Speech-recognition software can be useful in many applications, but its potential remains unfulfilled.
More than 35 years ago, the original "Star Trek" series and "2001: A Space Odyssey" conjured up sweet dreams of communicating with a computer through speech. Fifteen years later, I finally owned my first computer, but the speech part of my fantasy has remained an odyssey in progress. Most recently, that journey includes evaluating the current best-of-breed speech-recognition software, Dragon NaturallySpeaking. Alas, while the state of the art has vastly improved and NaturallySpeaking can be useful for some of us, my personal dream remains unfulfilled.
NaturallySpeaking comes in various flavors beginning at about $100. I evaluated the Preferred version at twice that price. And I had a second evaluator, a person with 15 years as a technology publishing professional, take a look at it as well.
Speech-recognition software has many uses, including dictation and transcription of recordings, and user-interface and application control for things like surfing the Web and editing email. My primary interest in NaturallySpeaking was in dictating text using the product’s included headset. The technological results may be the best I’ve ever had with a headset that has been included with a desktop software product. But both my co-evaluator and I found the physical design of the headset awkward in practice. In my case, blame a super-sized head. While I experienced no special concerns whether wearing my glasses or not, my co-tester found her glasses and the headset were contentious.
NaturallySpeaking works with Microsoft Word, adding its own toolbar to the application. Correcting misrecognized words brings up a set of alternative choices that gives new meaning to the term "dialog box."
Click to Enlarge
Both of us found it reasonably easy to “train” NaturallySpeaking to our personal voices an exercise that takes up to an hour initially and a process you need to continue as you use the product. NaturallySpeaking plays well with most Windows products, such as Microsoft Word. Also, the dictation accuracy is about as good as it gets these days far better than in the past.
But the good news stops about there. While the product claims a three-fold increase in speed compared to keyboarding, in informal testing I could type about five times faster than dictating, even disregarding the many typos that had to be corrected after dictation. And NaturallySpeaking went crazy when faced with unexpected input, such as coughing.
My co-tester’s primary need was transcribing interviews from a digital recorder. She was delighted that getting her recorder to work with NaturallySpeaking was a relatively simple process. But perhaps partly because interviewees obviously can’t be asked to “train” NaturallySpeaking for their voices, the results, she said, were disappointing: “There were words, but [they] didn’t make any sense. And if you looked at the text while you listened to the recording, there were no similarities at all.”
NaturallySpeaking is powerful software, but it needs to matched with powerful hardware. You must have enough horsepower to avoid falling asleep as NaturallySpeaking applies its excellent translation algorithms to drive the buggy. The product’s minimum specs of a 500 MHz processor and 256MB of RAM simply aren’t realistic. Multiple each of those figures by four before even thinking about hitting the road.
Though NaturallySpeaking isn’t ready for prime time for the purposes for which we tested, it could be valuable for some users, such as those with slow hunt-and-peck typing skills or physical keyboarding limitations.
Dragon Naturally Speaking 8 Preferred Nuance Communications www.nuance.com/naturallyspeaking Price: $199.99 Summary: Alas, while the state of the art has vastly improved and NaturallySpeaking can be useful for some of us, my personal dream of conversing with my computer remains unfulfilled.
J.W. Olsen has been a full-time IT author, columnist, editor, and freelance book project manager with more than 1000 editorial credits since 1990, and has provided computer, Web site, and editorial services to other clients since 1985. He welcomes feedback via the response form at www.jwolsen.com.
The Business of Going DigitalDigital business isn't about changing code; it's about changing what legacy sales, distribution, customer service, and product groups do in the new digital age. It's about bringing big data analytics, mobile, social, marketing automation, cloud computing, and the app economy together to launch new products and services. We're seeing new titles in this digital revolution, new responsibilities, new business models, and major shifts in technology spending.
Join InformationWeek’s Lorna Garey and Mike Healey, president of Yeoman Technology Group, an engineering and research firm focused on maximizing technology investments, to discuss the right way to go digital.