Technology Products World™
ScanSoft's Dragon NaturallySpeaking speech recognition software enables voice-to-text applications
Peabody, Massachusetts
October 25, 2004
By Marc Strassman
This page and its contents are copyright © 2004 by Etopia Media News Networks. All rights in all media reserved.
Reporter
Technology Products World
Etopia Media Technology News Network
Etopia Media News Networks
Dragon NaturallySpeaking Preferred 7 package
Dragon NaturallySpeaking is a powerful tool for converting speech into text. You can visit the Dragon NaturallySpeaking homepage by clicking here. You can get its product datasheets by clicking here.
For Technology Products World's first coverage of exciting new technology products, we asked some questions of ScanSoft, which makes Dragon NaturallySpeaking. Here's what we found out:
o What is the history of Dragon NaturallySpeaking?
Dragon NaturallySpeaking has been the leading software for PC dictation since the mid 1990s. It was originally developed by Dragon Systems, which was bought by L&H. When L&H went into bankruptcy, the rights to Dragon NaturallySpeaking were bought by ScanSoft. V7 was the first release of Dragon NaturallySpeaking under ScanSoft managements.
o How did ScanSoft come up with the idea for the technology?
The idea was developed by Jim and Janet Baker, the founders of Dragon Systems.
o Who is ScanSoft trying to reach with Dragon NaturallySpeaking?
Home users who want to have more fun with PCs, people with disabilities who can’t type, physicians and other professionals in dictation intensive industries.
o How do people typically use Dragon NaturallySpeaking?
They dictate into a headset and have Dragon NaturallySpeaking automatically transcribe documents, e-mails and reports. Dragon NaturallySpeaking can also be used to navigate through applications, surf the web, and fill out forms, all by voice.
o What are the basic principles according to which DNS is able to convert sound waves into text?
It’s based on probabilities—the probability that an utterance captured by a spectrograph equals a specific phoneme, and then the probability that a string of phonemes equals a specific word, etc.
o How does DNS "learn" to improve its transcribing ability with a particular user?
DNS adapts to a user’s speaking habits to adjust the probabilities above, making sure there is a better alignment between what the computer “heard” and what the user “said”.
o Are there other real-world venues besides the office where DNS is being used or could be used?
Dragon NaturallySpeaking is also being used in medical, home, disability, and legal settings.