Yamaha has developed a "singing voice synthesis" program called Vocaloid (http://www.vocaloid.com/en/introduction.html) that can synthesize certain voices now. We've long had most instruments synthesized, but human voice was nigh impossible due to the wide range of sounds you could possibly make. I think Milli Vanilli would have loved this program.
Here's a few of samples of the voice synthesis on you tube:
Vocaloid Miriam - Every breath you take (http://www.youtube.com/watch?v=7SC8Knx8AMI&feature=related)
Vocaloid Miku (Japanese) (http://www.youtube.com/watch?v=abrhLQ3hTAk&feature=related)
Vocaloid Miku - Fly Me to the Moon (English, but with Japanese accent (http://www.youtube.com/watch?v=oGFAdQTGKzc&feature=related)
"Fffffrrrryyyy me to the moooon" Heee. Sorry, but that song being sung with a Japanese accent just cracks me up XD
Really impressed though with the accuracy of the imitation - aside from slight discrepancies (voices are still a bit tinny, notes get cut off early when they should be held a second or two longer) I wouldn't be able to tell it was a machine producing the sounds and not a human.
WIN
https://www.youtube.com/watch?v=kbbA9BhCTko
Wired mag had some article on the voice synthesizer a while back which had several links to various videos.
Kinda interesting :)
http://www.wired.com/culture/lifestyle/multimedia/2008/02/st_jsgw_ss
Ave Maria (http://www.youtube.com/watch?v=h-qDhj9ESao)
As an actor and singer, that makes me want to break something.