LEARNING a tonal language like Chinese is notoriously difficult 鈥� it鈥檚 easy to end up calling your mother a horse. But soon there could be a wearable headset that can help.
The system was created for people with autism who want help with social interactions, but it could be adapted to help with speech or anxiety problems 鈥� or even language learning, says at the that designed it.
Called SayWAT, it gives live feedback via Google Glass to the wearer when they are speaking too loudly or in a flat tone 鈥� things that can lead to autistic people unintentionally appearing bored or abrasive in conversation.
Advertisement
SayWAT uses Glass鈥檚 microphone to record speech, then displays real-time guidance on volume and tone. It shows a volume icon if the user鈥檚 voice is too loud and flashes the word 鈥渇lat鈥� if the user鈥檚 voice does not vary in pitch.
Fourteen autistic adults tested the system. Four used SayWAT to talk with non-autistic volunteers, and 10 used it at an employee training session. The simple feedback seemed to help the users modulate their volume, as they crossed the threshold that triggered feedback less often.
鈥淭he device gives real-time feedback to the wearer if they are speaking too loudly or in a flat tone鈥�
But the pitch feedback didn鈥檛 have the same effect. Boyd says this could be because pitch is more complicated and might need more specific feedback 鈥� something one of the volunteers requested. Her team will present the work at the conference for Human-Computer Interaction in San Jose, California, next week.
The technology could be adapted to work on other devices, says Boyd. A smartphone or watch could give haptic feedback to guide speech, for example.
Live speech feedback could help other groups too. 鈥淚t would certainly be applicable to anyone who had a speech disorder, any sort of anxiety disorder,鈥� says , a computer scientist at the University of Washington.
Real-time feedback when speaking a foreign language is also possible 鈥� but hard, says Boyd. Giving live feedback takes a lot of computational power, so the device could only focus on a few aspects of speech at once. But language learners could be alerted when they mess up a specific sound, like tones in Chinese or the Rs in French, for example.
Some autistic people don鈥檛 wish to change the way they interact, but for others, the technology seems to fill a need. In the corridor after the experiment, one autistic man told Boyd, 鈥淚 really, really want to be social, so if there are tools out there that can help me, I want to try them.鈥�
This article appeared in print under the headline 鈥淭one up your conversation鈥�
