'Sound' Synthesis

I was thinking of an approach to do synthesis of sound - 'synthesis of voice' to be precise.

I have made few basic assumptions upon which the sound synthesis approach will be based. They are -

• speech is not continuous but it is made up of discreet elements.
• each individual has his own 'library of sound elements' from which s/he assembles speech.
• a library has one atrribute - number of elements (size).
• each element also has one attribute - the sound (note) which we hear.
• the libraries of any two individuals are of same size and contain the same types of notes.
• two analogous notes from any two different libraries differ only slightly so as to accomodate the fact that no two voices are same.
  

More to come ...