Speech processing is an important technology for enhanced computing because it provides a natural and intuitive interface for the user. This chapter explains the Microsoft Speech Application SDK, shows you how to create, debug, and tune a speech application, and how to set up a telephony server.
Creating .NET Applications That Talk
2005-04-03 .NET 4 Comments
Microsoft, or rather Bill Gates, is really trying to push forward all this speech recognition thing. Am I the only one wondering why this is so important? I just don’t see myself sitting behind my PC talking to it. I guess this would be helpful in some businesses to “write” a quick report, but then again secretaries working there type faster than most men can speak
I could see lots of uses for speech recognition. Not necessarily inputing data into a typical PC but, things like robots… Telephone answering machines, androids 😀 Imagine a house that talks to you like the computer does on Star Trek. You could walk in an tell it to turn on the oven for you and we could be even lazier than we are now 😀
Though some of this is targeted towards adding speech interfaces to desktop applications that you use while sitting at your computer, one of the main uses for this technology is creating applications whose only interface is speech, or adding speech as an alternative interface for desktop and web apps so the same business logic can be used to allow customers to also access information via non-PC interfaces such as the phone.
Check out Microsoft Speech Server
and some demos
for more details.
Microsoft’s product activation call center uses this technology to help process activation requests.
Speech is quite a bit further advanced that it was in 2000. It’s quite performant these days.
I guess you won’t be able to type faster than you talk.
Besides, with this faster-than-light typing comes considerable strain on your hands and wrists. This develops into serious pain over time. From that perspective there are certainly good uses for speech technology.
Now, for Microsoft to develop this into anything worth using, they have to put a lot of people, time and money behind it.
And it’s going to take bloatware to new highs when it comes as part of the system. The speech recognition part of the system alone is easily going to devour a cool gigabyte of data to install. Then you have to park the cow on your hard drive too… The system builders are going to have a field day when Longhorn is introduced to the world as a finished product. The system requirements are going to read like the spec of the space shuttle compared to previous systems .