There’s a great AKAI sample set that comes on 5 or 6 cd’s called Symphony of Voices, it was used in the Sopranos…etc
EDIT* whoops, just read the above 
I completely feel ya man, I like doing everything completely from oscillators when possible. Physical Modeling is the dawgz balls.
There is also Virtual Singer, a plugin for Harmony or Melody Assistant as low budget alternative. Somehow it’s also possible to use your own voice with it.
http://www.myriad-online.com/en/products/virtualsinger.htm
Even if you additionally buy Melody Assistant to use it, it will only cost you 50 USD or 40 EUR. Try for yourself if it’s worth it.
I am filled with shame and guilt when using other’s samples.
No matter how excellent and unique the sample is, I feel like I am adding a cheesy “aaaaare yooooou reaaaaady?” sample…
Thanks. Listened to demos, not convincing enough.
So far it seems like Vocaloid is the only one that fits the criteria for non-sample based singing synthesis.
Although I liked the set of sample based instruments by that eastwest company.
Hey,
Wanted to thank you guys for the information.
At least now I know whats out there.
My conclusions are:
Cantor and Vocaloid have a very similar interface to the point where it is obvious that one copied from the other. Not sure who came first though.
sugar bytes sweet anne might do the job for female vocals? never used it myself but read a review the other month and it sounded alrite…
http://www.kvraudio.com/get/2849.html
thats a link you can click on.
Yes, thanks.
Sweet Ann is in fact a voice from the Vocaloid series, using the second version of the Vocaloid engine.
I know what you mean, I’m not a singer either… but there’s a trick in physics that allows you
to pretend harmonics, if you sing with three, five, six, etc voices. Anyway, I must admit I never
used any melody in the choirs I recorded, rather used its sound as an extra drama-element.
The track’s here: http://www.yomirecords.com/music/botb/botb_-_perelien.mp3
^ wow. 
beautiful drama
As I understand it, synthesing vocals is like the holy grail of synthesis because of the nature of the human voice.
Another option that I don’t think has been mentioned may be Melodyne. Even if you have a bad singing voice you should be able to get it in tune I understand you don’t want to sing, but you could also use Melodyne to further edit the results of any of these ‘virtual vocalists’.
Again, I know you don’t want to use samples, but when direct note access is available in Melodyne, I wonder how good it will be at detecting and stripping out the vocal in a piece of music?
Theoretically you should be able to extract the vocal then use Melodyne to rephrase and rearrange words to fit the music you are working on. If you need words that aren’t in a particular vocal performance you could strip the vocal from multiple songs by the same singer. Then edit them, rearrange them, repitch them and comp together your own vocal performance by your favourite singer. It would be like being able to get them to sing exactly what you wanted them to!
Listened. This is not the type of music I listen to normally, but the choir effect is very nicely added! Liked it a lot.
Sounds about right. You know, they are doing some pretty good job at it (the vocal synthesizers that are available today) - in a few years we may listen computer singers on the radio? who knows
From your description it sounds like one of those Auto Tune (aka tpain effect) softwares? I will check it out.
While on this subject of modifying your own vocal, I am VERY MUCH interested in a software that does vocal modifications, not like auto tune, but one that makes your voice sound deeper, or robotic - with my little understanding in sound, I guess it is dealing with the formants of the sound.
I have a sweet piece of hardware that does the job, but I looked for software like this with no luck (found some, but they were amateurish or unstable)
The Voice Transformer I have:
