I discovered the built-in speech recognition functions in Windows 7 last fall. Probably not the best thing for someone who likes playing around with computers! I had used speech recognition software way back in junior high. It left a thing or two to be desired. Fast forward a few years and Windows now comes bundled with built in speech-recognition abilities. While targeted as a accessibility feature, it does allow dictation, and could be useful for the average user.
I’ll post instructions on how to activate it soon. You’ll also need a microphone, preferably the headset type–it’s more accurate. (Something like this Koss Headset or this Logitech USB Headset, which is similar to what I use)
My intent using it was to dictate notes and papers, thus maybe saving myself some time. Did you know you can dictate faster than you can type? Probably no surprise, but still worth pondering if you write a lot. Speech recognition hast the potential to save fingers and time. My thoughts below are based on using it for a couple of months, including work on a major paper.
Positively, using Windows Speech Recognition is a nice change from typing, and is surprisingly accurate. My first encounter with computer speech-to-text was with a copy of Via Voice that came with an electronic encyclopaedia (yes, that was back in the early days before we had wikipedia!). I don’t know if it was the software, or my tender child-like voice, but Via Voice had probably a 25% accuracy rate. Of course, I found that to be great fun in and of itself as it provided many hilarious bloopers. Windows 7 Speech Recognition has something like a 90% accuracy rate.
It was great not having to type. I’m a fairly fast typist, but do miss keys at times, and have to go to correct those errors. It was nice to think and speak my thoughts rather than to have to write them down. It saved a step.
Now the negatives.
First, as an academic writer, I found several issues. While I was surprised at how many words it recognized with relative ease, it still had trouble with some of them. My real problem, however, was really more of a personal drawback. While it is true that you can speak faster than you type, I find that the process of typing is an important lag time for my own thinking. I found myself struggling at times to fill complete sentences, and often going back and starting again. Somehow, typing I am able to more cohesively form my thoughts, and dictation seems to be more of a distraction than a help. This is obviously more of a user issue than a software fault. Still, a point for the good old keyboard.
Furthermore, and really my main problem: while Windows 7 Speech Recognition is actually surprisingly accurate, it still stumbles. And every stumble is something you have to go back and correct, which takes time, and often resulted in me losing track of my thoughts. I make mistakes typing too, but it doesn’t seem to throw my groove off like arguing with a computer over what the word should be. I tested this while writing a paper on history and memory, so the term ‘memory’ came up a lot. Windows 7 seemed to have trouble with that word in particular. It seemed to feel that ‘mammary’ was a better option, even after repeated corrections. Which leads to another problem: missing mistakes. Again, this can happen with a keyboard too, which is why you must must must proofread. Many thanks to one of my proofreaders for catching a few stray words!
If money were no object, I’d love to try Dragon NaturallySpeaking. I’ve read good reviews: most people say it is noticeably better than Windows 7 Speech Recognition. But I don’t think I can justify that expenditure at the moment. I’ll stick happily to my keyboard and perhaps on occasion to Windows 7’s speech recognition software.