voice tracing

Future Therapies

Advanced Speech Encoding

One system, being developed for DARPA by Rick Brown of Worcester Polytechnic Institute in  Massachusetts, relies on a sensor worn around the neck called a tuned electromagnetic resonator collar (TERC). Using sensing techniques developed for magnetic resonance imaging, the collar detects changes in capacitance caused by movement of the vocal cords, and is designed to allow speech to be heard above loud background noise.

DARPA is also pursuing an approach first developed at NASA's Ames lab, which involves placing electrodes called electromyographic sensors on the neck, to detect changes in impedance during speech. A neural network processes the data and identifies the pattern of words. The sensor can even detect subvocal or silent speech. The speech pattern is sent to a computerised voice generator that recreates the speaker's words.

DARPA envisages issuing the technology to soldiers on covert missions, crews in noisy vehicles or divers working underwater. But one day civilians might use a refined version to be heard over the din of a factory or engine room, or a loud bar or party. More importantly, perhaps, the technology would allow people to use phones in places such as train carriages, cinemas or libraries without disturbing others. Brown has produced a TERC prototype, and an electromyographic prototype is expected in 2008.

However, both systems come at a cost. Because the words are produced by a computer, the receiver of the call would hear the speaker talking with an artificial voice. But for some that may be a price worth paying for a little peace and quiet.

NASA scientists have begun to computerize human, silent reading using nerve signals in the throat that control speech. In preliminary experiments, NASA scientists found that small, button-sized sensors, stuck under the chin and on either side of the ‘Adam’s apple,’ could gather nerve signals, send them to a processor and then to a computer program that translates them into words.

"What is analyzed is silent, or sub-auditory, speech, such as when a person silently reads or talks to himself," said Chuck Jorgensen, a scientist whose team is developing silent, subvocal speech recognition at NASA Ames Research Center in California’s Silicon Valley.

To learn more about what is in the patterns of the nerve signals that control vocal chords, muscles and tongue position, NASA Ames scientists are studying the complex nerve signal patterns. "We use an amplifier to strengthen the electrical nerve signals. These are processed to remove noise, and then we process them to see useful parts of the signals to show one word from another," Jorgensen said.

the signals are amplified, computer software ‘reads’ the signals to recognize each word and sound. "We use neural network software to learn and classify the words," Jorgensen said. "It’s recognizing the pattern of a word in the signal."
In their first experiment, scientists ‘trained’ special software to recognize six words and 10 digits that the researchers ‘repeated’ subvocally. Initial word recognition results were an average of 92 percent accurate. The first sub-vocal words the system ‘learned’ were ‘stop,’ ‘go,’ ‘left,’ ‘right,’ ‘alpha’ and ‘omega’ and the digits ‘zero’ through ‘nine.’ Silently speaking these words, scientists conducted simple searches on the Internet by using a number chart that represents the alphabet to control a Web browser program.

"We took the alphabet and put it into a matrix -- like a calendar. We numbered the columns and rows, and we could identify each letter with a pair of single-digit numbers," Jorgensen said. "So we silently spelled out ‘NASA’ and then submitted it to a well-known Web search engine. We electronically numbered the Web pages that came up as search results. We used the numbers again to choose Web pages to examine. This proves we could browse the Web without touching a keyboard," Jorgensen explained.

Jorgensen and his team developed a system that captures and converts nerve signals in the vocal chords into computerized speech. It is hoped that the technology will help those who have lost the ability to speak, as well as improve interface communications for people working in spacesuits and noisy environments.

The work is similar in principle to how cochlear implants work. These implants capture acoustic information for the hearing impaired. In Jorgensen’s experiment the neural signals that tell the vocal chords how to move are intercepted and rerouted. Cochlear implants do it the other way round, by converting acoustic information into neural signals that the brain can process. Both methods capitalize on the fact that neural signals provide a link to the analog environment in which we live.