Talking to Your Computer

Many of us who use computers do a lot of typing and mouse-clicking — and pay for all those small, repetitive motions with serious aches, pains and stiffness in the joints.

Personally I have had problems with the finger I use for clicking the mouse. Another brother here developed carpal tunnel syndrome in both wrists and had to have surgery.

A few years back, we would have had just two basic choices: Live with the worsening pain or find another line of work. Today, however, there is a third option: voice-recognition software.

I'm dictating this article aloud, using Dragon Naturally Speaking Version 7 software. This approach to writing takes some getting used to, but it works quite well.

What makes this program particularly appealing to me is that that it lets me avoid not just the keypad but also the mouse. I've even found that it's not easily confused by ambient noises in the room such as a fan blowing or a bird chirping outside my open window.

It's easy to set up, too. When voice-recognition technology was new, the user had to speak for around 45 minutes to “train” the program to recognize his or her voice. I had this program trained in just five. I am able to bounce back and forth between dictation and commands. And I don't believe I have much of an accent, as I'm from Michigan, but a Bostonian has told me it works fairly well for him. Interestingly, the smart software analyzes documents you have already written. Apparently, this helps it recognize your style of writing.

Unlike Captain Picard of the Starship Enterprise, I'm not pacing the room while talking to my computer; I'm seated and wearing a headphone with a microphone. I'm sure it's possible to get a wireless headset for those who like to walk around. It's also possible to dictate into a pocket PC or use a digital recorder that can be plugged into your computer.

Oops — time to switch back to typing for awhile. Our next-door neighbor has just decided to mow his lawn. Ambient noise is one thing, but that kind of clatter is in a different category! (As we are temporarily living in a rented house, our monastery is located in a thickly settled neighborhood.)

For many of us, answering e-mail means doing a ton of typing. Speaking the text into the computer would speed things up and save our aching extremities. Other activities like chat and instant messaging would become faster and easier as well.

Okay, I'm back to dictating once again — in a different room. This time I'm using the voice-recognition software that comes with Microsoft Office XP. You turn it on by selecting “Speech” under the “Tools” menu on the Microsoft Word toolbar. I don't find this application as easy to use as its Dragon counterpart. But the dictation accuracy is fairly good, for five minutes of training. It also has some limited voice-command options for controlling Office programs, but it's not designed for mouse-free control.

Recently I tried using Dragon with Microsoft's Internet Explorer browser for surfing the Internet hands-free. I found that I was able to go to any Web page on my “Favorites” menu, enter a Web address in the address bar, scroll within a Web page and perform lots of other surfing-related functions.

This use of the software takes a little more practice than simply dictating an article. But for those who have had chronic and severe repetitive-strain problems, I believe it would be well worth the investment of time.

Those interested in making the switch to voice-recognition software will find three main contenders vying for their attention: IBM's ViaVoice, ScanSoft's Dragon NaturallySpeaking and Ultimate Interactive Desktops’ Voice Studio 2003. All these products let you speak commands to your PC, as well as dictate directly into most Windows applications. And they're all fairly inexpensive.

With Dragon, you can have text read to you aloud from e-mails or other documents. I'm not sure whether or not the other two offer that feature. Before buying, make sure your computer measures up to the system requirements. Dragon requires an Intel Pentium III/500 Mhz or equivalent with 128 megabytes of RAM and 300 megabytes of free hard-disk space. Usually, performance is better with more than the minimum system requirements.

Microsoft is looking into incorporating voice into future versions of the Windows operating system. That could only make the system, which is not without its hurdles for those who name “ease of use” as a top priority, truly user-friendly.

Several companies are working on speech recognition and you may already be experiencing the results. California's Nuance Communications and Boston's SpeechWorks are two of the market leaders in interactive voice-response systems. They have developed software that lets the computer understand — and respond to — routine natural-language requests. One way it's being used is on telephone help desks.

In fact, SpeechWorks’ call-center technology is used by such big businesses as Office Depot, the U.S. Postal Service, Thrifty Car Rental, United Airlines and Amtrak. The latter got back its $4 million investment in labor costs with this technology in 18 months. Nuance's software is being used by Schwab, Sprint PCS and Bell Canada. Intel is looking into audiovisual speech recognition.

With this kind of rapid deployment, maybe we will soon be witnessing real-life human-computer interaction like the kind the astronauts had with HAL 9000 in the movie 2001: A Space Odyssey. Stay within earshot!

Brother John Raymond, co-founder of the Monks of Adoration, writes from Venice, Florida.