Published on July 9, 2007

Hoping to apply technology to assist people with disabilities, the National Electronics and Computer Technology Centre (Nectec) has set a focus on the development of speech-synthesis technology

Since people with speech impairments have problems using their voices when they communicate with other people, speech-synthesis technology will help, said Wantanee Phantachat, programme director of the Assistive Technology Centre at Nectec.

The centre will push for the development of four speech synthesis projects. All projects are intended to give people with a speech disability better voice communication.

The first project is the development of speech-assessment software. This software is designed to help speech therapists evaluate speech capabilities and voice quality, allowing them to know each individual's degree of impairment to provide further training. Developed by researchers at Nectec, the software analyses speech capabilities and voice quality and gives a result to speech therapists.

Nectec has completed the first version, which Wantanee said still relied on a speech therapist's expertise. The team has also set 68 standard words used in the speech-assessment process.

The centre plans to distribute the software to hospitals and other related institutions which offer speech training. Wantanee said the centre also plans to develop the software to be able to make automatic analysis with no need to rely on speech therapists.

The team is in the process of collecting a speech database from children with speech disabilities and others so that when it comes to automatic analysis, the software can make comparison and give an accurate result. It's hoped that the prototype of the new automatic version will come out by the end of this year.

As people with speech disabilities require speech training, and the speech-training software is based on English language, Wantanee said the centre also plans to develop Thai speech-training software to train people to pronounce Thai words properly.

Speech-training software uses speech-recognition technology as a basis for teaching speech. It will also use visual and auditory feedback to analyse and improve the speech skills of those who have speech, language or hearing disorders. To encourage people, especially children, to continue the training, she said the software would be designed like a game so when a user pronounces the sound correctly, they will receive rewards for their efforts.

"Local development of this software will open opportunities for people with speech impairments to get better speech training," she said.

For those who have had their larynx removed, it's really difficult for them to speak naturally. To bring technology to help these people, the centre has worked with a research team from King Mongkut's University of Technology Thonburi to develop what it called a speech enhancer.

"Normally, people with no larynx cannot use the voice completely, so we had an idea to develop a system to analyse the voice performance of these patients and then generate missing voice components," Wantanee said.

The plan is to develop a system to help the speech impaired to generate a complete voice and be able to speak out loud, completely and naturally. The development began earlier this year and it's hoped the first prototype will be complete within two years.

The last project, she said, was the development of speech-recognition technology for voice-impaired children. This project aims to help children to use their voice, even if impaired, to control surrounding appliances.

The technology will allow children to instruct, for example, a wheelchair to move forwards, backwards, left or right, or command their computer system or other electrical appliances to turn on and off, using just their voice.

Wantanee said the centre has worked with a research team from Chulalongkorn University to develop the project and the prototype should come out by the end of next year.

Pongpen Sutharoj

The Nation