UTAU Voicebank Basics for Beginners
Creation/Recording
Before starting recording it is recommended to find a Reclist for the desired language in the desired format, and learning the Pronunciation of the used format.
The most basic for recording and recommended for beginners is a Japanese CV Library where each sample is a standalone file consisting only of consonant-vowel (Syllables) and few vowels + settings.
Oremo / Akorin / Recstar have its own tutorials how to set them up and use for recording.
By recording it is recomended to stay consistent with the voice, on the more advanced libraries it can be very significant if going out of pitch or tone.
For a first Library it is recommended to make a mono-pitch recording, adding aditional Pitch variations and/or Voice Colors can be done later after making the first steps.
After finishing the recordings it is optional to use an editing SW to align, cut, clean etc. the samples.
Some reclists have phomenes or other like breaths and glotal stops, these are not essential for a working Library, but can add some more liveness.
Some reclists includes a premade Oto file for starting, if not the UTAU, SetParam or Vlabeler can create them.
Library Files
There are few Files that have to be included in a Voicelibrary to berecognised and used by UTAU/OpenUTAU
character.txt (required)
txt file including information of the Voicebank's Character. These bases can be included and will be seen in the Profile window.
- "name=" = Character's Name in Romaji or Kana (in Shift-JIS), can be both - this one has to be added or the UTAU/OpenUTAU won't recognize the Library
- "author=" = Character/Voicebank Author (optional)
- "voice=" = Voice Provider (optional)
- "version=" = Library Version, Type or both (optional)
- "image=" = Profile Picture of the Character, in format "filename.bmp" (optional)
- "web=" = Character's/Creator's website (optional)
- "sample=" = wav sample to play, leaving blank it will use a random sound (optional)
- Other additional Information can be added but will be displayed as written underneath these.
oto.ini (required)
This file contains all the necessary data how to use Your samples to create singing.
readme.txt (optional)
Additional Details about the Voicebank, Character, Library Encoding details, Usage, Contacts etc.
Picture (semi-optional)
Profile Picture or Icon for the Character has to be in BMP format in resolution max 100x100px.
character.yaml (optional)
This File is used by OpenUTAU, this includes additional settings like the encoding, used phonemiser, Piano-roll Picture, Sub-Libraries etc.