there's a cvvc guide in the showcase
here! the guide also links this oto tutorial
here!
as for pitch, the voicebanks you're looking at are multipitch! multipitch banks are specifically designed to capture a larger range since singers have different tones for lower and higher registers! multipitch is a method people use on top of the otoing & alias method and is just a different way of making an utau! you don't need to do multipitch for a cvvc owo/
if you indeed aren't going for multipitch, you don't necessarily need to pick a single pitch, but i find it helps a lot. for example, you don't want to have one sample at D4 and another at B4! that will have too much of a tonal distance from the natural changes your voice makes as it goes higher & lower. so, i'd say it's best to pick either a single pitch to record at
or do a range where your tone doesn't change, usually between only a few semitones. but as long as your tone is the same throughout your samples, it really doesn't matter too much!
there's also a lot of cvvc plugins that exist that can automatically convert cv usts to cvvc for ease of use, but since you asked how the method works, here's the gist:
cvvc stands for consonant vowel, vowel consonant. that means that the beginning of each note is cv, as most japanese syllables are (e.g. ka/fu/shi), but it blends into the next consonant with an extra "note" that we use as the vc (at/ob/es). instead of having the standalone samples we have in cv, we use an extra note to blend so it doesn't sound as choppy!
so, for example, saying "watashi" would be split into 3 notes in cv, [wa] [ta] [shi] and in cvvc that goes instead to 5 notes; ([wa] [a t]) ([ta ash]) [shi] and, depending on the method used, there may be beginning and ending cv/vc's included, like [- wa] if it's at the beginning of a phrase and [i -] at the end. so it could even be 6 notes if you add the end breath!
you should definitely check the guide for how to oto and alias, though; if you want to use converter plugins, they won't pick it up if you do it a way the converter doesn't know.