Computer, Let Me Tell You Something…Voice Recognition Grows Up


From a July 2006 Reuters article describing an early Vista demo:

At Microsoft’s annual Financial Analyst Meeting…Vista product manager Shanen Boettcher set out to show just how easy to use the speech recognition technology built into upcoming Windows Vista software will be. Like, for example, dictating aloud a simple, heartfelt letter to mom, and having one’s voice automatically transcribed into a computer.

The result was a disaster.

Several tries at making the computer understand the simple salutation “Dear Mom” was read by Microsoft software as “Dear Aunt, let’s set so double the killer delete select all.” Attempts to correct or undo or delete the error only deepened the mess.

Remember the chuckles we had over that one? Here’s the video. Perfect, accurate voice recognition has always been a staple of science fiction, but until recently the reality has been closer to Microsoft’s embarrassing pre-Vista demo. As a web worker, you’re already used to having conversations with yourself. Now put it to good use.

According to Business 2.0 magazine, the market for voice recognition software is expected to quadruple by 2010.

Nuance has been leading the way with their speech recognition software. Their Dragon Naturally Speaking software for Windows has been improving steadily over the years. Version 9 achieves a level of accuracy that is almost uncanny. The trick is to take the time to train the software effectively, and to correct mistakes immediately. The software learns the more you use it. This involves reading paragraphs of text that can take up to 45 minutes or so. Last year I injured my shoulder and needed to lay off the keyboard and mouse for a while. I found that Naturally Speaking allowed me to continue to be productive. The key is to eliminate as much ambient noise and interference as you can. Ideal for the web worker who has more control over his/her working environment.

New York Times columnist David Pogue who works on both Macs and PCs regularly, wrote an in-depth review of Natually Speaking 9 which is what originally prompted me to buy it. Unfortunately, none of the developers working on speech recognition software offer demo licenses. You need to leap in with both feet, or lips, as the case may be.

Despite the laughs we got from the early demo, the release version of Microsoft Vista’s speech recognition software included in the operating system is getting some decent reviews. Maybe it works a little too well.

If you’re running OS X, the most recommended application is MacSpeech’s iListen, which seems to be replacing IBM ViaVoice as the option Mac users turn to. Still, neither application is getting the raves that Naturally Speaking does.

Not content to simply dominate speech recognition technology on the office front, Nuance is making bold moves into the market that really needs hand-free communication…mobile devices. They recently completed their purchase of MobileVoiceControl, software for Blackberrys and some Windows Mobile devices that let’s you navigate your phone by speech only. Anyone who has struggled with training clunky contact voice tags can appreciate the need for this technology. My Cingular 8125 is not currently supported by the Nuance product. However, I’ve worked with Microsoft’s similar Voice Command software and it does the job quite nicely. Provided your device can handle the heavy CPU and RAM requirements.

Ready to ditch the keyboard and save your manicure? First, get a good headset. makes some recommendations for USB microphones based on their use with voice recognition software. Some are pricey. But if you want to dictate your emails most of the time to save the wear and tear on your wrists, it’s worth the investment. The site also features excellent articles to help you get started with speech recognition. I found their advice on microphone positioning to be very helpful when I was getting started. Moving the microphone even a quarter inch increases accuracy.

Have you tried voice recognition software? Are there microphones that you recommend? Tips or tricks to increase accuracy? Take your fingers off the keyboard and share!


Chuck Lai

I’ve been using Dragon NaturallySpeaking 9 for the last several weeks. I’m actually using it right now to dictate this comment. I’ve been pretty impressed with its accuracy out of the box, but I have to make sure that I enunciate all my words. And I’m still getting used to the editing commands and dictating in general.

If I had to rely solely on speech recognition for dictating and editing I would be less productive than just typing right now. However, I believe I’m more efficient with a combination of speech recognition and typing.

As far as a microphone recommendation, I’m using a Plantronics DSP-400 headset. It works very well with Dragon NaturallySpeaking 9, but I wouldn’t recommend it for the mobile web worker. The DSP unit attached to the USB cord is very bulky.

Here’s a short overview of Dragon NaturallySpeaking 9 from my blog.

Comments are closed.