From Keyboard To Microphone

Systems builders on the lookout for new products and services to add to their portfolios should consider a technology that has come of age: speech recognition, or SR for short.

Dragon NaturallySpeaking, from Nuance Communications, is a good package for many SR applications.

SR offers an alternative to keyboard-and-mouse interaction with a PC. SR works by letting systems equipped with a sound input source (such as a microphone) interpret human speech for transcription.

While installing SR software is simple enough, most SR users need a good deal of help transitioning from the keyboard to the microphone. And that's where the business opportunity for systems builders comes in.

With SR software, a user can potentially input text at about 130 words per minutes (wpm) with 90 percent accuracy.

Doctors' and lawyers' offices, which have traditionally used dictation, are good places to start when selling this technology. Only about 2 percent of this market currently uses SR, estimates Howard Parks, founder of Highland Park, Ill.-based Microref Systems, an SR consulting firm. That means the professional space is fertile ground for systems builders.

Another viable market for SR? Helping people with motion impairment--carpal tunnel syndrome, for example.

User support and assistance is key, because without proper coaching, about three-fourths of SR users go back to keyboarding, Parks estimates.

Any systems builder who plans to pitch SR to customers should get comfortable with the technology first by using it on a regular basis for several weeks. By thoroughly learning the software, you'll be able to offer combined installation, configuration and coaching services to those who can benefit from SR. To build an SR system, you'll need the SR software and an adequate PC system. The software comes with a microphone, although a higher-quality microphone is recommended for better results.

Installation

Install the Dragon software from the CD. You will then be asked to don the headset mike and read aloud from one of several canned texts. This training will take about five minutes.

The Dragon software can analyze text files to adapt to the user's writing style. The software will scan all Word, WordPerfect, HTML, RTF and TXT files it finds in the My Documents directory. For users who intend to compose e-mail via voice, the Dragon software can also perform similar scans of e-mail files. Be aware that this process can easily take 30 minutes or so.

If response times are slow when dictating to an outside word-processing program (such as Word), dictating directly to Dragon's own word processor, DragonPad, usually yields faster results.

Coaching Services

Helping users learn to correct recognition errors will enhance recognition accuracy. Learning to use the voice commands to control applications is important, too, although the mouse is always an option for visual controls, such as icons. After getting the user set up, expect at least two more follow-up sessions to help the user master the software.

Lamont Wood has spent more than two decades freelancing in the high-tech field. He lives in San Antonio, Texas.