Talking Your Stories


We all talk to our computers, but does yours answer back?  It can if you use some of the voice recognition software now on the market. Not only will your computer talk to you, but it can listen to the words you say, analyze them, and repeat them back to you.  Holy Star Trek, Batman!

Why use voice recognition (VR) software instead of a tape recorder?  The software types the words for you as you speak.  And the end of the day you have your work “on paper”. And, if you want a sound file for some reason, you get that too with voice recognition software. The computer can even read your work back to you using your own voice, just like a tape, or you may have the option to use someone else’s voice. In other words, VR software gives you everything you’d get from a tape but adds much more.

Voice recognition software can be very useful to anyone who is a poor typist or who has carpal tunnel syndrome or other physical limitations that prevent long hours of typing. The advanced versions of the software will even let you run your computer with voice commands.  Imagine working using nothing but your voice to work.

Open Word. File menu, Open, Chapter 12. Go To end of document. New paragraph. Start quote. Don’t open that door comma John exclamation mark. End quote.

Or you can skip all the fussy bits and just dump words onto a page for proper editing later on.

New paragraph. Don’t open that door John-   New paragraph.  Why not-   New paragraph. Because I said so-

And so on.

Voice recognition software works something like this. You read pre-selected text into a microphone plugged into your computer. The VR software “learns” how you pronounce certain words and sounds. Actually, since very few people will pronounce the same word exactly the same way twice in a row, the software is really learning your range of mispronunciation. It’s important to keep the background noise to a minimum while you’re training your computer since any sounds made by loud radios, TVs, screaming kids, and so forth will be interpreted as coming from you.

At some critical level, the software decides that it has enough prefab data to work with and you’re permitted to start with free-range words and phrases. The software still has lots of learning to do early on so you have to speak slowly and as distinctly as you can for a while. As the database grows larger, you can speed up, especially if you don’t throw a lot of new vocabulary choices at the computer.  

When presented with a new word not in its database the software uses a standard dictionary to try to guess what word you mean.  This also includes referencing some rules of syntax and grammar to help the computer guess the correct word based on sentence context. This is not as easy as it sounds. For example, not only are the words red and read pronounced the same, but they can also be used more or less the same way. A sentence like, “Which book was read to the class?”, is clear to humans, but may confuse voice recognition software.

Even with the best voice recognition software there will always be errors.  For example, if you say the name of the Carthaginian general, Hasdrubal, a name the computer is almost certain not to have in its dictionary, it will probably type something like, “has drew ball”.  But that’s OK, because then all you have to do is a find-and-replace procedure with your word processor. Or you can stop, correct the line and add the name to your computer’s dictionary at that moment.

If voice recognition software’s so great, why isn’t everyone using it? Because of an interesting characteristic of writers: not everyone writes the same way. There are professional authors out there who must use pen and paper for at least their first draft. And when I say “must”, I mean that those particular authors cannot write if they try to use a typewriter or computer. Pen and paper is, for them, an integral part of the creative process. Many other authors are still using typewriters for the same reason. Even a computer keyboard is simply too intrusive for some.  

A quick check with Amazon shows over 70 voice recognition software products for sale ranging from a few dollars to hundreds of dollars.  If you’re lucky enough to be able to dictate your work, fine, but be sure to try some samples before you buy any of them. Regardless of how appealing the idea may be, it’s quite possible that the software will be useless to you.

Of the four commons methods of creative writing, dictating to a scribe is the oldest.  All you need is the right hardware and software and you can carry on that tradition.


First published November 2002
Copyright 2002
Fred Askew