Bless this newbie with your wisdom!

Rakgnarok

Teto's Territory
Hi! Haha, sorry for the silly title, basically I'm pretty new to utauloid making and using and Im wanting to make an english voicebank. I've heard that english VB are harder than japanese ones but considering I know squat diddly about the japanese language for now I really have no choice.
Anyway so, I've done some investigating and (I think) I know how to use audacity but I'm really unsure about a reclist, a friend sent me a reclist but its honestly very confusing, it looks like
4f2df21547.png
and I honestly don't know how to even pronounce some of those.. things... So, basically I'm just asking you all for any and all advice you can provide me about what type of reclist to use and such and, well, if im honest I don't really know that well what "CV" or "VCCV" are exactly, I mean, I know it stands for stuff like "Consonant, vowel" or something like that but thats pretty much it.

So, yeah any advice you can provide would be greatly appreciated by a noob like me.
Thank you for your time and I hope you have a great day!
 

WinterdrivE

Ritsu's Renegades
Defender of Defoko
CV and VCV are pretty specific to Japanese and are moot otherwise. CV (Consonant Vowel) means all the recordings consist of only 1 consonant and one vowel, eg, ka, se, ru, etc. Good for Japanese where all syllables are only 1 consonant and 1 vowel, not so much for English since English can have upwards of 9 consonants in one syllable.

The problem with CV is that it only records the consonant to vowel transition, leaving it often sounding choppy. eg, in the word "aka" I have the initial, a, the a to k, and then the k to a. CV only gets you the first a and the ka. VCV alleviates this by recording all the syllables of Japanese with all 5 vowels before it. So instead of just "ka" you have "a ka" "i ka" "u ka" "e ka" and "o ka." Whereas "sayonara" is written "sa" "yo" "na" "ra" in CV, its "- sa" "a yo" "o na" "a ra" in VCV. This also means its literally 5 times as much recording as CV. Its also moot for English since again, English allows more than one consonant in a row and it also has, like, 4 times as many vowels as Japanese.

For English, you basically have two choices: Painted Cz's VCCV (which is what your friend sent you) and Arpasing. VCCV is more or less the standard if only due to it being the most well-established English method. Arpasing is a recent phenomenon. In a nutshell, though, the primary differences are a) how its written: CVVC uses Cz's own phonetic system whereas Arpasing is based off of Arpabet, and b) the amount of oto entries per recording: VCCV is 2-4 oto entries per recording but is 1066 recordings, while Arpasing at its most basic is 120 recordings with ~10 oto entries each. (In layman's terms, that means VCCV is shorter recordings but more of them, and Arpasing is longer/more complex recordings, but fewer overall)

Cz has a video going over how to pronounce the phonemes in the VCCV list here (along with other guides on their channel):

And the arpasing website is here: http://arpasing.tubs.moe/en/index.html
 
Last edited:

heta-tan

Genderless Goon
Global Mod
Supporter
Defender of Defoko

Rakgnarok

Teto's Territory
Thread starter
Ah, thank you both a lot! This helps make more sense of it though, I think I'd rather play it safe for now and simply ask what you guys think would be the best for me to start out with.
 

Neku

Teto's Territory
Ah, thank you both a lot! This helps make more sense of it though, I think I'd rather play it safe for now and simply ask what you guys think would be the best for me to start out with.

Hi!

For starters I'd usually recommend CV voicebanks since they're easy to configure & are not too time consuming. Also they give you a notion of how the whole program works. VCV voicebanks are far more realistic & natural, but take more time and are a pain to oto.

If you wanna be safe and put out a voicebank, record a Japanese CV because it's the most "basic" thing and where everyone starts. If you feel adventurous, try to make a Japanese VCV, but I really wouldn't recommend it just yet :sad:
 

Rakgnarok

Teto's Territory
Thread starter
Ah, well thank you for the advice, I'll definetly take that into mind when I make a japanese vb though for now I'd like to figure out an english vb first since thats a language I can understand.
 

Kiyoteru

UtaForum power user
Supporter
Defender of Defoko
If you would like someone to help guide you through the entire process of making an UTAU you can contact the tutors from St. Defoko's School of UTAU. We have a discord server here: https://discord.gg/rSzZD9P
But contact information for all of the tutors is in this resource post. http://utaforum.net/resources/st-defokos-school-of-utau-1-to-1-tutoring.291/

I am among those tutors. I would like to agree with Neku that, if you are completely new to UTAU as a whole, you should start with a basic Japanese CV voicebank. Although it's a language you don't understand, the process of recording and configuring the samples is MUCH simpler than English (a nightmare!), and it's an excellent introduction to how UTAU works.
 
Similar threads

Similar threads