I freely admit that I tend to be attracted to gadgets, especially computer gadgets. For instance, how many people do you know who bought one of those disk punches that supposedly turns double-density disks into high-density disks? Well, now at least one ...
But for a more than a year, I had my eye on perhaps the ultimate computer gadget, something right out of Star Trek, 2001: A Space Odyssey and Tron. You know what I’m talking about.
And now, so does my Macintosh.
I first saw the Voice Navigator II in action one-and-a-half years ago at Macworld Expo in San Francisco. As I watched a suit talk into his computer — and was flabbergasted when it listened and responded — my jaw dropped and drool began to flow. I knew I must have one.
As my co-workers know, I tend to talk a good game about the latest technology, but usually can’t follow through financially. But Voice Navigator II was too special to put aside. I had to chomp down on the financial bullet. Really cool technology tends to cost a pretty good chunk of change, but I knew Uncle Sam was going to return a healthy refund last spring and began to plan. Within 24 hours of that Treasury Department check arriving in my mailbox, I had all of it — and more — spent and Voice Navigator II hooked to my SE/30.
When you spend nearly $650 on a contraption, you expect it to work. But I wasn’t sure if it would really work.
Let’s just say Articulate Systems Inc. has hit high on the astonishment scale. The dang thing works just as advertised.
How does it work?
Voice Navigator II is a small computer system interface (SCSI) device, so it hooks either directly into the back of your Macintosh or anywhere along a daisy chain. It works on at least a Mac Plus with 2MB RAM and System 6.0 or newer.
A few of its basic features include:
• Unlimited number of voice commands
• Real-time recognition performance
• Recognition of any accent and any language
• Speaker-dependent training
• A somewhat limited macro system
• An extremely flat learning curve
At first, Voice Navigator II seems like an amazing bit of technology — and it is — but as I learned more and more, I realized that this hardware/software combination was less like HAL of 2001 and more like QuicKeys (from CE Software). Like a keystroke combination, Voice Navigator II simply is a macro system that uses sounds instead of keyboards or pointing devices. Looking at it that way takes away some of the mystery, though none of the amazement.
My boss, for instance, is British and despite living in the United States for years he has retained a strong accent. Voice Navigator II could care less because it doesn’t listen to the words; it listens to the sounds. As an example, I can train Voice Navigator II to undo my last action every time I say, “Oh no!” Or it might send a document to the printer each time I say, “Get out of here.” Upon hooking up Voice Navigator II, the various body noises and expletives that could be used as verbal commands came to mind rather quickly and caused my thoughts to race.
Getting started
Hooking up Voice Navigator II is very simple, depending upon how lucky you are with SCSI devices. I tend to be notoriously poor at trying to successfully make a daisy chain work on the first time, but I got lucky with Voice Navigator II and once I found the power switch (it was hidden on the top front of the contraption) I began to roll. Voice Navigator’s hardware consists of the previously mentioned SCSI device and a microphone. A unidirectional desktop microphone accompanies Voice Navigator II and handles home-computing tasks very well. Articulate Systems also sells headset microphones that work much more accurately in noisier office situations. The hardware installation guide is very detailed and setup took less than 15 minutes.
Installing the software was even easier. It meant copying the contents of four floppy disks to a hard drive, dropping four files into the System folder and rebooting.
Training Voice Navigator II
The first key to success with a voice-controlled computer is the software languages. Voice Navigator II comes with a number of languages. These have nothing to do with French, Spanish, Fortran or Pascal. Instead, Voice Navigator II languages are predefined, customizable sets of commands that can be issued by voice from within given applications. For instance, Voice Navigator II comes with two disks full of languages ready to go for such applications as Aldus PageMaker, Adobe Illustrator, Microsoft Word, QuarkXPress, HyperCard and the Finder for both systems 6 and 7. If a favorite application is missing, such as America Online, a small utility called Language Maker can quickly create one.
The second key to success is Voice Navigator’s voice files. These files hold the specific voice commands and add a great deal of flexibility in an office environment. Since everybody’s voice sounds different, everybody needs a separate voice file. My boss is an example. With his accent, I’d never be able to successfully issue commands with his voice file loaded., at least not with any useful consistency. If several people use a Macintosh at work, a voice file could be loaded for each. A good macro program such as QuicKeys would allow users to hit hot keys to quickly switch to their voice files.
These voice files are surprisingly small, too. I have loaded commands for about a dozen applications into my voice file, yet it only takes up about 150 kilobytes of disk space. Part of this is because certain commands are used by nearly every Macintosh application. For instance, WriteNow, PageMaker, Lotus 1-2-3 and ResEdit all use undo, cut, copy, paste, print, quit and many others. Once Voice Navigator II knows how you say commands for one application, it remembers them as they apply in other programs.
To train Voice Navigator II, we first must tell it which language to use, which is done with a small system extension called Voice Control. This lets us set our preferences, voice confidence levels, languages and voice files. After we have loaded, say, the System 7 Finder language, we move to another application called Voice Commands. If Voice Control is the brains behind Voice Navigator II, Voice Commands is its larynx. By clicking on the Train button, Voice Commands will present us with all of the available commands for the Finder application. We are asked to speak the commands three times so Voice Commands can get a read on our voice patterns. Surprisingly, each application takes very little time to train because of the number of repeat commands.
I learned an important lesson during my first voice training session: The desktop microphone is very sensitive to other noises besides my voice. I generally like to listen to music while working at the computer and Bob Dylan’s singing irritated Voice Navigator II at least as much as he irritates most humans. I also had to oil my chair because Voice Navigator II interprets some squeaks as commands. I’m not sure if that means I sound Bob Dylan.
The really cool stuff
Within about two hours, I had trained a half-dozen application languages that came with Voice Navigator II and created and trained three others from scratch.
But I was only beginning to tap into Voice Navigator II’s power, which lies in the previously mentioned Language Maker desk accessory. As noted, Language Maker can create application languages that don’t come with Voice Navigator II. It also can create voice commands to:
• access any type of menu item
• select any tool, pattern or color from an application’s palettes
• control buttons, check boxes and pop-up menus that appear in dialogue and alert boxes
• control a window’s size, zoom or close box or to click in other parts of the window, including the scroll bar
• trigger existing macros (such as CE Software’s powerful QuicKeys), enter text strings (such as your name or other phrases), press keystroke combinations or select multiple menu, dialogue or palette items
• move the mouse pointer.
Language Maker also can create the ability to launch any application with a voice command. For instance, when I say, “McMUG,” TeleFinder launches so I can dial into the McMUG bulletin board.
Quite by accident, I discovered I could play sounds with voice commands, too. It was getting quite late when I stumbled upon this ability and things got a bit goofy. I found some sounds from 2001: A Space Odyssey and now when I say, “How are you, Hal?” my Mac responds (in Hal’s voice from the movie): “I’m completely operational and all my circuits are functioning perfectly. Would you like to play a game of chess?” And when I say, “Open the pod bay doors, Hal!” my little Mac retorts: “I’m sorry, Dave. I’m afraid I can’t do that.”
It’s sick, I know, but it had to be done.
Language Maker’s macro capabilities are admirable, but don’t match the power QuicKeys offers. Fortunately, Voice Navigator II’s macros can be used to launch other macros and this makes for a potent combination. If you plan to purchase Voice Navigator II or any other voice-recognition system sure to appear in the near future, be sure to buy QuicKeys, too, if you aren’t already tapping into it.
Technical support
The folks at Articulate Systems aren’t dumb. They know they have one of the potentially hottest products on the market and are ready to nurture their swelling user base. Therefore, Articulate Systems not only offers lifetime technical support for free (no annual fees, no 900 numbers), but also does it with a toll-free number (a technique that has worked wonders for WordPerfect).
Articulate Systems also has a frequently used bulletin board on America Online, which holds a wealth of tips and tricks from other users, as well as additional software utilities and languages.
Conclusions
Voice recognition is one of computerdom’s Holy Grails and Articulate Systems has done it right. I wasn’t sure what to expect when I ordered Voice Navigator II, but I undoubtedly got more than $650 worth of very cool hardware and software that until now was only available in the science fiction section of the used bookstore.
Expect voice recognition to become a big part of our lives, not only with computer use but with everyday household and workplace objects such as telephones, televisions, VCRs, microwave ovens and climate controls. If the technology is here, the applications can’t be far behind.
About the time I purchased my Voice Navigator II, Articulate Systems introduced Voice Navigator SW, a software-based voice-recognition application that provides voice control for Macintoshes that come equipped with microphones.
So if you own such Macs as the Classic II, LC, LCII, IIsi, Quadra 700, 900 or 950 or the higher-end PowerBooks, you can have voice recognition for about $350.
Articulate Systems Inc.
99 Erie St.
Cambridge, MA 02139
(617) 876-5236
Reprinted from The Finder, monthly newsletter of the Mid-Columbia Macintosh Users Group. Non-profit computer users groups have permission to reprint this article providing the author and the Mid-Columbia Macintosh Users Group in Tri-Cities, Wash., are fully credited and a copy of the newsletter in which the article appears is sent to:
