How do you add multiple pitches to a CV voicebank?

Ivy!

Ruko's Ruffians
Defender of Defoko
Is it possible to do and have both kana and romaji aliasing? I want to have two for my Utau but I'm not sure how that would work in terms of aliasing and picking which one gets used. I've seen it done in VCV but I'm not sure about CV. Could someone explain it to me?
(This might sound vague, if it does I'll try to explain in more detail but for now that's my question)
 

Kiyoteru

UtaForum power user
Supporter
Defender of Defoko
To distinguish between pitches, all you need is suffixes in the OTO. For the sake of organization, people often separate them into separate subfolders, but you could theoretically just dump all the audio in a single folder.
Let's say that you decide to use a standard CV reclist, and record into multiple folders. Here's what that would look like.
Code:
voicebank (folder)
    readme.txt
    character.txt
    low (folder)
        oto.ini
        a.wav
        ba.wav
        ... etc ...
        zu.wav
    high (folder)
        oto.ini
        a.wav
        ba.wav
        ... etc ...
        zu.wav
When you load the whole voicebank in UTAU and open the OTO, you'll be able to see all of the OTOs at once, but each folder will still have their own OTO.ini file. (This only applies to Windows. On Mac, using UTAU-Synth, there's only one oto_ini.txt for the whole voicebank which has entries for every folder.) Let's look at the OTO for the low pitch only.
Code:
a.wav=aL,0,0,0,0,0
a.wav=あL,0,0,0,0,0
... etc ...
zu.wav=zuL,0,0,0,0,0
zu.wav=ずL,0,0,0,0,0
The high pitch would look much the same, but with an H instead of an L. Suffixes don't need to be H and L, they could easily be the actual pitch being recorded (such as G3) or really anything you want. The important part is to distinguish between the pitches using unique suffixes. However, I would not recommend using special characters like kanji or unicode arrows. These can easily be mojibaked, so stick to things that are easy to type.

Every wav file needs to be duplicated in order to have both romaji and hiragana. The easiest way to do this is to OTO the bank entirely one way, then duplicate all the lines to add aliases for the other way, so that you don't need to OTO it twice. Alternatively, you can avoid having to make duplicate lines in the OTO by naming the files with suffixes in the first place. So if your file is "zuL.wav" you only need to have "ずL" in the OTO.

For UTAU to recognize which samples to use for which notes automatically, you need to set up a prefix map. This can be accessed from Tools > Edit prefix.map. Select all of the keys you want to use for a certain pitch, enter the suffix, and press "Set".

An example of a multipitch CV Japanese voicebank is KYE. You may want to use it for reference.
https://utaforum.net/showcase/kye.574/
 

Ivy!

Ruko's Ruffians
Defender of Defoko
Thread starter
To distinguish between pitches, all you need is suffixes in the OTO. For the sake of organization, people often separate them into separate subfolders, but you could theoretically just dump all the audio in a single folder.
Let's say that you decide to use a standard CV reclist, and record into multiple folders. Here's what that would look like.
Code:
voicebank (folder)
    readme.txt
    character.txt
    low (folder)
        oto.ini
        a.wav
        ba.wav
        ... etc ...
        zu.wav
    high (folder)
        oto.ini
        a.wav
        ba.wav
        ... etc ...
        zu.wav
When you load the whole voicebank in UTAU and open the OTO, you'll be able to see all of the OTOs at once, but each folder will still have their own OTO.ini file. (This only applies to Windows. On Mac, using UTAU-Synth, there's only one oto_ini.txt for the whole voicebank which has entries for every folder.) Let's look at the OTO for the low pitch only.
Code:
a.wav=aL,0,0,0,0,0
a.wav=あL,0,0,0,0,0
... etc ...
zu.wav=zuL,0,0,0,0,0
zu.wav=ずL,0,0,0,0,0
The high pitch would look much the same, but with an H instead of an L. Suffixes don't need to be H and L, they could easily be the actual pitch being recorded (such as G3) or really anything you want. The important part is to distinguish between the pitches using unique suffixes. However, I would not recommend using special characters like kanji or unicode arrows. These can easily be mojibaked, so stick to things that are easy to type.

Every wav file needs to be duplicated in order to have both romaji and hiragana. The easiest way to do this is to OTO the bank entirely one way, then duplicate all the lines to add aliases for the other way, so that you don't need to OTO it twice. Alternatively, you can avoid having to make duplicate lines in the OTO by naming the files with suffixes in the first place. So if your file is "zuL.wav" you only need to have "ずL" in the OTO.

For UTAU to recognize which samples to use for which notes automatically, you need to set up a prefix map. This can be accessed from Tools > Edit prefix.map. Select all of the keys you want to use for a certain pitch, enter the suffix, and press "Set".

An example of a multipitch CV Japanese voicebank is KYE. You may want to use it for reference.
https://utaforum.net/showcase/kye.574/
I'm definitely going to look at Kye's voicebank

How do you duplicate all the lines though? Is there an automatic way to do it? I'm also not too sure about how to use the prefix editor, would you be able to explain that a bit more?
 

Kiyoteru

UtaForum power user
Supporter
Defender of Defoko
How do you duplicate all the lines though? Is there an automatic way to do it?
Within UTAU's OTO editor itself, there's a button for duplicating the currently selected line.
Screen Shot 2018-10-07 at 1.05.03 PM.png
Alternatively, you can open the oto.ini file as plain text in something like Notepad, then select everything and copy+paste.
I'm also not too sure about how to use the prefix editor, would you be able to explain that a bit more?
Here's what the prefix map editor looks like.
upload_2018-10-7_13-6-58.png
To select multiple rows at once, click the first row, hold shift, and click the last row. Let's say I want to add an "H" suffix to all the notes from G4 to B7. First I select all the rows...
upload_2018-10-7_13-8-34.png
Enter the suffix...
upload_2018-10-7_13-8-54.png
And click "Set".
upload_2018-10-7_13-9-10.png
(You can also select other rows and set other suffixes.)
Once you click the "OK" button at the bottom, the entire prefix map is saved.
Whenever you have a note in a UST that's G4 or above, UTAU will automatically add the "H" suffix to the lyric without you needing to type it manually. This is how it will automatically use the high pitch.
 

Ivy!

Ruko's Ruffians
Defender of Defoko
Thread starter
I think I get it now. Will it still work if only one pitch has the suffixes but the other doesn't if it's only two? Like for example if only the lower one has the suffixes and it's entered into the prefix editor leaving the higher pitches with nothing entered, will it still work?
 

Kiyoteru

UtaForum power user
Supporter
Defender of Defoko
Yes, that'll also work!
While you're using the final voicebank, if you want to use a note without a suffix in a range that would normally have a suffix, put a question mark (?) at the beginning of the lyric.
 

Similar threads