The Dragon vs. The Q: The Battle for Voice-to-Text Supremacy

I was working with a girl yesterday whose speech was affected by early hearing impairment. My job was to find a voice-to-text solution to help her spell and write better. Dragon NaturallySpeaking version 9, which was ordered for her a few years ago, was not successful. SpeakQ was not successful. I had my doubts that anything would work for her because her speech was not very clear and she always spoke quietly and got many of the words wrong during voice training. Then we decided to try Dragon NaturallySpeaking version 10, which I recently got on my computer. We chose the "special" training option, where we don't get feedback during the training process; she would just read a passage for 10 minutes, mistakes, mumbles and all. I was not optimistic.

After voice training, Dragon spent SEVERAL minutes processing her voice sample. (I had to send her back to class because it seemed like it was going to take forever.) But after it was done, we tested it, and we were AMAZED! Very high accuracy, despite all the factors working against us.

I guess I should have had more faith. I use DNS v. 10 without a headset; that's right, I just use my laptop's built-in microphone (and I don't even know where on the laptop it's located). Granted, I myself have pretty clear speech, if I do say so myself, but still, using a built-in microphone for something that usually requires not just any headset, but a high-quality headset? That's impressive. Oh, did I mention, I did ZERO voice training for DNS 10? So, out of the box, using my built-in microphone, I can get an accuracy percentage in the high 90s.

So did the mighty Dragon smoke SpeakQ in my books since version 10 came out? Not exactly. There are still times when SpeakQ is better. The user experience with SpeakQ is so much simpler, making it my first choice for most elementary students and for secondary students with more pronounced reading difficulties. DNS, is, after all marketed towards adults, while SpeakQ was made specifically for special-needs students.

For a more detailed comparison between Dragon and SpeakQ, see my recent post at the following Diigo group: AT in SPED (EDIT: I'm not sure my post is accessible, so I've posted the contents in this blog at http://cybersert.blogspot.com/2010/03/dragon-vs-speakq-part-2.html)

If the makers of SpeakQ and/or Dragon happen to be reading this post, may I make a humble request? Please get together. Please combine the best of each product to make a DNS-powered SpeakQ or a simplified DNS! Our kids deserve it!