Vista Speech Macros Beta for Developers Released
- By Kurt Mackie
- April 29, 2008
Advanced developers who want to tell their computer what to do (literally), rather than punch keys or click on icon, may want to check out a "pre-beta" technical preview of Windows Speech Recognition Macros for Vista. The release of the technical preview was announced by the Windows Vista Team Blog on Saturday.
When speech macros are created, the user can speak a word or phrase into the PC's microphone and, if the speech is recognized, an action will be executed.
It's not exactly clear what actions this beta can carry out. Microsoft provides two demonstrations of the beta by Microsoft MVP Will DePalo showing how to program an API for the Windows Speech Recognition Macros using C++ in Visual Studio. The demo, after much work, shows text being returned at a command line prompt in response to spoken words.
Microsoft seems to have much more in mind for these macros than just returning text in response to speech. A blog post from late last year suggested users would be able to create speech macros "with no programming experience what-so-ever." Possible actions in response to an uttered word or phrase might be to "send keystrokes" and "launch programs," the blog stated. However, the coding complexity shown in DePalo's demos suggests such capabilities haven't arrived just yet -- at least in terms of having no programming experience to set up the macros.
Programmers who want to get started with the speech macros have to first run a microphone wizard to set up the system. Microsoft recommends using a headset with a high-quality microphone. You also need 1 GB of memory when running the macros with other applications.
Instead of using C++, as shown in the demos, programmers can access the built-in capabilities of the macros via JScript or VBScript, according to an MSDN post.
The current technical preview of Windows Speech Recognition Macros for Vista is available in English only. The application can be downloaded here.
Kurt Mackie is senior news producer for the 1105 Enterprise Computing Group.