On Friday, I posted that I was starting to experiment with Naturally Speaking. The results are in: here’s my first post written using NS.
However, this is my first on edited post using NaturallySpeaking. Whatever it types, I’m going to publish. That should give you a good idea of how accurate it is out-of-the-box.
Overall, he did quite well. After getting poor performance on my Mac book Pro, I installed it on my trusty ThinkPad. Right now, I’m dictating into Windows life writer, Microsoft’s data blogging tool. Given that I’ve only spent about 15 minutes training at before jumping into dictation, I’m quite pleased. I can talk faster than I type, so even though I have some correcting to do, the time balances out. It’s fun to stare reflectively at the ceiling while I talk instead of focusing my attention on the monitor as I type. It will be more fun when I can look out the window of my new attic office instead of at the rafters and insulation above my head in the basement.
What I really want to do is what the people on the parallels message ports have been talking about: using CNC to take keyboard input from a Windows session and parallels and feeding it into MacOS applications. For now though I’m happy being able to talk into a microphone and have my text appear, more or less like I intended, onscreen. This is a huge improvement over my past experience, which dates back to using Dragon power secretary to write my second book. Okay, I admit it: parts of my second book. Power secretary was way too painful to use to write a whole book with. Naturally speaking, on the other hand, could certainly be a contender. I’m sure as I train at it will get better. The acid test will be using it with Word 2007 and Outlook 2007, which are my primary applications in Windows. We’ll see how it goes!
