Operation Guides
CVVC Japanese
Before you begin:
- This guide was written and coordinated by members of the UTAU community and has no affiliation with the UTAU software itself or its creator.
- This guide was written with users of WINDOWS 10 in mind, advice does not immediately translate to other operating systems or other versions of UTAU, such as UTAU-Synth for MacOS.
- This guide relates to the original UTAU software by Ameya which released in 2008 and NOT OpenUTAU, the fanmade UTAU alternative, as such the utility of this resource may vary.
- While the process of operating UTAU is ultimately safe when done correctly, JOEZCafe and other parties involved in JOEZUTAU projects take no responsibility for any incidents, loss or damage to users or property from following these instructions.
Making your CV UST CVVC compatible:
The process of UST-making that is specific to CVVC often occurs after making a UST using the CV method, as such, it's recommended that the CV Japanese Guide is read beforehand to get you up to speed, as a lot of essential steps already covered in that guide will be ommitted here for the purpose of streamlining.
From this point onwards, you should have a CV format UST that's ready to be adapted into CVVC.
Your average CVVC voicebank will be able to read your CV UST just fine, but taking advantage of CVVC's exclusive features will improve your vocals greatly.
For this next step, try setting your Quantize to "L64 64th note"
to resize notes more precisely!
With our CV UST notated, we're now going to insert VCs into the UST to make it CVVC compatible.
To start off, take the tail ends of each CV note in your UST and shorten them to make room for your VCs, if you're unsure how much shorter you should make the note, a good starting point is making the gap as large as the following note's overlap.
Then double click on each rest note to replace it with the Alias of a VC.
As the name suggests, you want to enter the Vowel of the previous note in Roman lettering, then add a space, then enter the Consonant of the following note, also in Roman lettering.
As an example, if the phrase in your sequence is:
ど/do - れ/re - み/mi - ふぁ/fa
Then the same sequence in CVVC becomes:
do - o r - re - e m - mi - i f - fa
And just like that, you have an UTAU UST formatted for CVVC Japanese!
From here, you can use the RESET+P2P3 method to crossfade the CVs and VCs together, smoothing the vocals. If any of the transitions sound off, a good way of tweaking the delivery is to adjust the size of a VC note using the CTRL Dragging method, which affects the level of emphasis and delivery when changing between samples.
Troubleshooting - How Vowels change in CVVC:
Voicebank types like CVVC are very much a vague and rough framework that can be adapted from, making each CVVC bank truly unique to one another, and one big example of this is how different banks handle Vowel samples.
For the most part, how you managed vowels in a CV UST goes relatively unchanged in CVVC, but knowing about these variations can be helpful in reducing the likelihood of problems later.
Depending on the CVVC vocalist you're using, some have two types of Vowel sample represented by two different aliases.
"- a"
If the CVVC bank has romanised vowels that are aliased with a Hyphen (-) at the beginning, these are "Opening Vowels" or "Attack Vowels", these have very little overlap and begin with the vocalist opening their voice, these are for placing at the beginning of vocal phrases if a phrase starts with a vowel sound.
"あ"
CVVC banks with vowel samples of just the Hiragana character on its own can be one of two types of vowel sample.
In some banks, it's an ordinary vowel sample that functions the same as vowels do in CV banks, where it can be used in every vowel note in your vocal phrase.
However, if the CVVC bank ALSO contains "Opening Vowels" (- a), these vowels are "Overlap Vowels" that play a recording PARTWAY THROUGH the vocalist singing that vowel sound with a large overlap value.
These samples are meant to be used in the MIDDLE of a vocal phrase for crossfading, and will sound mistimed if used at the beginning of a phrase.
If you're unsure which vowel system the CVVC bank you're using has, perform the following process:
1. Enter the Alias of an Opening Vowel to see if the voicebank recognises it, depending on the bank, it will either use Roman characters (- a) or Kana (- あ).
2a. If the bank recognises one of those Aliases, the bank has Opening Vowels, any vowels at the BEGINNING of a vocal phrase must use these samples so the sample is timed properly, while any vowels PARTWAY THROUGH the UST can just use the standard alias of the vowel character on its own.
2b. If the bank does not recognise one of those Aliases, it just has standard Vowel Samples similar to CV voicebanks, every vowel in your UST can use these samples and you shouldn't have any problems with timing.
Example: While JYOZE's CV banks and the CV banks in Cafe Series like Claude Ulisse and Saga just have standard Vowel Samples, CVVC banks in Cafe Series like SIX have Opening and Overlap Vowels.
Troubleshooting - VV Samples Vs. Crossfading:
Depending on the CVVC voicebank you use, each one can handle transitioning from one Vowel sound to another in a different way, most CVVC banks use one of two Vowel - Vowel methods that you should familiarise yourself with while using the bank.
Method 1 - VV Samples
Some CVVC banks have specially recorded Vowel To Vowel (VV) recordings of the vocalist transitioning between the two sounds manually.
If a CVVC bank has VV Samples, then Vowel Samples that are partway through the UST must be formatted with a VV Alias, wherein the Vowel of the prior note is entered in Roman lettering, followed by a space, then the Vowel of the note itself in Hiragana.
Method 2 - Crossfading
Some CVVC banks do not have VV samples and handle vowels similarly to CV.
Simply enter the Hiragana into the notes and then run the Crossfade Tool with the U key to crossfade the notes together.
Example: All CVVC banks in Cafe Series do not use VV samples, and instead opt for the Crossfading method so that they can double as CV voicebanks.
