| Tweet |
| Welcome to the UTAforum! UTAU is a Japanese voice-synthesis program used to create singing vocals for music. It is much like the more popular, commercially sold Vocaloid products, but UTAU is freeware! To learn more about UTAU and various UTAU voices/characters, you can click here. You're currently viewing our forum as a guest. This means you are limited to certain areas of the board and there are some features you can't use. If you join our community, you'll be able to access member-only sections, and use many member-only features such as customizing your profile, sending personal messages, and voting in polls. Registration is simple, fast, and completely free. Join our community! If you're already a member please log in to your account to access all of our features: |
| OTOing like a boss | |
|---|---|
| Tweet Topic Started: Nov 29 2011, 10:18 PM (501 Views) | |
| Savy | Nov 29 2011, 10:18 PM Post #1 |
|
It boggles little Savy's crazy brain how many people do not understand how to create an oto.ini. I was there once, yes, and I honestly don't remember how I didn't get it. I didn't understand what I was doing either, and messing with the oto scared the crap out of me. And now I edit every oto into oblivion. I can't teach you how to oto flawlessly, as a good chunk of people here still won't understand when I'm done, but I'll be damned if I don't teach at least one person. So! Let's get on with it! Fair warning, we're about to get image-heavy. First, open up your voicebank's voice configuration. (Second-to last tab, first option, should read "voice bank settings" for English versions) I'm re-OTOing Fuu. She's CV, but this tutorial should at least help a little bit with other types of banks. ![]() I'll assume you already have freqency maps generated, so select the sound you want to oto and listen to it two or three times beforehand, because this will help. Set all those numbers to the right to '0' if they aren't already. Now click "Launch Editor". You'll get something that looks like this: ![]() Don't worry if it doesn't look the same, no two sounds do. I have my view zoomed out so you can see the whole thing. First, we'll deal with the cutoff. It's that part at the end you don't want. Perhaps your mic picked up your mouse clicking, or you exhaled, whatever it is, you need it gone in order for your sound to mesh with the others. For Fuu's 'go', it's this part. It probably looks like your sample's end, too. ![]() Just drag the end to where you want your sound to cut off. It'll be highlighted in blue, like this: ![]() For those who didn't quite catch it the first time (like me), this blue part WILL NOT PLAY. Don't be afraid to clip the end of the sample. I like to do it because my sounds mesh together better, but it's your preference. You don't need to cut a metric crapton off, but if by some chance you do, leave AT LEAST half a second (0.5) of your sample left. I hear this is the best length for a sample, but if it's longer, it won't hurt. It will actually cut down on the metallic scream your UTAU makes when the note they're singing is being stretched farther than the sample's length. Now we'll deal with the offset. It's the same thing as cutoff, but for the front of the note; cut out everything before the beginning of your sound, or else your UTAU will sing it. For Fuu's 'go', the offset looks like this: ![]() Now, some banks don't have any sound before the sample. Maybe yours doesn't, I don't know, but you do. If your sample is cut where there's nothing but your voice, leave the offset at 0. Same thing goes for the cutoff, unless you want to trim the end of the sample. Next, click the S button in the top left. Right next to the P (play, double-click to hear what's left after offset and cutoff). You'll get a spectrogram view that looks like this: ![]() SWEET JEBUS THIS IS TERRIFYING. Don't worry about the part before the pink line. That's what we wanted cut off. S'all good. We'll deal with that green line now, the overlap. OHHH MESSING WITH IT. SCARY. Well, calm down, get out from behind the couch, it's okay. Go to a happy place and we'll resume. Now, as I said, this is the overlap. Anything behind this will overlap your envelopes when you fit your UST to your UTAU VB. It sounds smoother because it's placing the beginning of your sound's envelope over the end of the last envelope. It really depends how much overlap you'll need with each sound. For something like 'chi', you'll want to overlap around half the 'ch'. For something like 'tsi', you'll want to overlap almost the whole 's' after cutting some of the beginning off. 'z'...will make this really weird sound if you don't offset part of the 'zz' and add the right amount of overlap. Some sounds like 'd' or the vowels don't need overlap. There is no right way to do this, only a wrong one, so don't twist yourself up about it. It's only done right when you decide it sounds right. Now, for something like Fuu's 'go', or anyone's 'go', you'll want to overlap the beginning of the 'g'-part. Using the blue lines (or in some cases, blocks) can really help you. ![]() Next we'll deal with the red line. It's listed as the preutter on the right side of the options box. This line, I can't really explain that well. Just place it between the middle of the sound. Example, for 'go', the sample will sound like 'gooooooo', and the red line sound go between the G and the O (where it's 'g|ooooooo). Like so: ![]() Reading a spectrogram like this isn't the easiest thing, so make sure to double-click the P button plenty of times to listen to your sound. Eventually it'll dawn on you where the red line should go. Okay! We're nearly done! Last we have the consonant; it's that section in the pink area! You should only select what you don't want stretched. The closer to the red line this is, the better your UTAU will sing fast songs. What gets selected in the pink cannot and will not be stretched when UTAU needs a longer note. It's the part that gets played first when your envelope is down, and if your pink section is longer than your note envelope, it's going to sound really, really bad when your UTAU sings. I like to oto extra-tight, like this: ![]() That's just how I do it. You can totally do the same, thought! In general, you only need a little bit of the whole sound selected in pink ('[go]oooooo'), but go ahead and play with it. Whatever sounds best. Some people speak in such a way when they record that you just have straight lines. If your VB is like this, or you're otoing a VB like this, I hope your soul has pity taken on it, because it's REALLY hard. If it is your VB that's like this and you can't tell what you are doing, just re-record it. It'll save you a lot of pain in the long run. If it's someone else' VB though, you might want to leave it to someone with a bit more OTOing experience. That's it for this tutorial, hopefully I didn't offend anyone because I was being silly, that tends to happen. Now go! BE FREE~! OTO UNTIL YOUR HEART IS CONTENT WITH YOUR NEW-FOUND KNOWLEDGE! (Also, could someone maybe pin this? It'll help a lot of the newbies and cut down on topics asking for oto help.) Edited by Savy, Nov 29 2011, 10:28 PM.
|
![]() Art by Subarashiki | Click for VB download Guess the music, win a chibi | |
![]() |
|
| LyrieYumi | Dec 3 2011, 09:22 AM Post #2 |
|
god this is what i need. ... i still don't get it. *stares at ACT 2 VB which is now in shreds. |
![]() download act 2.1 here | |
![]() |
|
| The_Eternal_Child | Dec 17 2011, 05:01 PM Post #3 |
|
Is there an advantage to using the spectrogram over the normal editing screen? It seems to me like it just complicates things. o.o |
![]() |
|
| Aleksandr | Dec 17 2011, 05:14 PM Post #4 |
|
Kanaya West
|
Spectrogram view is great for sounds that are hard to pinpoint; like r's, y's, and w's, as well as diphthongs. I thought it was weird at first too, but now I can't do without it~ |
|
I will push my ships on you, always. Aleks' World of Fun // Ask!Zhenya | |
![]() |
|
| 1 user reading this topic (1 Guest and 0 Anonymous) | |
| « Previous Topic · Tutorials and Resources Directory · Next Topic » |



















2:12 AM Jul 11



