I’ve been an artist for more than 8 years having composed and produced music, and am also a Computer Engineer with siginificant research and direct industry experience in Digital Signal Processing; so trust me when I say this.
Other than very high-end software that uses parametric techniques to isolate human voice from the background (which, as a technology is still in its infancy), production engineers use manual techniques, such as phase inversion in stereo recordings, to isolate and delete the vocals (which is the opposite of what you want). Usually the reverse is not really possible for the general case unless it was mixed down in a certain fashion.
So until I come up with an algorithm or technique to synthesize profiled human voice, stay tuned.