There are plenty of people on this planet for whom speaking is impossible. Injury, illness and other misfortunes have robbed them of their voice. Yes, there are electronic/mechanical aids that can help, but you end up sounding like Stephen Hawking.
Now an American company called VocaliD would like to help. They’re looking for “surrogate vocal donors” to sample for a new library of synthetic voices. Their technology will allow patients to say any sentence in any voice from this library–even without the donor having to record a million different phrases.
What VocaliD does is extract the properties that make your voice sound like you, a process that takes two or three hours in a studio. Software then blends that voice into “speech units” which then can be combined by the “target talker” (i.e. the person who can’t speak) to create a satisfying and realistic synthetic speaking voice.
Cool, no? Read more here.