OK.
As someone who actively worked with CVVC and had already recorded some banks with this method, I will give my two cents.
CVVC is pretty much an easier way to get a continuous sound. With the release of presamp, you don't have to convert the usts anymore to a CVVC format. It reads CV usts right away in a smooth motion.
The smoothness of CVVC to VCV is the same. You can record your sounds as low as 2moora, up to 8moora (Tatsumi list, voicer of Giga Lacan). With presamp in use, the way the transitions crossfades in between sounds makes them really smooth and natural.
Another pro of the method is that you can keep your bank/tones more consistant, since your voice will get less tired from recording. The time you'll need to record will also be less.
Another pro is that your VB will weight less MB, reducing the space on your computer. For example: Namine Ritsu KIRE has around 288MB unzipped. Tsukishiro Hakupo CVVC, with the same amount of tones (4) weights only 72 MB.
We all see CVVC as "VC must be an ending consonant sound", which is correct for most other lenguages as English, Korean, and Spanish. In the case of JP-CVVC, the "VC" particle is a "connecting one" which smoothes the transition.
I am quite in love with CVVC banks for being versatile, stable, and more interesting to tune.
This is an example of a CVVC rendered with presamp:
[soundcloud]https://soundcloud.com/yue-nagareboshi/utau-splatter-party-e-type[/soundcloud]