Project UTAU

on

Project Utau

Utau or Utauloid is a free open source version of Vocaloid.

Despite the many stereotype Utauloid creator get for the childish or cringeful content they put out. It is really impressive that the hardwork they put in, in order to create their voice bank. So I pretty much sought out in making my own voicebank that I could possible use for other things, such as making robotic voiceline and TTS(thinking about it will be hard to make the (.ust) for each assignment, I guess I can reuse audio asset for a TTS [potential next project]).
Update[7/10/22]: Apparently there's an application called Boyomi-chan and Utayomi that could do this. So I'll definitely check it out

So far I have made 2 attempts at making voice since, first one following tutorial step-by-step using their resource (this one covered in this run through), second attempt being just use whatever the program initially gave.

PROJECT DETAILS

Basic Required: 

  • Hardware:
    • A computer
    • A microphone
  • Software:
    • UTAU
    • Oremo
    • setParam
    • Any type of DAW (FlStudio)
  • Patience and the vocal power to record more than 100 audio takes

Difficulty :

Exponential to how impatient you are

Brief Process Overview :

This is just a rough overview, I'll link important tutorial links regarding each of the steps, later I'll update my personal explanation of the step in my own understanding and some tips from my sensei.

INSTALLING UTAU [1]
Refer the tutorial video in the reference section.


CREATING A VOICEBANK
The voice bank is what the Utau software will take the voice sampling and stitch it all together to create the synthetic singing
  1. Install all require software
  2. Record voice bank using Oremo based on what type of voicebank you want to create (IMPORTANT TO TAKE NOTE OF). You'll end up with a whole folder of wav file
  3. Set the pre-utterance and other parameter using setParam, this will result in the creation of an oto.ni file. [2]
  4. You're all set, just need to create a detail text file and add an image.
Property of my Voice Bank

A glimpse of the voicebank folder


CREATING @ DOWNLOADING AN UST
Now that we have the voice bank finished now we need to make a pattern to put those voice samples to make it sounds like a singing voice. Which mean we need to make an UST or download one. 
    UST[3], short for Utau Sequence Text, is similar to what VSQ(Vocaloid SeQuence file) is to an Vocaloid Software. 
    As for now I downloaded the UST from an existing utau project and replace the voicebank with the one we made, for the sake of accomplishing a milestone and get motivated to do more. You can pretty much find bunch of ust of various song shared by others. [4]
    1. Open the ust project
    2. Replace the voice bank with yours
    3. Readjust the pitching and crossfade to suit your voicebank
    4. Render and export your UST

AND WE FINISH OUR FIRST PROJECT
As for now we focus on getting our voicebank to work, later down the line , we'll try to create our own song and ust. As for now enjoy the attempts even though it is very scuffed.

If you wanna refer tutorials, I would suggest Yunari's Tutorial. Its been really useful for me. I left the link to her UTAU playlist in the reference section.[5]

n(Attempt):
First Attempt

Second Attempt 

References

Comments

Post a Comment