I have separated out the speech signal into vocal tract spectrum and excitation spectrum. Now i want the original speech signal (ie., the input voice) to be reproduced as output. Ideas please..