#utau tutorial

Oh that could be cool! And it kind of depends. It's difficult at first, but once you get the hang of it I think it's fairly easy. It's a long process though.

The first thing you'll need to do is to extract the Wizard101 WAD files. I use QuickBMS for this. http://aluigi.altervista.org/quickbms.htm (Scroll down to find the Wizard101 KIWad file)

To make sure you're getting the right file you can go to this link: http://versionec.us.wizard101.com/WizPatcher/REVISION/LatestBuild/Data/GameData/Root.wad

This is what the person who sent me this told me to do: Swap REVISION with the version from the Credits, like V_r696969.Wizard_1_69. Then change the Root.wad to the respective WAD you are looking to download the full version of. (In your case it would be the dialogue file for whatever world Grandfather Spider is from.)

Next, after you extract the WAD and find all of the dialogue you're going to want to go through and make a folder for the UTAU bank.

Using Audacity, cut out the syllables you want to use. (A good CV bank reclist is here: https://www.deviantart.com/grayoblivion/art/UTAU-CV-Japanese-Reclist-381067704

Save the syllables as .WAV files. You can also splice, by taking the consonant for one and attaching the vowel from another syllable to it and cutting tiny bits off until it sounds right.

Once you do that, you'll open it in UTAU, go to tools - "voice bank settings" and then "launch editor".

This link will show you how to OTO your voicebank! It can be difficult but with some practice you'll get the hang of it.

Also one thing I'd recommend doing is selecting all the phonemes and generating the frequency files. Then go through and make sure that the top left number isn't zero - if it is zero just copy and paste the number from any of them and then click the top middle button. It should give you a message like "x.wav (something in Japanese)."

Both test have identical tuning but the first one I didn’t vocal shift the second I tried to achieve the effect of it.

I changed the g flag lower or higher depending on how stressed the note was. 

Unstressed notes: higher. 

Stressed notes: lower. 

By “stressed I mean more emphasis, not length. A long note with too much g flag would sound weird

I feel like it could be better but the subtle difference dose changed the tone the singing.

the only ones ive seen are like “copy an oto from someone else” and uh i havent seen one that actually teaches you how to oto by hand? so heres one

NOTE: this is based on my very limited knowledge of vcv if i get somethin wrong Blease dont kill me

so! you wanna oto? heres how

i’m using a 7 mora reclist for this but it still applies for other mora lists!

lets take somethin from my wip voicebank

lets use the highlighted file for our tutorial

so youre gonna launch your editor

and cut out the empty space in the beginning like in cv

but at the end! instead of just the blank space at the end we’re gonna go all the way to the first syllable, in this case “u”

(the pictures sideways but thats ok)

edit the pink stuff and the lines how you would in cv (? would you?)

add your alias, which in this case would be “- う”. since thats the first syllable, theres nothin before it, and because of that theres a dash in the beginning

congratulations! you have one note done

now duplicate it

now we’re gonna oto another note.

this time itll be “u う” (i’m not doing that because i just realized i already had that)

oto the thing so it plays the end of the first u and the entirety of the second one

alias that too

congrats! you have another one done.

duplicate that too

now we’re doing “u うい”. fun fact: the beginning of a note is NEVER a regular cv syllable. we’re not doing “wi wo” next. we’re doing “i wo”. remember that!!

do the same thing as with “u う”

repeat until you get to the Very End.

the last note! in this case its “i R”. we’re gonna. chop off everything but the end. we’re gonna keep the fading out this time. its,, Breathe

alias that as “i R”

there youve got it! now do that with every single file youve recorded

A tutorial for installing ENUNU (a plugin that lets NNSVS AI voicebanks work with UTAU) into OpenUTAU.

Hello everyone! I’ve wanted more resource/tutorial posts on here for a while, and this is one of the most common problems I see among new UTAU users! Unfortunately, there are dozens of reasons why UTAU might not be making sound on your device. Note that everything I’m saying applies to UTAU for PC and not necessarily UTAU-Synth for Mac. I don’t own a Mac, so I don’t have much advice there. Here is a basic checklist/set of tips to get UTAU to work.

Installing UTAU

If you’re on Windows, installing UTAU in the Documents folder (or somewhere else) is better than installing it in Program Files. Installing it in Program Files is known to cause issues like lacking sound.

Setting your locale to Japanese

There isn’t really any way around it-on non-Japanese systems, not having Japanese locale installed will cause issues with using UTAU. That goes double for downloading any fan created voicebanks. Here is a basic tutorial for changing locale on Windows. 

Many people don’t want to change locale because they think it will change the language of everything on their computer. That isn’t true! Changing locale and changing interface language are two different things. What locale actually does in the context of UTAU is give your computer the ability to understand Japanese text. It doesn’t have the ability by default on most non-Japanese computers. Unless you’re using UTAU purely for other languages like English, French, etc, changing locale is a top priority and a large reason why it might not be making sound.

Installing and selecting voicebanks

“Installing” UTAU voicebanks incorrectly is another big reason for there not being sound. Voicebanks should be one folder deep inside the “Voice” directory/folder in order to be selected-not in a subfolder.

For multipitch voicebanks, make sure to select the main folder and not one of the subfolders.

Here is a properly installed UTAU voicebank. The icon is in the top left, and the name of the voicebank is in the Project Property menu. Clicking the three dots button next to the white bar that says “voice/(vb name)” will let you manually select a folder to use, but it takes less time to just place the voicebank in the directory.

Pre-made USTs

Pre-made USTs are a common source of sound not working right. Make sure the UST and the voicebank are in the same format. If the UST is in romaji, convert it to hiragana. If a UST is CV, convert it to VCV format and it should work. Iroiro is a plugin that can do both of those things. Some timing issues might still be present, but it will at least make sound.

Converting USTs takes more time by hand, but here are some Japanese voicebank types and how they usually look. Remember that other languages follow different rules!

CV

VCV

CVVC

Resamplers

Resamplers are what gets UTAU from a bunch of recorded samples to a synthesized result. Some UTAUs don’t work with certain resamplers. Going to Project(in the top bar)->Project Property->Clicking “Initialize tools” should reset the resampler to its default state. If you still don’t get any sound coming from UTAU after this, the problem is probably something else.

Someone asked for this a while back, so I decided to deliver! Many flags aren’t commonly used or well known, but all of them can be useful (or you can just mess around with them).

A lot of this information is from tady159′s guide, but I revised some of it. I’ll also add the flags for various other  resamplers. Most of them work off of these flags or have similar ones. Moresampler has a totally unique flag system which I think deserves its own post.

Flags for the default resampler

a: stretches or compresses the consonant part (pink part in oto) of a sample. The default value is 100, so lower than that stretches, and higher than that compresses.  b: applies breathiness after synthesis, so it isn’t affected by the F flag. Default value is 0, range 0-100 B: applies breathiness before synthesis, and is affected by other flags. Default value is 50, range 0-100 c: changes value of C flag before formant filter adaptation. Default value: 50, range: 0-100. C: Cuts high frequencies. Default value is 0, range: 0-100. D: Cuts middle frequencies. Default value is 0, range: 0-100. E: Cuts low frequencies. Default value is 0, range: 0-100. F: Formant filter strength. Default value: 3. I think it can go up to 5 or so, beyond that it all starts to give similar results. g: gender factor flag. Positive values will deepen the voice, while negative ones will make it higher. Default value is 0, range: -100-100 G: forces resampler to render .frq files before playing. This makes the synthesis process slower, but gives you new frqs.  Flag is used by itself without a number added.  H: Emphasizes low tones, and cuts higher ones. Default value is 0, range: 0-99. h: Emphasizes higher tones, and cuts low ones. Default value is 0, range: 0-99. L: Sets a fixed frequency for F flag. Has no default value. Probably has the same limits as F flag. N: turns the formant filter off.  Flag is used by itself without a number added.  P: Peak compressor. Compresses the voice to make the volume more consistent. Default value is 86, Range: 0-100. t: Shifts the pitch by 10 cents for each point added. A cent is a unit in music. Default value is 0. W: Produces a sharper and more metallic sounding voice. Flag is used by itself without a number added. x: affects a note depending on how the sample it takes from was recorded. Higher notes make the voice brighter, while low notes make the voice more muffled. Negative values affect the formant. Default value is 0, Range: -100-100. Y: Controls the breathiness for the "vowel" part of a note. Default value is 100, range: 0-100.

Other resamplers’ flags are under the cut!