Training a new piece of software is a strange idea. Other computer programs don’t need to be trained. When you get a new word processor, it doesn’t have to watch you type for awhile before it catches on. New spreadsheets do their adding and subtracting perfectly well straight out of the box, without any instruction from you.

So why does NaturallySpeaking need to be trained before it understands your speech? The simple answer is that speech recognition is probably one of the hardest things your computer does. Humans may not think speech recognition is hard, but that’s because they are good at it.

NaturallySpeaking comes out of the box not knowing anything about you. It has to work as well for a baritone with a Scottish accent as for a mezzo-soprano with a slight lisp. It needs time to figure out how you talk.

How long? If everything goes smoothly, as it probably will if you have compatible hardware and follow the installation instructions, something like 40 minutes will probably pass between the time you take the shrink-wrap off the box and the time you start dictating. Allocate 10 minutes or so to read to your computer so it can analyze your voice and pronunciation. Nuance has shortened this training and has succeeded admirably!

Training continues for as long as you keep using NaturallySpeaking. It makes mistakes, you correct them, and it learns. That’s the process. It gets better and better the more frequently you use it.

The exact error rate depends on many factors: how fast your computer is, how much memory it has, how good your microphone is, how quiet the environment is, how well you speak, what sound card your computer has, and so on.

NaturallySpeaking is 99 percent accurate out of the box and it gets better as long as you keep correcting it. Don’t be lazy.