Become a Contributer!

Windows Aero

Authentic. Energetic. Reflective. Open.

  • Jul
    8

    Recently, I woke up with a throbbing pain in my right hand.

    That was worrisome, but to be honest, I do abuse my hands quite a bit: I type all day at work and spend my spare time playing drums. So I fumbled my way through the workday, yelping in pain each time I was forced to bend a finger.

    But eventually I found myself in the emergency room. When I left, my hand was encased in a cast. What does someone who types for a living do in a situation like this? I grabbed the dusty, neglected headset that came with my PC and fired up Windows Vista Speech Recognition.

    I’ve always been something of a fanatic about speech recognition. As a freelance tech journalist in the late 90s, I’d go to trade shows and hang out like a groupie at the booths of early innovators like Dragon Systems and Lernout & Hauspie.

    Say, what?

    But whenever I tested their latest offerings (DragonDictate and Voice Xpress), I’d soon get frustrated, tossing my mic in the corner until the following year. Why? The programs made too many mistakes, especially ones that were hard to detect when proofreading. One typical glitch was the confusion of homonyms such as “be” and “bee” or “to,” “two,” and “too.” But I’ve got other horror stories.

    I remember one product demo back in 1997. The company president, wearing a headset, was trying to show the audience how he could page through a calendar program with his voice.

    “Next day,” he told the computer.

    Nothing happened.

    Next day.”

    Still nothing.

    Eventually, he tried a different approach—he got louder. “NEXT DAY!”

    Suddenly, the computer sprang into action. It opened a calculator on top of the calendar.

    Clearly, speech was not yet ready for prime time.

    So I wondered, as I caressed my cast, if the speech recognition in Windows Vista was going to remind me of my bad experiences from days gone by. Was it usable? Or would it go as badly for me as it did for the guy who gave the now infamous July 2006 demo of an early beta version?  Would I rue my cast and curse my immobile fingers?

    "Nothing short of startling"

    As it turned out, I did very little ruing.

    I found the accuracy of Vista’s Speech Recognition to be nothing short of startling.

    It takes about twenty minutes to set up and train the software, and then you can start dictating directly into your favorite programs.  I created documents in Microsoft Word, e-mails in Outlook, and random notes in OneNote.  If the software made a mistake, it was easy to correct.

    You can also give Windows voice commands. I launched programs, opened the Control Panel, and even browsed the web, all without touching the mouse. (If you’re ever at a loss what to tell Windows, just ask “what can I say?” to bring up a list of commands.) 

    I wore my cast for a total of about five days. During that time I managed to continue working on the PC about as well as usual. To be on the safe side, I added a disclaimer to my e-mail messages, warning recipients I was using speech recognition software so they might see some weird errors.

    Most people told me if I hadn’t mentioned it, they’d never have known. That, I think, is perhaps the best endorsement any speech recognition program could hope for. (It’s not just me: See what Extreme Tech had to say about Windows Speech Recognition.)

    At home, things didn’t go as smoothly. Bored, I decided one afternoon to try playing drums by wedging a drum stick between the cast and my palm. Unlike using Vista’s Speech Recognition, it turns out that was not an especially good idea….

    <Dave Johnson>

     

    TRY IT

    Like to test Windows Speech Recognition for yourself? This short article on Windows Help and How to explains how to get started.


    No Comments
  • Jul
    8

    Recently, I woke up with a throbbing pain in my right hand.

    That was worrisome, but to be honest, I do abuse my hands quite a bit: I type all day at work and spend my spare time playing drums. So I fumbled my way through the workday, yelping in pain each time I was forced to bend a finger.

    But eventually I found myself in the emergency room. When I left, my hand was encased in a cast. What does someone who types for a living do in a situation like this? I grabbed the dusty, neglected headset that came with my PC and fired up Windows Vista Speech Recognition.

    I’ve always been something of a fanatic about speech recognition. As a freelance tech journalist in the late 90s, I’d go to trade shows and hang out like a groupie at the booths of early innovators like Dragon Systems and Lernout & Hauspie.

    Say, what?

    But whenever I tested their latest offerings (DragonDictate and Voice Xpress), I’d soon get frustrated, tossing my mic in the corner until the following year. Why? The programs made too many mistakes, especially ones that were hard to detect when proofreading. One typical glitch was the confusion of homonyms such as “be” and “bee” or “to,” “two,” and “too.” But I’ve got other horror stories.

    I remember one product demo back in 1997. The company president, wearing a headset, was trying to show the audience how he could page through a calendar program with his voice.

    “Next day,” he told the computer.

    Nothing happened.

    Next day.”

    Still nothing.

    Eventually, he tried a different approach—he got louder. “NEXT DAY!”

    Suddenly, the computer sprang into action. It opened a calculator on top of the calendar.

    Clearly, speech was not yet ready for prime time.

    So I wondered, as I caressed my cast, if the speech recognition in Windows Vista was going to remind me of my bad experiences from days gone by. Was it usable? Or would it go as badly for me as it did for the guy who gave the now infamous July 2006 demo of an early beta version?  Would I rue my cast and curse my immobile fingers?

    "Nothing short of startling"

    As it turned out, I did very little ruing.

    I found the accuracy of Vista’s Speech Recognition to be nothing short of startling.

    It takes about twenty minutes to set up and train the software, and then you can start dictating directly into your favorite programs.  I created documents in Microsoft Word, e-mails in Outlook, and random notes in OneNote.  If the software made a mistake, it was easy to correct.

    You can also give Windows voice commands. I launched programs, opened the Control Panel, and even browsed the web, all without touching the mouse. (If you’re ever at a loss what to tell Windows, just ask “what can I say?” to bring up a list of commands.) 

    I wore my cast for a total of about five days. During that time I managed to continue working on the PC about as well as usual. To be on the safe side, I added a disclaimer to my e-mail messages, warning recipients I was using speech recognition software so they might see some weird errors.

    Most people told me if I hadn’t mentioned it, they’d never have known. That, I think, is perhaps the best endorsement any speech recognition program could hope for. (It’s not just me: See what Extreme Tech had to say about Windows Speech Recognition.)

    At home, things didn’t go as smoothly. Bored, I decided one afternoon to try playing drums by wedging a drum stick between the cast and my palm. Unlike using Vista’s Speech Recognition, it turns out that was not an especially good idea….

    <Dave Johnson>

     

    TRY IT

    Like to test Windows Speech Recognition for yourself? This short article on Windows Help and How to explains how to get started.


    No Comments

Categories

Archives