVoice.

IBM SMAPI supports only speech recognition functions. The SMAPI interface set is the native interface for the ViaVoice engine. This section contains a description of the overall architecture of ViaVoice.

The heart of a speech recognition system is known as the speech recognition engine. The speech recognition engine recognizes speech input and translates it into text that an application understands. The application decides what to do with the recognized text. It can transcribe it literally for dictation, or it can act on it for commands.

Applications can access the speech recognition engine through a speech recognition API. For ViaVoice, this API is known as the Speech Manager API, or SMAPI, for short. SMAPI is a conventional API. This means that the API is defined as part of the resource; in this case, SMAPI is defined as part of the speech engine. With an API, speech becomes a resource to all applications, just like any system resource (mouse, video, and so on). Any of the ViaVoice SDK interfaces can be used in this manner, but SMAPI is the focus of the material in this guide. Windows developers note: The SMAPI interface set cannot be used in conjunction with other speech interfaces, such as SAPI, from within the same source code.

Download pdf IBM ViaVoice SDK SMAPI Developer’s Guide