Larkspur (larkspurlazuli) wrote,

Day 19 of 40

Today, I've been investigating one of the windows 7 cutting edge features: speech recognition. I had always thought of speech recognition as a rather primitive kind of software that was still error prone. Was I wrong! I've practiced using it to dictate text today, and I found that even though it had received a minimum of training on my voice, it still correctly picked up some quite unusual vocabulary words, including a couple of proper nouns (but it still did not pick up derren brown on the first try! :( ) I'm using it right now to dictate this entry.

Besides the fact that speech recognition software could be very useful to any non handicapped users who want to use their hands while simultaneously using the computer (can anyone say knitting?), it makes me feel like I'm on Star Trek! Not only can users dictate any kind of writing, but speech recognition could also allowed them to do things like turn pages on a PDF while using their hands for other things like cooking, knitting, embroidering, etc. other positive side effect that I've seen is that dictating my writing makes me think very carefully about the words that I choose. Written text generally uses much more formal language than spoken communication. I had never dictated before, and it's been a useful exercise to put the same amount of thought into my speech that I usually put into my writing. I consider myself a relatively articulate person, but I also think it is a psychologically more difficult task to create clear and eloquent sentences in on the fly speech vs. in a text that is concrete, and can be mulled over and edited.

However, I have a couple of privacy concerns relating to this software. To improve speech recognition accuracy, the software scans the contents of your hard drive, I suppose, to learn the words and topics that the user is interested in. And, of course there is the fact that anything dictated is being recorded by the speech recognition software, and there's no real assurance that this software is not sending everything it dictates to Microsoft.

Overall I have been quite impressed with the accuracy and utility of this tool. However, occasionally it will translate background noise into words, which can be quite problematic when dictating or doing non dictation activities. Secondly although the tutorial claims that it is relatively easy to navigate the web using speech recognition, I found it nearly impossible to navigate webpage is using speech recognition.

I'm also quite lucky to be able to use this technology, because I live alone, in a quiet environment. I don't have to worry about anyone listening to me when I use it!
  • Post a new comment


    default userpic
    When you submit the form an invisible reCAPTCHA check will be performed.
    You must follow the Privacy Policy and Google Terms of use.