Writing by Talking Instead of Typing

I’ve always had a hidden desire to be good at dictation, thinking if I could dictate coherent thoughts it could speed up my writing. However, after using earlier iterations of voice recognition software, I decided to face the facts and accept that I speak with all the clarity of King George VI.

(What you’re about to read was dictated by me into a machine.)

Through the years, I’ve purchased numerous iterations of dictation software from the company now called Nuance. For some reason, dictating has never worked for me. Perhaps it’s because I was born after the Don Draper era. When I graduated from college, I could type 80 words a minute, which is probably faster than I can think. I can probably type faster than that now, but as anyone who has read this blog knows, the faster I type the more goofy things I say. And one of my rules of blogging is to not work over the text as much as I would if this were, say,  a final edit of something  that was going to be read by more than 12 people.  (I love you, but you’re the only one reading this.)

In other words, a lot of what you see on this blog is  more like a first draft  than any kind of finished writing.  Bottomline: I’ve always been able to type fast, so I never really learned how to dictate.

But I’ve always had a hidden desire to be good at dictation, thinking if I could dictate coherent thoughts it could speed up my writing. However, after  using earlier iterations of voice recognition software, I decided to face the facts and accept that I speak with all the clarity of King George VI.

But recently, something amazing happened. Siri started understanding me. At least in the text messages I dictate, the only time  I’ve even tried to use speech recognition since discovering my inability to talk with machines. As I have heard along the way (whether it is rumor or fact doesn’t matter) that Nuance licenses to Apple some of the technology that goes into Siri, I began to think I should give dictation a another shot.

An opportunity for trying it again came after I had an encounter with a sideview mirror of a car while riding my bike home from work on December 1.

I ended up on the pavement with a few injuries including what, after a couple of months, I learned was a finger with a chipped bone floating around inside it. That led to some hand surgery and a couple of pins inserted in a finger and a few days during which I am unable to type in the way I’ve done nearly all day, every day for the past several decades.

Rather than go with the index finger typing method, I decided to use the six days of having my hand wrapped up in a way that makes it look like a sock puppet to give speech recognition software yet another chance. So once more, I purchased the Mac version of Nuance Dragon Dictate software and am using it to dictate and command my MacBook air.

Observations:

1. Dictation slows down my first draft. Typically, I start writing with a really rough draft in which my primary goal is getting words on the screen. It’s not until I’m through a few rounds of editing that all those words start to turn into something that sounds remotely like writing. (Again, I view blogging more like sharing class notes than publishing a final draft. I give myself permission to mess up when first getting ideas off my chest. ) I haven’t learned yet to be comfortable enough with dictation to let strange looking words stay on the screen for later edit. I still don’t trust dictation to keep up with me. I haven’t yet reached the point where I don’t want to stop and go backwards to correct something.

2 The accuracy of recognized words in Dragon Dictate is far superior today than it was the last time I tried it. Then again, I guess most technology is. Nuance has made several purchases and seems to be consolidating a lot of the expertise in the speech recognition niche of technology. While talking to my computer and having what I say recognized consistently is welcomed, I still haven’t grasped the subtleties of moving among the different “modes” one uses in writing on the screen. In some ways, dictation on the screen is a bit like simple coding. You have to think of the context in which the next block of words will be used, and then think enough like the software to know when to open a mode that tells the software, “hey this is a number!” or “hey this is a command to go check something on Google.”

3. Dictation commands can be counterintuitive. Even after printing out a list of commands so that I can refer to them, I’ve discovered I can’t find the specific command for the function that I’m looking for. And typically, what I need is something as simple as “back up and fix that word.” I quickly discovered that backup means something different than go backwards. In other words, I think it’s fairly easy to learn how to dictate, but learning commands requires a lot of time. I guess I should say, there’s a feature that allows you to create your own commands, but I would probably limit those two highly technical terms or terms you use regularly. I think it would be less productive in the long run to come up with replacements for the global commands.

4. I’ve only used this software for about 48 hours but I now know more about the accessibility features of my computer than I’ve ever known. Knowing how to talk with your computer and how to get it to play back something that appears on the screen as text, has many benefits for all of us, not just those who cannot hear or see. For example, merely listening to text when someone else is reading it helps to uncover indecipherable language and just dumb typos. As I’m sure you can tell from this post, it doesn’t do away with all of them, it at least slows you down and helps you listen to what you have thought you said.