Wouldn t it be nice if you could just think what you wanted to say without actually saying it and have your thoughts picked up and converted into spoken words? The possibilities for this technology are practically endless. Imagine engaging in a conference call with your colleagues where no one actually speaks out loud, yet each participant hears and takes part in the conversation. Of course, there are a couple of downsides. What if your wife and kids could hear what you are thinking during one of those heated family discussions?
Well along those lines, scientists at the NASA Ames Research Center in California have developed a system that captures signals resulting when a person speaks inaudible words, processes the signals, and sends them to a computer that translates them into audible speech. Small, button-sized sensors attached to the chin and on either side of the Adam's apple detect subauditory signals such as when a person talks so softly as to be unheard by others. Using subvocal speech recognition software, the nerve signals can be translated into speech. The principal behind the patent-pending technology is that the tongue and vocal cords are influenced by speech signals from the brain. These signals can be detected even though the person does not utter words out loud.
An initial experiment conducted by the scientists trained the software to recognize several words and digits, achieving a 92% recognition level. Repeatedly the researchers softly spoke the words stop, go, left, right, alpha, and omega and the digits zero to nine to teach the software to recognize the words and numbers. Taking the experiment one step further, the scientists, by arranging the alphabet into a matrix with each letter identified by two single-digit numbers, were able to search the Web without using a computer keyboard.
A difficult part of the project is analyzing and quantifying the complex nerve signals that control the tongue and vocal cords. Because the signals are low level and noisy, they must be amplified and filtered to extract the usable information. Further processing then is necessary to accurately differentiate one word from another. The researchers plan to continue expanding the learned library to include more English sounds.
What I like about this device is its applicability to everyday phone conversations. There is nothing more annoying than to be in a public place surrounded by people talking on their cell phones. Invariably, these people assume they must talk rather loudly to be heard by the person on the other end of the conversation. Using these devices particularly in airport gate areas could make the wait much more relaxing. Now, if we can find some way to silence those pesky gate announcements that are made all too frequently by airline personnel.
Paul Milo
Editor
[email protected]
All contents • 2004 Nelson Publishing Inc.
No reprint, distribution, or reuse in any medium is permitted
without the express written consent of the publisher.