Hey, so I run a tutorial channel on Youtube, UTAUtorial, and many of my viewers have mentioned that they like the fact that I provide closed captioning, since English is not their first language. So, I've had the thought to try and provide subtitles in languages other than English, but I am only bilingual [English and French], so I am looking for more people who might be interested in translating my videos.
Since my tutorials are free to access and I do not monetize off of them, I will not be able to pay you, so that is why I am simply asking for volunteers who want to help me make my videos more accessible.
I will provide the .sbv files, which can be edited in either Notepad or Notepad++, and I will credit you in my videos.
I'm just going to post the English transcript here. I might move them to Pastebin later though.
Reply to this thread or PM me if you are interested <3
Since my tutorials are free to access and I do not monetize off of them, I will not be able to pay you, so that is why I am simply asking for volunteers who want to help me make my videos more accessible.
I'm just going to post the English transcript here. I might move them to Pastebin later though.
Code:
Hey everyone this is TheEndIsNearUs with UTAUtorial, and today is the start of the brand new series, UTAU 101.
UTAU 101 is a free online course designed to teach new or inexperienced users the basics of the singing synthesis software UTAU.
Follow along as you learn to make your very own UTAU voicebank, including recording, configuration, and creating simple USTs.
These are my own personal methods, and are by no means the end-all-be-all way of doing things, and I hope that many of you will go on to find your own styles.
For this course, you will need two programs: UTAU, or course, and Oremo, a recording software specifically designed for UTAU voicebanks.
I will provide download links to both of these in the description.
You will also need the following sound equipment: A microphone, preferable one of decent quality, a headset or equivalent, and a quiet space with minimal background noise.
If you have access to a recording studio, amateur, professional, or otherwise, that would be optimal.
Lastly, you will need to download the tutorial kit, which contains everything else you will use for this tutorial, including the sample voicebank and UST that I will be making.
Let's get started, shall we?
First thing's first, we need to find a reclist, or "recording list."
A reclist is simply a list of the sounds that need to be recorded for a voicebank.
A full understanding of reclists requires knowledge of both phonetics and phonology, but for the purpose of this course, I will be keeping my explanations simple and straightforward.
There are many different styles, often named for what kind of samples they contain.
For instance, today we will be making a Japanese CV, or "consonant-vowel," voicebank, meaning that most of the samples will be diphones consisting of two sounds, a consonant followed by a vowel, optimized for the Japanese language.
This is arguably the simplest kind of voicebank, but because it is very limited, it can be more difficult to get high quality results.
Here we have a basic Japanese CV reclist.
This one is written in both hiragana and romaji, two of the Japanese alphabets, and it includes a few extra sounds.
If you don't know hiragana, but are planning to use UTAU for Japanese songs, I would highly recommend learning it; it will make using the software much easier.
Since I have an English computer, I prefer to encode my voicebanks in romaji, so I that is the reclist that I will be using.
Now, pronouncing Japanese is fairly easy, but there are a few sounds that can be more difficult for English speakers such as myself.
I feel I have decent pronunciation, but I still have an American accent, so, if that is important to you, I would suggest listening to native speakers and trying to mimic them.
First, we have the lone consonant /n/.
Then, the vowels, /a/, /i/, /ɯ/, /e/, and /o/.
Be care on /e/ and /o/ that they do not become the diphthongs /ej/ and /ow/, as they are in English.
It is a subtle difference, but a noticeable one.
Also, the "u" should be pronounced more as /ɯ/, rather than /u/.
Most consonants are pronounced as they would be in English, with a few exceptions.
First, the voiceless plosives "k," "t," and "p," should be unaspirated, if possible.
My voice is not capable of doing that as of now, but it is worth noting.
Next, "r" should be pronounced as /ɾa/, not /ɹa/.
It is the same consonant that is found in the English word "ladder," so feel the way your tongue moves in the middle of that word and try to recreate it at the start of the sample.
If you can't pronounce it properly, it is best to record /ra/ or /la/ instead, as they are the closest common equivalents.
The "ts" sound is pronounced as /tsa/, like the end of the word "cats."
The "f" should be pronounced with the lips, not the teeth, as /ɸa/ rather than /fa/.
Lastly, the CCV samples which include the liquid "y" should be pronounced with a very short /i/ before the main vowel.
For instance, /kja/ or /nja/.
We have a few of these in English, such as the word "cue," so they are not as tricky as they look.
Alright, so now we're going to learn how to use Oremo.
So go ahead and open up the software - which, I just have it my program files; your's will be wherever you downloaded it to - and you should get a window like this.
The layout's pretty easy to understand.
Over here you have your reclist, which you can load any of them in, as long as they're a txt file.
Here is where the suffixes will be listed.
I don't actually have any loaded, but just keep it on the blank one; we'll mess with those later.
Here is where the waveform will be displayed once we've actually recorded something, and here is the folder that all of our wavs will be saved to.
So the good - the cool thing about Oremo is that, after you record something, it will automatically save it as a wav file under that name to this folder, so it saves you a lot of time, because you don't have to manually export anything; it does it all automatically.
So what we're going to do first is go to file > load recordlist and find the one that I included in the tutorial kit.
I'm going to be using the romaji one, because I figure most of you probably don't know how to read hiragana, and I prefer to encode in romaji anyways, but you can use the hiragana one if you want to; it's the same list.
So open that up, and now it's displayed here at the left, and click on the first one, because that's the one we're going to be recording.
So, next what you do is open up view > pitch guide, and this will open up the tuning fork, which, what this does is it will play a long continuous note on the set pitch until you tell it to stop, like this.
[tuning fork plays]
And it will just play that.
It's really useful for when you feel like you might be getting out of tune, and you need to find your pitch again, which, for what we're doing, we're going to be just recording one pitch, so it's good to be familiar with your own vocal range, and know what pitches that you're comfortable with.
If you don't have any professional training, or you're just not aware, then go through the different pitches and such that they have listed and just try to match it, like this
[matching pitch on an "a"]
See, I started off really sharp, but I brought it down.
And then just figure out one that you're comfortable with, and use that to record.
So, I'm going to be recording at a G3, which is very close to the the pitch that I speak at; it's usually what I record at, just because I'm comfortable with it, and it's just a decent pitch for me, but it's going to be different for everyone, so just, like I said, experiment, figure out what your range is, figure out where you feel comfortable singing.
So next, you're going to go to options > show BGM guide, which I just opened up.
Now what a BGM guide is is an audio file that will play as you record, and what this does is it helps you keep in time and keep in pitch, so this will actually play, like -
[BGM guide plays]
It sounds like that, at least the one's that I've included.
They sound like that, and they'll play as you record, and you record along with them.
So, once you've found your pitch, go through the folder of the guide BGMs that I've included in the kit, and find where it's listed.
I have all them from C3 to C5 included, so I've already got mine set to G3.
And make sure you're on semi-automatic recording, which means that when you press "r," the BGM will start to play, and then it will stop playing once it's finished, and stop recording once the audio has finished.
But don't go to automatic recording unless you want to just keep going, like, it will start playing it again and start recording it again immediately after it stops, so I definitely prefer semi-auto recording, so I can use the guide BGM, but I can still go at my own pace.
So that's basically all the setup that we need, so we're ready to start recording.
Make sure that you're wearing headphones, so that none of the guide BGM or the tuning fork gets picked up in your recording, and also make sure that you're in a quiet location with minimal background noise.
Alright, so, let's get started, and I'll just show you - I'll just go through and record a couple of these and just show you how it's done.
So first I'm going to match my pitch, and then I'll record a couple things.
[matching pitch on an "a"]
♪ n a i u e o ♪
I think I'm really sharp, actually.
So, if you make a mistake like I just did, you can actually just record again, and it will do it over it.
So I'm going to match my pitch again, because I think I was really sharp.
[matching pitch on an "a"]
There we go, open up my throat a little bit.
♪ n a ♪
Yeah these are turning out much better.
♪ i u ♪
That one was still kind of weird.
So, also it's okay if you have a little bit of stuff going on at the end there, because that won't be - we can cut that out in the software, so you don't have to worry about having perfect silence at the beginning and at the end.
Also notice how I'm breathing in when the chime plays, and then holding the note for two beats, and then breathing out.
You can hold it for longer if you want, but I've found that two beats, which is approximately 1000 milliseconds, is a pretty good length.
♪ u ♪
Hmm, sharp again.
[matching pitch on an "u"]
♪ u ♪
Yeah, so sometimes it takes several attempts to get it how you want.
You don't have to be nearly as much of a perfectionist as I am, but, I mean, it's only going to make it higher quality.
[matching pitch on an "e"]
♪ e o ka ♪
So yeah, I mean, I'm going to go ahead and record the rest of these off camera, just because I figure you guys probably don't want to sit there for a half hour or however long it's going to take me to record all these, so I'm going to go ahead and cut out, and then come back in.Reply to this thread or PM me if you are interested <3
Last edited:
![[K] of Honor](/data/medal/1_1406930084l.jpg)