Need help understanding how to record JP cvc

[Atcho]

Hello ! I'm new here, and I have recently relapsed into the Utau fandom, and especially into the voicebank creation side.

I do have experience with recording and otoing cv voicebanks, but I would like to go for a cvc voicebank for my current project. I found a reclist (this one) but the more I look at it, the less I understand how I am supposed to "voice" it, if that makes any sense, and I hope some people here could help me figure things out.

I remember that I would record my cv samples to be around 2 seconds long, with the vowels being longer (ex: I would pronounce "tsu" tsuuu, and not tsssu). Should I also do the same while recording a vcv ?

There's also some confusing parts about the reclist itself :



Some of the lines go like that a_a_e_... Should I pronounce both a ? Should I prounouce the line like aae ? I think I saw someone saying something about the first sound of each lines being for the alphabetical order in Oremo but I'm not sure about it so I'd rather ask if that's the case or not.
(I also noticed some very short lines beginning with n and having only one sound afterwards (ex: n_i, n_gya n_nyo,..). Should I ignore the n at the beginning ?)

And some line have the letter n in them (ex: n_n_a) and that somewhat confuse me. In the line e_e_o_e_n_e, n is both before and after an e. Does that mean that I should pronounce e_n like en (as in energy) and n_e like ne (as in nepal), or should I pronounce each sounds separately ?



The first tsu sound is the only one in the reclist that is "all alone". I am supposed to pronounce it too or not ?

And that's about all the question I have (and that I could think of) ! I'd really appreciate it if anyone could help me with all of that. Thanks in andvance ! ^^

 

[Halo]

Good questions! Just to clarify, this is VCV (vowel-consonant-vowel).

I remember that I would record my cv samples to be around 2 seconds long, with the vowels being longer (ex: I would pronounce "tsu" tsuuu, and not tsssu). Should I also do the same while recording a vcv ?

Similarly! Usually because there are multiple syllables in a VCV recording, you record to a tempo/BPM or GuideBGM music piece rather than in seconds. There are many reasons but for now it suffices to say that it's just a bit easier to keep track of.

I think I saw someone saying something about the first sound of each lines being for the alphabetical order in Oremo but I'm not sure about it so I'd rather ask if that's the case or not.

That is incorrect. They are for the transition between the same vowel, imo you should try your best to pronounce both. Sometimes it'll just sound like the extended vowel almost, and that's fine too.

(I also noticed some very short lines beginning with n and having only one sound afterwards (ex: n_i, n_gya n_nyo,..). Should I ignore the n at the beginning ?)

They're for missing transitions, in this case the n vowel to less common sounds. I don't think recording these as 1-syllable sounds is efficient personally but as long as it has the sounds it's fine. But, no, you do need to pronounce the n at the beginning-- n for one beat, then the CV for one beat.

And some line have the letter n in them (ex: n_n_a) and that somewhat confuse me. In the line e_e_o_e_n_e, n is both before and after an e. Does that mean that I should pronounce e_n like en (as in energy) and n_e like ne (as in nepal), or should I pronounce each sounds separately ?

I think you should probably listen to an example from another VCV bank. It's how it looks! just the vowels smoothly one after the other. It maybe makes more sense looking at it in hiragana-- ええおえんえ, not "en ne", not "e ne", just eeehh nnnnn eeehh

And for the tsu sounds, again this feels annoying to me personally but yeah you just need the "- tsu" sample (tsu at the start of a phrase), so you need to record tsu by itself as well. Good luck!

 

[Atcho]

Thanks a ton, that helps me a lot ! And yeah I misstyped vcv in the title oops-