Talkitt demo Talkitt demo In this demonstration of Talkitt, the app allows Ray to say "I love you" to his mother in a clear voice. (Photo: Screenshot courtesy of Voiceitt)

Giving a voice to people with speech disabilities

New app can be taught to speak in a clear version of the user's voice.

In a world where people can text, IM, email and tweet – sometimes all at the same time – verbal communication is still important. For those who have speech difficulties, communication is frustrating in the extreme; in your mind, it's crystal clear, but what comes out of your mouth is completely different. An Israeli startup called Voiceitt is attempting to bridge that gap with a new app called Talkitt.

The goal of the app is to use devices such as tablets and smartphones to translate indistinguishable speech into understandable language and allow users to express themselves easily. According to Voiceitt, about 1.5 percent of the world’s population have speech difficulties as a result of numerous conditions, such as autism, ALS, brain injury, cerebral palsy, Parkinson’s disease, stroke and more, meaning the app could help millions.

Unlike text-to-speech software, such as the speech synthesis used by Stephen Hawking, Talkitt is user-friendly to people who have motor disorders. The app will recognize the user’s voice and can generate clear and natural speech in the original voice of the user, as opposed to a robotic voice, such as that associated with Hawking. The app will enable people with speech difficulties to communicate easily and also make use of voice-activated technology such as Apple's Siri.

Danny Weissberg, co-founder and CEO of Voiceitt, explained to From The Grapevine that the app “translates text to voice and connects text to speech. It helps people who cannot speak clearly communicate with the world in different ways, whether it’s face to face, via phone or social media.”

Voiceitt was founded two years ago by Weissberg and Stas Tiomkin, who holds a doctorate in voice recognition. Weissberg and Tiomkin wanted “to make a change in people’s lives by helping them be able to communicate, be understood and feel connected.” Weissberg said that “most people take communication for granted, but it is one of the hardest things people with speech disabilities deal with. How can we communicate without speech? We can use our hands and our expressions, we can write, but without speech, people often feel disconnected.”

Talkitt works by learning words and recognizing the user’s voice. Before the user can start using the software fluently, however, they need to teach the software a vocabulary, as well as demonstrate how they pronounce their words. The user will say the word and then type it. The software is based on consistency; it will learn the user’s pronunciation of each word and will be able to generate an understandable pronunciation of that and similar words in the user’s voice.

Since the software is built to recognize consistent patterns of speech, it can't help people who have inconsistent speech patterns, such as those who suffer from aphasia, a speech condition usually caused by stroke.

It was designed to be used by young children with speech disabilities, with the help of their parents. “Parents usually understand their children even if their speech isn’t clear,” Weissberg explained. "The idea is that the mother will select a word and type it or show the child a picture, the child will then say the word out loud, and the app will pick up the word and recognize the voice of the child."

"Talkitt opens up the ability to enjoy a broad range of social activities and have a better life." Orit Kelner, Talkitt user with cerebral palsy

The software has several stages of development. In the first stage, Talkitt is still basic: “If the person says a word that was not used during the learning phase, the software will not recognize it,” Weissberg said. In the second stage, “the software will try to guess unknown words and ask the person if they meant this or that. The software will present questions to the person, and this way, the vocabulary will grow over time.” In the third stage, the goal is that the software will become more adaptable and will preserve the user’s voice.

The basic app is in its final testing before being released, but users who've tested the prototype are delighted. Orit Kelner, an Israeli who is living with cerebral palsy, describes her experience of using the app: “The ability to speak and be understood is basic for most people, yet for me it is a daily struggle," she told From The Grapevine. "Talkitt opens up the ability to enjoy a broad range of social activities and have a better life."

It turns out Talkitt can also make people laugh, as in the case of "sit-down" comedian Chris Fonseca, who has cerebral palsy. Fonseca joked to From The Grapevine, “When I order sushi, I get pastrami. Now I have high cholesterol and cerebral palsy, but with Talkitt they can understand me. Thank you, Talkitt.”

To further develop the app, Weissberg said, “Voiceitt is cooperating with numerous hospitals and associations, both in Israel and in the U.S., testing algorithms, working on beta tests and gathering voices.”

Dr. Morit Beeri, CEO of Alyn Pediatric Hospital in Israel, was optimistic about the impact of the app: "We ... believe Talkitt will change the lives of speech-impaired children, empowering them to do things they never could before. We are dedicated to helping bring this solution to the lives of the disabled, enabling them to communicate in the most natural way possible and enjoy social interaction between friends and family."


Photos and SlideshowsPhotos and Slideshows

Related Topics: Apps

Giving a voice to people with speech disabilities
New app can be taught to speak in a clear version of the user's voice.