Perhaps it’s the coolness factor, or perhaps I’ve just watched one too many sci-fi movies . . . whatever the reason, I’ve always been interested in the idea of talking to the computer, giving it commands, and having the computer execute functions based on voice input.
Over the years I’ve experimented with Microsoft’s built-in speech recognition software a number of times. I used it occasionally under Windows XP, and after moving to Windows Vista I decided to give the latest version a try. I worked through some of Vista’s initial speech training, but I have not yet completed any of the supplementary exercises that are supposed to improve speech recognition accuracy. Even without these additional exercises my initial impression is that Vista’s speech recognition is much better than its XP predecessor. I certainly found the text editing commands to be superior to those of the previous version.
Nevertheless speech recognition brings with it a certain level of frustration. There are times when the words appearing in my document bear absolutely no resemblance to my utterances. Sometimes the software seems to just get stuck. A voice command that has successfully worked several times previously simply stops working, and no amount of repetition (in any tone of voice) will encourage the software to start accepting my commands again. Perhaps it’s just tired of listening to me.
Beyond just wanting to experiment with the software to see how far it has progressed, a student recently asked if the library had voice recognition software. Under traditional library mentality it’s difficult to imagine a student sitting in a public area speaking aloud to a computer. Under the current model of information commons areas and group collaboration spaces, the idea is not so far-fetched. Somehow though, I imagine this student quickly becoming an annoyance to other patrons – not from speaking aloud, but rather from the endless repetition of some word or command that the computer stubbornly refuses to recognize.
In trying to create this post and the previous one with speech recognition software, Windows slyly pretended that it didn’t recognize the word “and.” Instead it pretended that it was the word “end,” and responded by pressing the End key when I simply wanted to type the word “and.” There were other pages and glitches along the way, and at times it was easier to simply type corrections than to talk Windows through them. Ultimately, I can still type a document faster than Windows and I can complete it through speech recognition. Needless to say some additional training is still required – for both me and the computer. The coolness factor is still there, so perhaps between my improved “broadcaster” voice and a few more training sessions the software and I will arrive at a happy medium.