Researchers from the University of Edinburgh are helping to restore speech to people who have lost their voice through conditions such as Motor Neurone Disease (MND).
They are using the latest technology to make synthetic voices for patients which have been generated by “voice donors” across Scotland.
The goal is to make communication aids for those who lose their voice through illness – but instead of a “robotic” Stephen Hawking sound, they want to get as close to the patient’s own voice as possible.
The project is a team effort, with clinical researchers from the Euan MacDonald Centre for Motor Neurone Disease Research working with speech scientists at the University of Edinburgh’s Centre for Speech Technology Research.
The inspiration for the project came from Euan MacDonald, who at 29 was diagnosed with MND. He approached researchers at the University of Edinburgh because he knew he was at risk of losing his ability to speak and he wanted his children to grow up knowing the sound of their father’s voice.
Patients record just 20 minutes of their own voice. Researchers then blend this with donor recordings to produce a high quality, synthetic version of the patient’s own voice. Vocal characteristics such as, gender, age and accent can be incorporated into the digital voice to reproduce personality and tone.
The team has already banked donor voices from across the breadth of Scotland and will soon begin collecting accents from the rest of the UK.