Utter Command: impressed first impressions

Kim Patch has asked me to take a look at the aptly-named Utter Command by Redstart Systems, a voice command add-on for Dragon NaturallySpeaking that comes with commands already built in. I wouldn’t be able to use it, because it only works with the Professional versions of Dragon 8 through 10, not Preferred. However, I strongly recommend visiting their website regardless, because it’s well done and just excellent technical reading by itself if you’re an aspiring self-taught geek like me. Nice and low-key and well argued — almost exhaustive, in fact. And who knows — if you have a Pro version of Dragon and sufficient funds and want ready-made commands, you might be persuaded.

Particularly eye-opening is its chart comparing the number of steps for an action by input device: keyboard, mouse, Dragon alone, and Dragon with Utter Command. You can see charts for such common tasks as moving the mouse or browsing the Internet. (Internet Explorer is the default for their chart, but Utter Command has a demo for Firefox too.) Right there is the reason for any add-on command program whatsoever — seeing exactly how many utterances can add up, fast, especially with programs that Nuance only partially supports or with unsupported, heavily Java or Flash applications. (*singsong* “Oh Mouse Grid…”) Seriously, it looks like Utter Command could allow you to scrap the Mouse Grid altogether if you wanted. Solitaire, anyone?

You can also download a tutorial directly from the website, so you can get a feel for it before you decide to buy it. This is a particularly good thing because some of the commands aren’t structured in the way that you’re used to — these are known as human machine grammars, rather than natural language commands. The various slides and videos explain this. I hadn’t thought about it, but their explanation of why certain syntaxes can be inconvenient for commands makes a certain amount of sense as far as improving accuracy. Really, Redstart needed to be eavesdropping when Nuance introduced the maddening new “Cut <word>” shortcut in Dragon 10, which I have happily disabled with the service pack. I can now freely “Cap <words>” again without words being axed, but that service pack wasn’t a moment too soon. “<Word> cut” would’ve made a lot more sense there, Nuance. Thankfully, though, you can still use natural language structure if you want for a lot of things. I don’t think I’d want to give up natural language entirely.

If you don’t feel like reading, they have video tutorials for almost every aspect of the program, which speak for themselves. You can also find these on YouTube. The website also supplies you with a list of common commands. I won’t go into all of them — most of them are standard macros. However, depending on your needs, a few look particularly intriguing.

1. The “3 minutes break” command will apparently turn off your microphone, wait 3 minutes, and then turn the microphone back on. This command is valuable for those of us who can’t always press keys to turn the microphone back on. You could always put the microphone to sleep and then wake it back up, but the microphone is then still listening for the “wake-up” command and, depending on your microphone sensitivity, this command can be triggered accidentally. I don’t know if the 3 minutes can be increased to, say, half an hour or more, but if so, this could be quite useful.

2. The “word screen down return” command says it can let you scroll the page of one document, but keep your cursor in the other.

3. You can also change media player tracks even if your focus is in another program. (Yes, you can listen to music while dictating.)

4. I swear I saw something that implied you could turn the microphone on by voice.

Also well done is their Community section, covering everything from sharing user commands to troubleshooting. This could be invaluable for Utter Command users — what SpeechWiki was to Vocola before it disappeared. Nice move.

From what I’ve read, the makers of Utter Command have themselves very well put together. Their website is a huge point in their favor, including explanations of human machine grammar, research papers, and item by item comparisons. Furthermore, I like that they included videos and brochures directly on the website, so you could actually see how it works before committing to a purchase. (Nuance’s flashy little pitch of Dragon 10 still irks me for not explicitly listing its native limitations with Firefox; it borders on misleading. I don’t care if the respective programmers haven’t figured it out yet — well, I do, but it’s understandable — but at least tell us that we still need Mouseless Browsing if we want to do pretty much anything reliably.) This program is out of my league for several reasons, but I respect a product that’s tempered with communicated research and explanation, not just hyperbole or “we’re better than so-and-so.” I appreciate their time and effort and wish them well.