Voice AI Nyaa~
Super Simple Tutorial To Get Started!!!
Because, I just literally learn an hour ago and have no idea what I'm doing, ehehe~
Step 1 - Get Some Voices!!!
You can find the voice models here nyaa -> Voice Models
Just click one of the sheets!
Now, pick one of the model you want. For the purpose of this tutorial, please pick the Gura model.
(because shark is cute, ehehe~)
See the arrow? Scroll to the right a little and copy the link.
Once you have that. Time for step 2!!!
Step 2 - Open Collab!!
Open this link nyaaa~ (No, I don't make this colab, someone else did, ehehe~)
The WebUI but for Voice!
Now press the play button!
After pressing the play button scroll down just a bit~

Put the model link there! Also, do you have cute crab in your colab too? Ehehe~
After you put the model link, press the play button! Then wait.
Once done, press play button! At the left, forgot to add arrow! If you are confused on what to click in this step, let me know on discord, humu humu~

In this part, press the play button again and wait until you get the link!

Press the gradio link! As seen in the arrow, humu humu~
It will send you to the webui! Finally~ Ehehe...
Step 3 Let's Synthesize!!!

Okay! So, first choose the model/voice you like. There should be only one there~
Now press convert! It will sound terrible... But hey, at least it works???
Now, let's work on making it sound good!
First off, scroll down a little and you will see the test audio being used.
As you can see, it's using some dude's voice! The problem is, in this example, we are trying to generate Gura's voice, who is a very cute shark girl.
To fix that, we need to meddle with the pitch!

Yes, up here, there's the pitch! You need to change it
Basically, if the voice is very deep, you might wanna go up to 16
Now press convert! Now it should sound loads better, humu humu~
There are more advanced stuff, but I will leave you to experiment for yourself. This is to get started.
Each pitch increase is one semitone up! So 12 is one octave up. Useful if you wanna make AI Cover.
But really, it's all just trial and error.
You can even press the recording button to record yourself playing~
Important
If you find any error, try clicking refresh button near the audio select (the one at the bottom).
Some model doesn't seem to work and that's because of my skill issue. (Again, just started using it one hour before making this rentry.)
In any case, hope you're all inspired to start your ai voice generation journey. Have fun now~
Have a nice day!!!