January 29, 2007

Microsoft Is Listening: Vista Speech Recognition Is Worth Talking About

As a professional programmer who also happens to be afflicted with spinal muscular atrophy (a severe neuromuscular disorder), PC accessibility is of paramount importance to me. Accessibility (or the lack of it) directly influences how efficiently I am able to work, which invariably influences my bottom line. More than that, it affects my state of mind. Being able to click that little red 'x' to close a window on your desktop may seem easy to most of you, but it can become quite tiresome or perhaps even be impossible to do for many users with disabilities. So when the world's most influential software maker introduces a new or updated accessibility feature, I take notice. And after test driving Windows Vista's speech recognition engine, it most certainly opened my eyes, er mouth!

Why am I so excited? Well, for one, speech recognition has finally become a first-class citizen in Windows. Before Vista, speech recognition was never installed by default in Windows (and for good reason). It used to only be effective in a very limited number of scenarios, like dictating in Microsoft Word, but, now, it is useful almost everywhere. Why is that? The short answer: It's truly integrated in the OS, which gives it much more power than ever before. The long answer: Nearly all Windows controls (text boxes, dropdown lists, menus, etc.) are now interfacing with the new Text Services Framework, but you can learn the details elsewhere from the experts.

So what does all of this really mean? Now, I can surf the web by voice without touching a mouse; I can click a point on the screen by speech alone; and I can dictate this article without typing on a keyboard. Pretty cool!

Of course, all of this has largely been available before in third-party applications, like Nuance Dragon NaturallySpeaking (DNS), but, in my opinion, never so elegantly and effectively with the entire user experience. Just try comparing Windows Speech Recognition and DNS when surfing the web in Internet Explorer or finding a file in Windows Explorer, and you'll quickly understand what I mean.

Windows Speech Recognition (WSR) still has room for improvement. One significant shortcoming of WSR is that there is no macro support yet. Also, my dictation is still more accurate in DNS, but the difference is minimal, and, with more use, WSR may very well eliminate that gap. Command-and-control is significantly superior with WSR, though, and the price is right (it's included in the OS). All in all, the speech recognition competition will definitely benefit consumers.

I, for one, am appreciative of all of Microsoft's effort put into speech recognition and am grateful it has become a mainstream feature in Windows. Indeed, I may have actually experienced a genuine "wow" moment because of it. ;-)