VCV OTO help

[Rakuei]

Hey guys, so I'm relatively new to UTAU and Vocaloid, specifically to various usage and semantics needed to properly use the program.

Regardless, I've started recording samples for my first VCV VB in Utau and have NO idea how to OTO it. Despite hours of research and tutorial reading online, nothing makes much sense to me. So let me show you my problem.

First off:

here is my list of current samples that I've already recorded. The oto I am using for reference, sample-wise is Ritsu's Kire VB. Notice I have no presets in any of my samples. I have no idea what any of those settings mean, so is there a preset number I should be using for most of them?

Second:

This is where I'm having the most trouble. I have four lines, and I have NO idea where to put them. I know the red line is supposed to go where the consonant starts, but where is the consonant in the "a" sample in an "aeauea" sample? I know I sound ridiculous.

See that big gaping spot? I can't tell where my a starts and ends from my "i" sample.

Someone please explain where I need to put these lines! Lol

 

[partial]

For preutterance and overlap I recommend using uniform values. Common numbers used for those are 300 for preutterance and 60 or 100 for overlap.

For recordings with aliases that start with - (like - か) it's totally fine if the preutterance and overlap values aren't uniform provided you only use it after a rest. Your parameters look fine in that first screenshot.

For something like "a あ" it's a bit trickier to tell, but you basically have to drag the blue line around to find that second "a" sound, if that makes sense.

I know the red line is supposed to go where the consonant starts, but where is the consonant in the "a" sample in an "aeauea" sample?

This depends. Do you mean - あ or a あ?
For - あ, the red line just goes as the start of the a sound.
For a あ, the red line goes at the start of the second a sound.

I hope I'm making some sense.

 

[Rakuei]

I understood the difference between -あ and あ。I don't know what the difference between the two are, but at least I can differentiate them, lol.

I don't know anything about parameters, preutterances, overlap, have no idea what any of that is.

I'm very new, lol.

What's the big gaping spot in the regular あ sample? I can't figure out where to drag the lines because I don't know where the あ starts and begins in the long vcv sample. The whole thing isn't just the あ sound, of course.

 

[Pupperoni Pizza]

I'm on mobile, so I can't post pretty images. But I'll try to explain as well as I can in words only.

Long wall of text incoming.

When OTOing in the UTAU software, there are arguably two of the most important settings, overlap and preutterence.

Preutterence = red line
Overlap = green line

Where the preutterence is placed, everything to the left of it will be adjusted to be played before the note, this also allows for space to place the overlap within that area.
The overlap dictates how far into the preutterence space two notes may blend together, it's important to find a proper balance as too much overlap can be worse than too little.
While you are able to invert the settings and have the overlap after the preutterence, I advise that you avoid doing so in most, if not all, situations.

In the case of VCV OTOing, those rules still apply.

The overlap in VCV OTOing ahould be a high enough value to create an adequate blending between two sounds, while also being low enough to cause quirks and glitches during track rendering.
My personal recommendation are values between 80 to 150 ms.
Based on your OTO reference you will be OTOing with an overlap of 100ms, a very common standard.

For easier configuration, it is best to keep the overlap uniform/the same for all VCV samples, deviation from that value is accepted and recommended for CV samples, but not required.
A good CV overlap for general blending are values between 30ms and 45ms, lower values aimed towards hard-type consonants, such as t/k/b/p/d/4/ch, and higher values being tailored for softer and more fluid consonants, such as s/sh/m/n/z.

with a VCV oto, the overlap should almost always be placed on/before the fade-off of the previous vowel.
So in "a ke", the overlap would be placed right before the fade-off in the previous "a" vowel.
Deviation from this is not recommended for novice users as it creates more need for fine adjustment.
The same rule applys in all cases.

For preutterence, the preutterence value is always placed after the consonant, but before the vowel.
In "a ke", it would be placed between the "k" consonant and the "e" vowel.
In situations where you are OTOing a fluid consonant, use of the spectrogram will provide a cleared view (this can be accessed via the [ s ] button in the oto configuration window.)

The final part that requires more work are VV samples, also known as vowel blends and/or vowel strings.
Vowel blends are otoed much the same as VCV samples, but you need to focus on the blending between the vowels.

For like-vowel blends, such as "a a" or "u u", simply setting you overlap and preutterence values the same within the like-vowel region, in your case this is 100ms, will provided sufficient smoothness with those blends without too much fuss.

For different vowel blends, such as "e u" or "a i", the transition between those sounds determines the placement of the values.
The overlap parameter would best be placed before the the transition and the preutterence is best placed after, this is the quickest and easiest way to determine the best locations for those two settings.

End-vowels/end breaths are a somewhat off topic subject.
End vowels are often aliased as "a -", "a R", or "a br"
The overlap goes on the vowel's fadeout; the preutterence can simply be placed right at the tip/end of the vowel and/or before the breath if one is present in the sample.
Increasing of the overlap value to 150ms and adjusting accordingly is an optional task to improve end breath blending.

The final part I will cover are starting vowels, like "- a" or "- o".
Quick and easy settings would be an overlap of 30ms and a preutterence of 65ms.
Simply line it up so the preutterence is placed at the very beginning of the vowel.

That should just about cover it!
If you have any questions, feel free to ask.
I can provide images shortly.
[DOUBLEPOST=1431233671][/DOUBLEPOST]Update: uniform values on the preutterence can be very sloppy and often causes blending issues with many short, successive, notes.

 

[Rakuei]

Please do!! I just read all that and it all went over my head, like I was reading from a textbook while still in the first day of classes.

Can we focus on my starting sample of "ああいあうえあ"? That is the sample I am oto'ing currently. So there are no consonants.

Can I put preset numbers in without having to move the lines, or will I still have to do that?

I'm sorry, I'm such a noob. I guess I need terminology I can chew in pieces first. I get this feeling that once I can understand one thing, it will all click into place for me.

I'm sorry to be such an obtuse novice!

 

[Pupperoni Pizza]

Okay, let me make an image from one of yours amd upload it

 

[Rakuei]

Okay, let me make an image from one of yours amd upload it

Thank you!!!

Again, I'm sorry! Like I said, I feel like once I understand one thing, I'll be able to understand everything else, too.

 

[Rakuei]

So here's my next problem. I put in some notes, and converted it to VCV, and my bank (still being oto'd) pronounces the notes just as they're listed, with the (a) before the い. I thought it was just my bank, so I tried Ritsu, and the same thing happens with that bank too. What is it and how to I fix it? I've looked everywhere lol