14 DAY TRIAL // A large number of people who cannot talk have ended up in their current situation on account of accidents and infections, which have affected their larynx and vocal folds. Now, researchers in South Africa are working on a new medical device, which could see many of these individuals regaining their ability to speak, also without the raspy voice that current aiding instruments produce. The new system, its creators say, works by tracking contacts between the tongue and the palate. This allows it to understand what word is being said, and then synthesize it into sound, Technology Review reports.
The find could come as a relief to the over 10,000 patients who get laryngeal cancer each year in the United States as well, and who then lose their ability to speak, following surgery. In most cases, these people need to have their voice boxes removed altogether, so as to prevent the disease from spreading throughout the body, and causing death. “All of the currently available devices produce such bad sound – it either sounds robotic or has a gruff speaking voice. We felt the technology was there for an artificial synthesized voice solution,” PhD candidate Megan Russell, from the Johannesburg, South Africa-based University of the Witwatersrand, says.
The main instrument behind the new system is a special type of palatometer, a device similar to the one used in speech therapy. The new variety, which is produced by Orem, Utah-based CompleteSpeech, features more than 115 embedded touch sensors, which are able to detect all motions of the palate, and of the larynx, and then infer the words a person is trying to say from a number of predefined algorithms and patterns, the team explains. The scientists will be providing a full explanation of their finds this week, at the International Conference on Biomedical and Pharmaceutical Engineering, held in Singapore.
Using the device is very simple. The patient just needs to place the palatometer inside their mouth, and then attempt to speak as they normally would. The instrument then attempts to identify which words are being mouthed, before reproducing them via a small speech synthesizer, which needs to be carried at all times too. The team says that the synthesizer is small, and that it could be easily tucked inside a shirt pocket, for example.
Russell says that, about 94.14 percent of the time, the instrument identifies the correct word. This statistic does not include instances in which the device skips unknown words, which happens 18 percent of the time. The expert explains that interpreting the wrong word could lead to some “very difficult social situations,” so the patients are better off this way.