About fifteen years ago, my dad purchased the first PC to grace the family home. It was pretty good for its time as it sported a Pentium 120MHz CPU, 16MB RAM and a massive 1GB hard drive. It also came with a pretty hefty price tag.
Shortly after we purchased the computer, we saw a number of speech recognition packages on sale that painted a picture of being able to quickly and easily dictate a slab of text which would then magically appear on your screen. What the box failed to mention was that you had to train the speech recognition engine to become more familiar with your particular voice and if you had a heavy accent your mileage would vary.
My experience at the time was mediocre at best. Our cutting edge computer found it difficult to keep up with the software and I also felt like a bit of a douche trying to dictate an assignment, especially when a lot of them were essays that required copious reviews and amendments.
Over the years, a few of the big names in the computer industry have touted speech recognition as the next big thing. Bill Gates has tried to push it a couple of times but it never seemed to be a runaway success despite Windows having the functionality out of the box. Mobile devices like the iPhone have built-in speech recognition that is limited to making phone calls, playing music and other basic functions.
Are we going to see full blown speech recognition in the mainstream?
I think there are specific settings and environments where speech recognition might work but I think we are yet to see the killer application for such technology. There’s no doubt that it has matured significantly over the years and the corresponding text-to-speech technology has made leaps and bounds over the last decade.
Personally, I believe that speech recognition is best suited to small phrases to aid in completing specific tasks. I don’t think I could ever dictate a whole assignment without feeling stupid or breaking the flow to correct a mistake or make a change.
I would love to be proven otherwise but I still think that speech recognition is still ahead of its time.