Dictated but Not Read: A Writer’s Perspective on Voice Recognition Software
22 Dec
About two weeks ago I bought a copy of MacSpeech Dictate. The idea was to find a tool that would help me be productive while also saving wear and tear on my delicate, sugary forearms.
So far so good. I’ve been able to keep up with my deadlines. I’m back on Facebook. Pretty soon I’ll be able to be up on this blog with more frequency.
I’ve also gotten a lot of interest from writers about how the software works, its accuracy, and in general what it is like to write with your mouth.
And so here is my take on using voice recognition software. With the exception of a few minor tweaks, I am letting the work stand.
How does it work?
It’s pretty simple. You install some software. You “train” the program so it recognizes your distinct speaking style (this takes about 10 minutes) and then you’re ready to go.
Included in the price is a headset with a specially designed microphone. So you pop that baby on and commence filling the air with your words.
You can speak about as fast as you normally speak, although you have to tell the program when you’re using punctuation, which can slow you down a little bit. But after a while you get pretty good at putting in the commas and periods and then you don’t notice that you’re even doing it.
If you make a mistake, then all you have to do is say “scratch that.” If you’ve just called for a piece of punctuation then the program will remove only that. Otherwise it will pick a word or phrase and delete that. More on how this affects your writing later.
Is it accurate?
Yes. For the most part it gets what you say right. It can handle homophones (full disclosure: it did take me a few takes so that word didn’t come out “humble phones”) although when it gets something wrong it can be frustrating and your best bet is to put in a placeholder word and move on.
For example, I would have preferred to write “sweet, delicate forearms” in my opening line, but I had to go with “sugary” because the program got the context wrong. It kept writing the word “suite” as an address. Still, the fact that the software takes context into account is a big plus.
If anything most of the problems have been my fault. One of the humbling aspects of using voice recognition software is that it lays naked your bad diction.
For example, I didn’t realize that I pronounce the word “already” like the man’s name Artie. As in “I Artie did that blog post.”
(I was also called out by a friend when I promised in an e-mail to “deftly” get in touch with her after the holidays.)
But this is easy enough to fix by over enunciating some words. When I need to say the word “disgust” I put more of a Z. sound into that first syllable, so it doesn’t come out “discussed.”
Does it feel weird?
At first I felt like a parody of the slick, self-involved artist who talks his thoughts into a micro recorder. But once I started enjoying using this new tool I got over any feelings of self-consciousness.
What’s nice about the software is that once you get into a groove you find that you can produce a lot of words in one sitting.
You can also talk your way through an idea and it captures your entire thought process. You are not always as aware of what you are writing as you are when you do it by hand, but I like that quality, especially when I’m in the discovery phase.
But that lack of awareness also brings us to the most important question:
That’s nice, but is it writing?
I’m going to say that it’s not writing, but that doesn’t mean the voice recognition is not an extremely useful tool for writers.
The reason I say it’s not writing is that the nature of the software rewards you for plowing forward. There are all kinds of commands for editing text but so far for me they’ve proved to be too cumbersome. Even saying “scratch that” a million times to undo the work can become a pain. Often it’s just much easier to spew and then fix it later.
I’m also going to say it’s not writing because it’s so easy. You just don’t agonize over the words, the structure, the internal rhythms, etc. of your work the way you do when you’re writing by hand or typing. It all just kind of comes out and that’s okay.
I’ve also noticed that after four hours of dictating I’m not at all tired. It reminds me a little bit of when I first started writing and everything that came out of my pen seemed like pure gold to me. And the reason was that while I was generating copy I wasn’t purposefully working the material.
Same goes for this software. I’ve had a 5000-word morning using MacSpeech Dictate, but when I went back and looked at those 5000 words I ended up cutting a lot more than I would have cut from 5000 typed words.
But given how I work that’s not a problem. I’ve always liked getting it all out in front of me and then working with it later. So this tool suits me well. But I definitely wouldn’t recommend it to the writer who agonizes their work one line at a time. I think you’d get too caught up in the mechanics of the program and you’d lose any benefits from being able to say it to the page.
One thing I do know is this: I’m working and my arms don’t hurt. So I will continue to use the software and perhaps even dig deeper into its many features.
Take care and have a great holiday break. I’ll be back up on the blog in the new year.
