Some may find speech a convenient way to dictate their messages, but for those who are busy, the time it takes to train speech recognition technology could be a major inconvenience.
I thought I would check out IBM’s ViaVoice and write this feature using the product. But because the software has a long learning curve, I found my story took me almost three times as long to write as it normally would.
To be fair, I didn’t try the software for very long. I trained it for about 40 minutes – reading it two stories. And I could see it improving a lot as I worked away at my story, but the progress was just too slow for my purpose. I was impressed at its ability to distinguish between words like “their” and “there,” but in every sentence I wrote, there was, without fail, at least one error. Usually more.
There was also an annoying time lapse before the words I said would actually appear on the screen. For now at least, my typing is faster, more accurate and more fluid.
This would probably change if I kept using the software, but I simply don’t have the time it would take for the software to reach an acceptable level of accuracy. I don’t particularly want to while away my nights at the office taking two to three times longer with each story I write.
I also found the software would pick up other people’s voices even though they were several feet away. Words would mysteriously appear on the screen even when I was not talking. The words “we will” would appear most often, and it soon became second-nature, as I trudged along with my feature, to stop at each sentence to correct errors.
It was also difficult writing a story by speaking it. Giving up my keyboard for the microphone was just as tough as it was a few years back when I gave up my pen for the keyboard.
Of course, now that I’m used to the keyboard, I wouldn’t dream of giving it up for my pen, and maybe the same would be true for voice recognition technology. But who’s got the time?