NaturallySpeaking 11 Promises More Natural Speaking
By Jim Bray
Not all Dragons are fierce creatures, mythical flying beasts of legend. Nuance's Dragon, in fact, is an excellent tool for taking your – or someone else's – spoken words and turning them into printed words or actual deeds.
And with version 11 of its Dragon NaturallySpeaking voice recognition software, Dragon-maker Nuance have given us what's easily the best I've tried.
Not everyone needs voice recognition technology, of course, but it's becoming increasingly ubiquitous in our lives – as many lucky folk who've bought new cars over the past few years can attest. You can use it to make phone calls or operate the vehicle's climate control and audio systems, and even beg it to show you the way home if you're lost and have a voice-activated navigation system.
Dragon NaturallySpeaking isn't for cars, though you could probably use its smartphone app to send text messages while you're driving – a frightening prospect. NaturallySpeaking's home is the computer, Windows or (now) Mac, and you can use it to operate the computer's interface, dictate a letter or email, surf the web, and much more.
It works very well, not that I was surprised. I've been using Dragon since I broke my hand a few years back and had a couple of months when I couldn't touch type – an awful thing for a writer (though perhaps a great thing for his audience!). Dragon NaturallySpeaking bailed me out big time, then, helping me get my work of avoiding the dreaded Pulitzer done.
Watching the transcription process is quite something. Your words appear on the screen as if by magic and, while it can be disconcerting to see the transcription lag as you speak, the software works better if you give it larger blocks of words to chew on. This means that, instead of blurting out a few words and waiting for the software to catch up (at which point you've forgotten what you wanted to say), you can ignore the transcription in process and just prattle on as you would normally – or nearly normally, anyway.
It's also a neat feeling of power to tell your computer to do something – like "Open Word Document" or many other commands – and have it do your bidding. It's more obedient than my kids ever were and probably the closest I'll ever come to being a dictator.
Here's a quick overview of some things Nuance says the software can do:
Version 11 trains a lot more quickly than it did before, too, down to only about four minutes, which is really nice. I remember one version where it seemed like I had to read in "War and Peace" to train it.
I'm not sure how relevant the training is to how I use Dragon NaturallySpeaking the most: transcribing voice files created when I interview people. Those files originate either on my digital voice recorder or are recorded via Skype's "Pamela" add-on. It's a way for me to get an interview transcript without having the type the whole darn thing in again – and for this use, NaturallySpeaking 11 is easily the best version yet.
And what a challenge that must be for the product! Think about it. It's being asked to recognize voices it's never been trained for – complete strangers – and the recordings I force it to work on are often made under really lousy conditions, with plenty of background noise. Yet despite asking the Dragon to work with one claw tied behind its virtual back, the interviews I've transcribed with Version 11 have been up to about 70% successful.
That includes an interview I did with a guy on a cellphone while he walked through an airport, the sound quality of which was absolutely appalling. It, "Naturally," wasn't up to the standards of a good quality file's transcription, but it still beat the pants off of having to transcribe the whole thing myself.
I can't honestly say if the newest Dragon performs such transcriptions 10 or 50 per cent better the last version, but the improvement is noticeable, substantial and welcome. The better Dragon gets, the less time I have to spend editing its transcripts and that means more time spent, well, partying.
"Smaug" does seem to have a bit of an issue with the formats of digital files I've given it, insisting on Mono, lower bit rates than I'd been using. It was easy enough to change the settings on my voice recorder and Skype, however, and I've noticed no loss in the quality of the transcription.
Another welcome fact is that Dragon will now transcribe digital media files in the background while you work on something else. It wouldn't do that in a previous version I tried, which meant I had to go have a coffee or play a videogame or something while the Dragon was having its way with my PC. Now I can keep working on other stuff.
If you want to automate the process, the Professional version of Dragon NaturallySpeaking comes with a kind of a companion application that monitors whatever folder you point it at and automatically transcribes the voice files it finds there. It works in the background and you can keep it running all the time if you want to, though I prefer shutting it down when I'm not using it.
Alas, I was sent the "Preferred" version of NaturallySpeaking 11 for this review and it doesn't have the agent. When I discovered that, my life started flashing before my eyes and I very nearly curled up into a little ball and whined softly to myself in desperation. Fortunately, that's when I noticed that Dragon transcribes in the background now and I've been very happy exploiting that ever since.
Apparently, the "Professional" edition is on the way, so I'll do an update a bit down the road on how the new version of the transcription agent works. Ooh, don't I lead an exciting life!
I like the new Dragon Sidebar, which puts popular features, commands and options right in front of your face where you'll trip over them.
You can switch easily between open Windows merely by telling the Dragon the number of the window you want, and you can launch programs on demand or scroll up or down pages of text inside Microsoft Office programs.You can correct transcription errors as you go, if you want, but I prefer pressing on to the end and fixing the few errors Dragon makes later. It doesn't interrupt my train of "thought" that way.
I wrote the "dragon's share" of this column using NaturallySpeaking 11, speaking into a microphone on my desk – all while a talk radio station and my wife's dog were blaring in the background. And Dragon never confused my voice (such as it is) with the other sounds; not one single word other than my mad utterances appeared on the screen as I dictated.
And while it may not have gotten the 99 per cent accuracy the manufacturer claims, it probably gave me close to 95 per cent accuracy, even with my SIRIUS satellite radio blaring, and that's nothing at which to sneeze.
The results I've obtained using Dragon NaturallySpeaking 11 so far have knocked my socks off – so much so that I'm tempted to do all my columns by voice; it would not only help me get my work done without having to worry about finding the "home row," it would also be excellent practice for my radio appearances, where I'm expected to string together entire sentences without tripping over my tongue.
Not going to happen, though. It would force me to keep my thoughts focused and I'm not sure that's even possible. But that isn't Dragon NaturallySpeaking's fault!
Copyright 2011 Jim Bray
Jim Bray's columns are available through the TechnoFile Syndicate.
We welcome your comments!