VocaliD is a organization that’s working on a smartphone app, a way of letting people donate their voice to those who can’t speak as well as creating customized synthetic voices for them. It’s perhaps more demanding than regular charitable causes because it demands your time, and not your money.
The way this works is quite simple – donors have to send a few hours of their recorded voice to the developers. The sentences that are to be recorded must cover all the combinations of sounds which occur in the English language and hence, the required information is provided to volunteers.
This is necessary because the donated sentences may be separated into words and even the syllables may be broken down and blended into a new synthetic voice. VocaliD takes certain parts of the recipient’s voice (if they have some control over their vocal source, that is) and combine it with some bits of the donor’s voice, using a software package called ModelTalker.
New Scientist reports that an application which allows people to donate their voice through the help of a mobile device is being tested currently. The alternative would be for charitable folks to go to a studio, record their voice and submit it to VocaliD. Surrogates similar in age and of the same sex as the recipient are chosen to donate their voice.
It apparently takes 800 sentences to stitch together a usable voice and almost 3000 lines for one that sounds somewhat natural to the human ears. The process of recording can take up several hours, even though it does not have to be done in one go. Have more money than time? Contributions are also welcome, says VocaliD.