Computing has more than its fair share of holy grails remaining to be found—the two-pound laptop, mouse pointer control by thought alone, crash-proof Windows—but perhaps the biggest of all is the idea of using our voices to free us from the keyboard.
Nuance has been plugging away at that goal since 1994, but in nearly 20 years of research and development, voice recognition software remains a niche product favored primarily by doctors, lawyers, and other people who do lots of dictation.
Why the holdup? A lack of computing power has been a big reason why dictation software has never taken off, and it's also been seen as complex, expensive, and inaccurate.
Now Nuance is back with the 11th version of the flagship dictation software product on the market: Dragon NaturallySpeaking. And Nuance has worked to address all of these issues with this release.
If nothing else, concerns about Dragon's accuracy issues are now completely resolved. In my testing Dragon never tripped up on words except for proper names and the (very) occasional homophone. Take proper names out of the equation and accuracy is functionally 100%. Training is simple — setup with all optional steps takes about an hour, only four minutes of which is actual dictation training — and the software is reasonably easy to use. Unless you start delving into using voice to complete Windows commands (Dragon can do virtually everything in Windows by voice), you rarely, if ever, need to delve into the program's menus or help features.


Comments
Back to top