VOICE PAD
An Introduction To Voice Recognition Technology
by
Gene Murray
Assistive Technology Specialist
Technology Resources for Education (TRE)
This article is about "writing" on your computer without using a keyboard. Instead of using my fingers to type, or click on icons, letters or words generated by an on-screen keyboard or word prediction program, I am talking to my computer through a software program called Kurzweil VoicePad. VoicePad recognizes my spoken words and translates them into written words in a word processing program. I've been telling people for years now that speech recognition technology, the ability to talk into your computer instead of typing into it, is coming. Not quite yet, I said. They're still working out the kinks and bringing the price to a reasonable level.
Well, it's here. Or at least the first generation is here. Okay, if you want to get technical, it's been here for a few years. Dragon Systems has been offering their "Dragon Dictate" for several years, and continually improving the quality while reducing the price. VoicePad doesn't do as much as DragonDictate. For example, it won't run with other programs or allow you to work with data bases or spreadsheets. Kurzweil has narrowed the focus, and developed a voice recognition word processor that is easy to set up, easy to learn and use, and doesn't eat up your technology budget for the next two years.
Voice Pad works with IBM compatible computers. It has a 20,000 word vocabulary, of which 3,000 can be programmed and customized to meet the users needs. It is very user-friendly, has a high recognition ratio, and requires very little training. Standard computer commands (open, save, print, quit...) can be spoken, and text or numbers entered via a microphone input. Mis-recognized words can be corrected through either a menu of word choices or by spelling the desired word using the International Alphabet (Alpha, Bravo, Charlie, Delta, Echo, etc...). During all of these operations, the mouse and keyboard are still operational. Computer requirements for this technological wizardry are pretty standard; 8mb of Ram, a 16 bit sound card, and a VGA monitor. It will run efficiently under either the Windows 3.1 or Windows '95 operating systems. The people at Kurzweil recommend a pentium if you are using Windows '95. A compatible headset microphone is part of the VoicePad package.
And, here is the most impressive number of all. 79.95. No, that's not the square root of the IQ needed to install and use this program. It's the price. Yep, seventy nine dollars and ninety five cents. Well, okay, you need to add tax, shipping and handling, but you get the idea.
Now, you're probably wondering about setup. Do you actually need an IQ with four digits and 48 hours of serenity to get it up and running? No, although that would be nice. It installed in 10 minutes from a CD Rom, and recognized about 70% of what I said with minimal training.
Is this the perfect input method or what? Well, no, it's not perfect. For one thing, you need to speak slowly and pause between each word, which, can, become, a, little, tedious. But as I said, we're talking first generation here. They're already working on a "next generation" program that will recognize continuous speech, (and probably go where no one has gone before in a much bigger spaceship). There are now several products out there capable of what is called "continuous speech" or "natural language" recognition. Both Dragon Systems, the folks who brought you DragonDictate, and IBM, the folks who brought you "bad command or file name," have new continuous speech recognition programs on the market. Kurzweil is due to come out with one sometime in the near future. These programs are, however, more expensive, and require more training as well as more computer fire power. VoicePad fills the role of an entry level program, an introduction to the brave new world of voice recognition software.
What VoicePad does, basically, is create a voice file that takes into account all of that motor speech stuff, like volume, rate, pitch and resonance, and then continuously compares that file with what you are saying into the microphone. The more you use it, the more raw data it has to compare, and the better it recognizes your unique vocal and speech patterns. A lot of variability, like the kind of speech produced by most dysarthric speakers, doesn't allow the program to "do, do that voodoo that it do so well", so it is less efficient with people with speech impairments. It will also have a hard time recognizing your voice when you have a cold or laryngitis, or the quality of your voice is somehow altered. Athletes foot probably won't affect it much, unless you are one of those people who frequently puts his foot in his mouth.
The value of all of this is that it bypasses the keyboard/mouse combination as the sole method of getting information into a computer. Kids, or adults, with severe learning disabilities, difficulties with upper extremity function, visual motor dysfunction, or fine motor problems, are often frustrated by a keyboard. Their ability to produce written work, so important in classrooms and offices alike, is often short-circuited by their inability to type. Voice recognition can provide a better way.
So, are you working with any kids with a lot to say but no way to get it down on paper? Or, do you consider yourself "keyboard impaired"? doaz yuur tyypng fre2wn;ty look lilke thisss? The Kurzweil VoicePad is a cost effective, user friendly way to beat those keyboard blues. For more information regarding this product, please contact the manufacturer:
Kurzweil Applied Intelligence
411 Waverley Oaks Rd., Suite 330
Waltham, MA 02154
(617) 893- 5151
Fax#: (617) 893-6525
E-mail: erikak@Kurzweil.com